From ad8ec177efe5ea92811611fbb483781105c26889 Mon Sep 17 00:00:00 2001
From: Yang Zhou <goat.zhou@qq.com>
Date: Thu, 24 Mar 2022 10:51:59 +0800
Subject: [PATCH 1/9] add tlg decoder

---
 speechx/examples/decoder/CMakeLists.txt       |    4 +
 .../decoder/offline_wfst_decoder_main.cc      |  101 +
 speechx/speechx/decoder/CMakeLists.txt        |    3 +-
 .../speechx/decoder/ctc_beam_search_decoder.h |    2 +-
 speechx/speechx/decoder/ctc_tlg_decoder.cc    |   50 +
 speechx/speechx/decoder/ctc_tlg_decoder.h     |   47 +
 speechx/speechx/kaldi/CMakeLists.txt          |    3 +
 speechx/speechx/kaldi/decoder/CMakeLists.txt  |    6 +
 .../{nnet => kaldi/decoder}/decodable-itf.h   |    2 +-
 .../kaldi/decoder/lattice-faster-decoder.cc   |    4 -
 .../kaldi/decoder/lattice-faster-decoder.h    |    3 +-
 .../decoder/lattice-faster-online-decoder.cc  |    4 +-
 .../decoder/lattice-faster-online-decoder.h   |    2 +-
 speechx/speechx/kaldi/fstext/CMakeLists.txt   |    5 +
 .../kaldi/fstext/determinize-lattice-inl.h    | 1357 ++++++
 .../kaldi/fstext/determinize-lattice.h        |  144 +
 .../kaldi/fstext/determinize-star-inl.h       | 1204 ++++++
 .../speechx/kaldi/fstext/determinize-star.h   |  116 +
 speechx/speechx/kaldi/fstext/fstext-lib.h     |   34 +
 .../speechx/kaldi/fstext/fstext-utils-inl.h   | 1265 ++++++
 speechx/speechx/kaldi/fstext/fstext-utils.h   |  386 ++
 .../speechx/kaldi/fstext/kaldi-fst-io-inl.h   |  208 +
 speechx/speechx/kaldi/fstext/kaldi-fst-io.cc  |  148 +
 speechx/speechx/kaldi/fstext/kaldi-fst-io.h   |  158 +
 .../speechx/kaldi/fstext/lattice-utils-inl.h  |  267 ++
 speechx/speechx/kaldi/fstext/lattice-utils.h  |  259 ++
 speechx/speechx/kaldi/fstext/lattice-weight.h |  892 ++++
 .../kaldi/fstext/pre-determinize-inl.h        |  798 ++++
 .../speechx/kaldi/fstext/pre-determinize.h    |   98 +
 .../kaldi/fstext/remove-eps-local-inl.h       |  318 ++
 .../speechx/kaldi/fstext/remove-eps-local.h   |   57 +
 speechx/speechx/kaldi/fstext/table-matcher.h  |  387 ++
 speechx/speechx/kaldi/lat/CMakeLists.txt      |    6 +
 .../lat/determinize-lattice-pruned-test.cc    |  147 -
 .../kaldi/lat/determinize-lattice-pruned.cc   |  468 +--
 .../kaldi/lat/determinize-lattice-pruned.h    |  156 +-
 speechx/speechx/kaldi/lat/kaldi-lattice.h     |   16 +-
 .../speechx/kaldi/lat/lattice-functions.cc    | 3632 +++++++++--------
 speechx/speechx/kaldi/lat/lattice-functions.h |  767 ++--
 speechx/speechx/nnet/decodable.cc             |   12 +-
 speechx/speechx/nnet/decodable.h              |    6 +-
 41 files changed, 10943 insertions(+), 2599 deletions(-)
 create mode 100644 speechx/examples/decoder/offline_wfst_decoder_main.cc
 create mode 100644 speechx/speechx/decoder/ctc_tlg_decoder.cc
 create mode 100644 speechx/speechx/decoder/ctc_tlg_decoder.h
 create mode 100644 speechx/speechx/kaldi/decoder/CMakeLists.txt
 rename speechx/speechx/{nnet => kaldi/decoder}/decodable-itf.h (99%)
 create mode 100644 speechx/speechx/kaldi/fstext/CMakeLists.txt
 create mode 100644 speechx/speechx/kaldi/fstext/determinize-lattice-inl.h
 create mode 100644 speechx/speechx/kaldi/fstext/determinize-lattice.h
 create mode 100644 speechx/speechx/kaldi/fstext/determinize-star-inl.h
 create mode 100644 speechx/speechx/kaldi/fstext/determinize-star.h
 create mode 100644 speechx/speechx/kaldi/fstext/fstext-lib.h
 create mode 100644 speechx/speechx/kaldi/fstext/fstext-utils-inl.h
 create mode 100644 speechx/speechx/kaldi/fstext/fstext-utils.h
 create mode 100644 speechx/speechx/kaldi/fstext/kaldi-fst-io-inl.h
 create mode 100644 speechx/speechx/kaldi/fstext/kaldi-fst-io.cc
 create mode 100644 speechx/speechx/kaldi/fstext/kaldi-fst-io.h
 create mode 100644 speechx/speechx/kaldi/fstext/lattice-utils-inl.h
 create mode 100644 speechx/speechx/kaldi/fstext/lattice-utils.h
 create mode 100644 speechx/speechx/kaldi/fstext/lattice-weight.h
 create mode 100644 speechx/speechx/kaldi/fstext/pre-determinize-inl.h
 create mode 100644 speechx/speechx/kaldi/fstext/pre-determinize.h
 create mode 100644 speechx/speechx/kaldi/fstext/remove-eps-local-inl.h
 create mode 100644 speechx/speechx/kaldi/fstext/remove-eps-local.h
 create mode 100644 speechx/speechx/kaldi/fstext/table-matcher.h
 create mode 100644 speechx/speechx/kaldi/lat/CMakeLists.txt
 delete mode 100644 speechx/speechx/kaldi/lat/determinize-lattice-pruned-test.cc

diff --git a/speechx/examples/decoder/CMakeLists.txt b/speechx/examples/decoder/CMakeLists.txt
index 4bd5c6cf..11e2ca91 100644
--- a/speechx/examples/decoder/CMakeLists.txt
+++ b/speechx/examples/decoder/CMakeLists.txt
@@ -3,3 +3,7 @@ cmake_minimum_required(VERSION 3.14 FATAL_ERROR)
 add_executable(offline_decoder_main ${CMAKE_CURRENT_SOURCE_DIR}/offline_decoder_main.cc)
 target_include_directories(offline_decoder_main PRIVATE ${SPEECHX_ROOT} ${SPEECHX_ROOT}/kaldi)
 target_link_libraries(offline_decoder_main PUBLIC nnet decoder fst utils gflags glog kaldi-base kaldi-matrix kaldi-util ${DEPS})
+
+add_executable(offline_wfst_decoder_main ${CMAKE_CURRENT_SOURCE_DIR}/offline_wfst_decoder_main.cc)
+target_include_directories(offline_wfst_decoder_main PRIVATE ${SPEECHX_ROOT} ${SPEECHX_ROOT}/kaldi)
+target_link_libraries(offline_wfst_decoder_main PUBLIC nnet decoder fst utils gflags glog kaldi-base kaldi-matrix kaldi-util kaldi-decoder ${DEPS})
diff --git a/speechx/examples/decoder/offline_wfst_decoder_main.cc b/speechx/examples/decoder/offline_wfst_decoder_main.cc
new file mode 100644
index 00000000..7c1f6226
--- /dev/null
+++ b/speechx/examples/decoder/offline_wfst_decoder_main.cc
@@ -0,0 +1,101 @@
+// Copyright (c) 2022 PaddlePaddle Authors. All Rights Reserved.
+//
+// Licensed under the Apache License, Version 2.0 (the "License");
+// you may not use this file except in compliance with the License.
+// You may obtain a copy of the License at
+//
+//     http://www.apache.org/licenses/LICENSE-2.0
+//
+// Unless required by applicable law or agreed to in writing, software
+// distributed under the License is distributed on an "AS IS" BASIS,
+// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+// See the License for the specific language governing permissions and
+// limitations under the License.
+
+// todo refactor, repalce with gtest
+
+#include "base/flags.h"
+#include "base/log.h"
+#include "decoder/ctc_tlg_decoder.h"
+#include "frontend/raw_audio.h"
+#include "kaldi/util/table-types.h"
+#include "nnet/decodable.h"
+#include "nnet/paddle_nnet.h"
+
+DEFINE_string(feature_respecifier, "", "test feature rspecifier");
+DEFINE_string(model_path, "avg_1.jit.pdmodel", "paddle nnet model");
+DEFINE_string(param_path, "avg_1.jit.pdiparams", "paddle nnet model param");
+DEFINE_string(word_symbol_table, "vocab.txt", "word symbol table");
+DEFINE_string(graph_path, "TLG", "decoder graph");
+
+
+using kaldi::BaseFloat;
+using kaldi::Matrix;
+using std::vector;
+
+int main(int argc, char* argv[]) {
+    gflags::ParseCommandLineFlags(&argc, &argv, false);
+    google::InitGoogleLogging(argv[0]);
+
+    kaldi::SequentialBaseFloatMatrixReader feature_reader(
+        FLAGS_feature_respecifier);
+    std::string model_graph = FLAGS_model_path;
+    std::string model_params = FLAGS_param_path;
+    std::string word_symbol_table = FLAGS_word_symbol_table;
+    std::string graph_path = FLAGS_graph_path;
+
+    int32 num_done = 0, num_err = 0;
+
+    ppspeech::TLGDecoderOptions opts;
+    opts.word_symbol_table = word_symbol_table;
+    opts.fst_path = graph_path;
+    ppspeech::TLGDecoder decoder(opts);
+
+    ppspeech::ModelOptions model_opts;
+    model_opts.model_path = model_graph;
+    model_opts.params_path = model_params;
+    std::shared_ptr<ppspeech::PaddleNnet> nnet(
+        new ppspeech::PaddleNnet(model_opts));
+    std::shared_ptr<ppspeech::RawDataCache> raw_data(
+        new ppspeech::RawDataCache());
+    std::shared_ptr<ppspeech::Decodable> decodable(
+        new ppspeech::Decodable(nnet, raw_data));
+
+    int32 chunk_size = 35;
+    decoder.InitDecoder();
+
+    for (; !feature_reader.Done(); feature_reader.Next()) {
+        string utt = feature_reader.Key();
+        const kaldi::Matrix<BaseFloat> feature = feature_reader.Value();
+        raw_data->SetDim(feature.NumCols());
+        int32 row_idx = 0;
+        int32 num_chunks = feature.NumRows() / chunk_size;
+        for (int chunk_idx = 0; chunk_idx < num_chunks; ++chunk_idx) {
+            kaldi::Vector<kaldi::BaseFloat> feature_chunk(chunk_size *
+                                                          feature.NumCols());
+            for (int row_id = 0; row_id < chunk_size; ++row_id) {
+                kaldi::SubVector<kaldi::BaseFloat> tmp(feature, row_idx);
+                kaldi::SubVector<kaldi::BaseFloat> f_chunk_tmp(
+                    feature_chunk.Data() + row_id * feature.NumCols(),
+                    feature.NumCols());
+                f_chunk_tmp.CopyFromVec(tmp);
+                row_idx++;
+            }
+            raw_data->Accept(feature_chunk);
+            if (chunk_idx == num_chunks - 1) {
+                raw_data->SetFinished();
+            }
+            decoder.AdvanceDecode(decodable);
+        }
+        std::string result;
+        result = decoder.GetFinalBestPath();
+        KALDI_LOG << " the result of " << utt << " is " << result;
+        decodable->Reset();
+        decoder.Reset();
+        ++num_done;
+    }
+
+    KALDI_LOG << "Done " << num_done << " utterances, " << num_err
+              << " with errors.";
+    return (num_done != 0 ? 0 : 1);
+}
diff --git a/speechx/speechx/decoder/CMakeLists.txt b/speechx/speechx/decoder/CMakeLists.txt
index 7cd281b6..ee0863fd 100644
--- a/speechx/speechx/decoder/CMakeLists.txt
+++ b/speechx/speechx/decoder/CMakeLists.txt
@@ -6,5 +6,6 @@ add_library(decoder STATIC
   ctc_decoders/decoder_utils.cpp
   ctc_decoders/path_trie.cpp
   ctc_decoders/scorer.cpp
+  ctc_tlg_decoder.cc
 )
-target_link_libraries(decoder PUBLIC kenlm utils fst)
\ No newline at end of file
+target_link_libraries(decoder PUBLIC kenlm utils fst)
diff --git a/speechx/speechx/decoder/ctc_beam_search_decoder.h b/speechx/speechx/decoder/ctc_beam_search_decoder.h
index 451f33c0..cf1824c6 100644
--- a/speechx/speechx/decoder/ctc_beam_search_decoder.h
+++ b/speechx/speechx/decoder/ctc_beam_search_decoder.h
@@ -15,7 +15,7 @@
 #include "base/common.h"
 #include "decoder/ctc_decoders/path_trie.h"
 #include "decoder/ctc_decoders/scorer.h"
-#include "nnet/decodable-itf.h"
+#include "kaldi/decoder/decodable-itf.h"
 #include "util/parse-options.h"
 
 #pragma once
diff --git a/speechx/speechx/decoder/ctc_tlg_decoder.cc b/speechx/speechx/decoder/ctc_tlg_decoder.cc
new file mode 100644
index 00000000..c08a7d5b
--- /dev/null
+++ b/speechx/speechx/decoder/ctc_tlg_decoder.cc
@@ -0,0 +1,50 @@
+#include "decoder/ctc_tlg_decoder.h"
+namespace ppspeech {
+
+TLGDecoder::TLGDecoder(TLGDecoderOptions opts) {
+    fst_.reset(fst::Fst<fst::StdArc>::Read(opts.fst_path));
+    CHECK(fst_ != nullptr);
+    word_symbol_table_.reset(fst::SymbolTable::ReadText(opts.word_symbol_table));
+    decoder_.reset(new kaldi::LatticeFasterOnlineDecoder(*fst_, opts.opts));
+    decoder_->InitDecoding();
+}
+
+void TLGDecoder::InitDecoder() {
+    decoder_->InitDecoding();
+}
+
+void TLGDecoder::AdvanceDecode(const std::shared_ptr<kaldi::DecodableInterface>& decodable) {
+    while (1) {
+      AdvanceDecoding(decodable.get());
+      if (decodable->IsLastFrame(num_frame_decoded_)) break;
+    }
+}
+
+void TLGDecoder::AdvanceDecoding(kaldi::DecodableInterface* decodable) {
+  // skip blank frame?
+  decoder_->AdvanceDecoding(decodable, 1);
+  num_frame_decoded_++;
+}
+
+void TLGDecoder::Reset() {
+  decoder_->InitDecoding();
+  return;
+}
+
+std::string TLGDecoder::GetFinalBestPath() {
+  decoder_->FinalizeDecoding();
+  kaldi::Lattice lat;
+  kaldi::LatticeWeight weight;
+  std::vector<int> alignment;
+  std::vector<int> words_id;
+  decoder_->GetBestPath(&lat, true);
+  fst::GetLinearSymbolSequence(lat, &alignment, &words_id, &weight);
+  std::string words;
+  for (int32 idx = 0; idx < words_id.size(); ++idx) {
+    std::string word = word_symbol_table_->Find(words_id[idx]);
+    words += word;
+  }
+  return words;
+}
+
+}
\ No newline at end of file
diff --git a/speechx/speechx/decoder/ctc_tlg_decoder.h b/speechx/speechx/decoder/ctc_tlg_decoder.h
new file mode 100644
index 00000000..b4cd8c34
--- /dev/null
+++ b/speechx/speechx/decoder/ctc_tlg_decoder.h
@@ -0,0 +1,47 @@
+#pragma once
+
+#include "kaldi/decoder/lattice-faster-online-decoder.h"
+#include "kaldi/decoder/decodable-itf.h"
+#include "util/parse-options.h"
+#include "base/basic_types.h"
+
+namespace ppspeech {
+
+struct TLGDecoderOptions {
+   kaldi::LatticeFasterDecoderConfig opts; 
+   // todo remove later, add into decode resource 
+   std::string word_symbol_table;
+   std::string fst_path;
+   
+   TLGDecoderOptions() 
+       : word_symbol_table(""),
+         fst_path("") {}
+};
+
+class TLGDecoder {
+  public:
+    explicit TLGDecoder(TLGDecoderOptions opts);
+    void InitDecoder();
+    void Decode();
+    std::string GetBestPath();
+    std::vector<std::pair<double, std::string>> GetNBestPath();
+    std::string GetFinalBestPath();
+    int NumFrameDecoded();
+    int DecodeLikelihoods(const std::vector<std::vector<BaseFloat>>& probs,
+                          std::vector<std::string>& nbest_words);
+    void AdvanceDecode(
+        const std::shared_ptr<kaldi::DecodableInterface>& decodable);
+    void Reset();
+
+  private:
+    void AdvanceDecoding(kaldi::DecodableInterface* decodable);    
+
+    std::shared_ptr<kaldi::LatticeFasterOnlineDecoder> decoder_;
+    std::shared_ptr<fst::Fst<fst::StdArc>> fst_;   
+    std::shared_ptr<fst::SymbolTable> word_symbol_table_;
+    int32 num_frame_decoded_;
+  };
+
+    
+
+}  // namespace ppspeech
\ No newline at end of file
diff --git a/speechx/speechx/kaldi/CMakeLists.txt b/speechx/speechx/kaldi/CMakeLists.txt
index 414a6fa0..6f7398cd 100644
--- a/speechx/speechx/kaldi/CMakeLists.txt
+++ b/speechx/speechx/kaldi/CMakeLists.txt
@@ -4,3 +4,6 @@ add_subdirectory(base)
 add_subdirectory(util)
 add_subdirectory(feat)
 add_subdirectory(matrix)
+add_subdirectory(lat)
+add_subdirectory(fstext)
+add_subdirectory(decoder)
diff --git a/speechx/speechx/kaldi/decoder/CMakeLists.txt b/speechx/speechx/kaldi/decoder/CMakeLists.txt
new file mode 100644
index 00000000..f1ee6eab
--- /dev/null
+++ b/speechx/speechx/kaldi/decoder/CMakeLists.txt
@@ -0,0 +1,6 @@
+
+add_library(kaldi-decoder
+lattice-faster-decoder.cc
+lattice-faster-online-decoder.cc
+)
+target_link_libraries(kaldi-decoder PUBLIC kaldi-lat)
diff --git a/speechx/speechx/nnet/decodable-itf.h b/speechx/speechx/kaldi/decoder/decodable-itf.h
similarity index 99%
rename from speechx/speechx/nnet/decodable-itf.h
rename to speechx/speechx/kaldi/decoder/decodable-itf.h
index 8e9a5a72..19e07498 100644
--- a/speechx/speechx/nnet/decodable-itf.h
+++ b/speechx/speechx/kaldi/decoder/decodable-itf.h
@@ -121,7 +121,7 @@ class DecodableInterface {
     /// decoding-from-matrix setting where we want to allow the last delta or
     /// LDA
     /// features to be flushed out for compatibility with the baseline setup.
-    virtual bool IsLastFrame(int32 frame) const = 0;
+    virtual bool IsLastFrame(int32 frame) = 0;
 
     /// The call NumFramesReady() will return the number of frames currently
     /// available
diff --git a/speechx/speechx/kaldi/decoder/lattice-faster-decoder.cc b/speechx/speechx/kaldi/decoder/lattice-faster-decoder.cc
index 42d1d2af..ae6b7160 100644
--- a/speechx/speechx/kaldi/decoder/lattice-faster-decoder.cc
+++ b/speechx/speechx/kaldi/decoder/lattice-faster-decoder.cc
@@ -1007,14 +1007,10 @@ template class LatticeFasterDecoderTpl<fst::Fst<fst::StdArc>, decoder::StdToken>
 template class LatticeFasterDecoderTpl<fst::VectorFst<fst::StdArc>, decoder::StdToken >;
 template class LatticeFasterDecoderTpl<fst::ConstFst<fst::StdArc>, decoder::StdToken >;
 
-template class LatticeFasterDecoderTpl<fst::ConstGrammarFst, decoder::StdToken>;
-template class LatticeFasterDecoderTpl<fst::VectorGrammarFst, decoder::StdToken>;
 
 template class LatticeFasterDecoderTpl<fst::Fst<fst::StdArc> , decoder::BackpointerToken>;
 template class LatticeFasterDecoderTpl<fst::VectorFst<fst::StdArc>, decoder::BackpointerToken >;
 template class LatticeFasterDecoderTpl<fst::ConstFst<fst::StdArc>, decoder::BackpointerToken >;
-template class LatticeFasterDecoderTpl<fst::ConstGrammarFst, decoder::BackpointerToken>;
-template class LatticeFasterDecoderTpl<fst::VectorGrammarFst, decoder::BackpointerToken>;
 
 
 } // end namespace kaldi.
diff --git a/speechx/speechx/kaldi/decoder/lattice-faster-decoder.h b/speechx/speechx/kaldi/decoder/lattice-faster-decoder.h
index 2016ad57..d142a8c7 100644
--- a/speechx/speechx/kaldi/decoder/lattice-faster-decoder.h
+++ b/speechx/speechx/kaldi/decoder/lattice-faster-decoder.h
@@ -23,11 +23,10 @@
 #ifndef KALDI_DECODER_LATTICE_FASTER_DECODER_H_
 #define KALDI_DECODER_LATTICE_FASTER_DECODER_H_
 
-#include "decoder/grammar-fst.h"
 #include "fst/fstlib.h"
 #include "fst/memory.h"
 #include "fstext/fstext-lib.h"
-#include "itf/decodable-itf.h"
+#include "decoder/decodable-itf.h"
 #include "lat/determinize-lattice-pruned.h"
 #include "lat/kaldi-lattice.h"
 #include "util/hash-list.h"
diff --git a/speechx/speechx/kaldi/decoder/lattice-faster-online-decoder.cc b/speechx/speechx/kaldi/decoder/lattice-faster-online-decoder.cc
index ebdace7e..b5261503 100644
--- a/speechx/speechx/kaldi/decoder/lattice-faster-online-decoder.cc
+++ b/speechx/speechx/kaldi/decoder/lattice-faster-online-decoder.cc
@@ -278,8 +278,8 @@ bool LatticeFasterOnlineDecoderTpl<FST>::GetRawLatticePruned(
 template class LatticeFasterOnlineDecoderTpl<fst::Fst<fst::StdArc> >;
 template class LatticeFasterOnlineDecoderTpl<fst::VectorFst<fst::StdArc> >;
 template class LatticeFasterOnlineDecoderTpl<fst::ConstFst<fst::StdArc> >;
-template class LatticeFasterOnlineDecoderTpl<fst::ConstGrammarFst >;
-template class LatticeFasterOnlineDecoderTpl<fst::VectorGrammarFst >;
+//template class LatticeFasterOnlineDecoderTpl<fst::ConstGrammarFst >;
+//template class LatticeFasterOnlineDecoderTpl<fst::VectorGrammarFst >;
 
 
 } // end namespace kaldi.
diff --git a/speechx/speechx/kaldi/decoder/lattice-faster-online-decoder.h b/speechx/speechx/kaldi/decoder/lattice-faster-online-decoder.h
index 8b10996f..f57368a4 100644
--- a/speechx/speechx/kaldi/decoder/lattice-faster-online-decoder.h
+++ b/speechx/speechx/kaldi/decoder/lattice-faster-online-decoder.h
@@ -30,7 +30,7 @@
 #include "util/stl-utils.h"
 #include "util/hash-list.h"
 #include "fst/fstlib.h"
-#include "itf/decodable-itf.h"
+#include "decoder/decodable-itf.h"
 #include "fstext/fstext-lib.h"
 #include "lat/determinize-lattice-pruned.h"
 #include "lat/kaldi-lattice.h"
diff --git a/speechx/speechx/kaldi/fstext/CMakeLists.txt b/speechx/speechx/kaldi/fstext/CMakeLists.txt
new file mode 100644
index 00000000..af91fd98
--- /dev/null
+++ b/speechx/speechx/kaldi/fstext/CMakeLists.txt
@@ -0,0 +1,5 @@
+
+add_library(kaldi-fstext
+kaldi-fst-io.cc
+)
+target_link_libraries(kaldi-fstext PUBLIC kaldi-util)
diff --git a/speechx/speechx/kaldi/fstext/determinize-lattice-inl.h b/speechx/speechx/kaldi/fstext/determinize-lattice-inl.h
new file mode 100644
index 00000000..0bfbc8f4
--- /dev/null
+++ b/speechx/speechx/kaldi/fstext/determinize-lattice-inl.h
@@ -0,0 +1,1357 @@
+// fstext/determinize-lattice-inl.h
+
+// Copyright 2009-2012  Microsoft Corporation
+//           2012-2013  Johns Hopkins University (Author: Daniel Povey)
+
+// See ../../COPYING for clarification regarding multiple authors
+//
+// Licensed under the Apache License, Version 2.0 (the "License");
+// you may not use this file except in compliance with the License.
+// You may obtain a copy of the License at
+//
+//  http://www.apache.org/licenses/LICENSE-2.0
+//
+// THIS CODE IS PROVIDED *AS IS* BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY
+// KIND, EITHER EXPRESS OR IMPLIED, INCLUDING WITHOUT LIMITATION ANY IMPLIED
+// WARRANTIES OR CONDITIONS OF TITLE, FITNESS FOR A PARTICULAR PURPOSE,
+// MERCHANTABLITY OR NON-INFRINGEMENT.
+// See the Apache 2 License for the specific language governing permissions and
+// limitations under the License.
+
+#ifndef KALDI_FSTEXT_DETERMINIZE_LATTICE_INL_H_
+#define KALDI_FSTEXT_DETERMINIZE_LATTICE_INL_H_
+// Do not include this file directly.  It is included by determinize-lattice.h
+
+#include <algorithm>
+#include <climits>
+#include <deque>
+#include <string>
+#include <unordered_map>
+#include <unordered_set>
+#include <utility>
+#include <vector>
+
+namespace fst {
+
+// This class maps back and forth from/to integer id's to sequences of strings.
+// used in determinization algorithm.  It is constructed in such a way that
+// finding the string-id of the successor of (string, next-label) has constant
+// time.
+
+// Note: class IntType, typically int32, is the type of the element in the
+// string (typically a template argument of the CompactLatticeWeightTpl).
+
+template <class IntType>
+class LatticeStringRepository {
+ public:
+  struct Entry {
+    const Entry *parent;  // NULL for empty string.
+    IntType i;
+    inline bool operator==(const Entry &other) const {
+      return (parent == other.parent && i == other.i);
+    }
+    Entry() {}
+    Entry(const Entry &e) : parent(e.parent), i(e.i) {}
+  };
+  // Note: all Entry* pointers returned in function calls are
+  // owned by the repository itself, not by the caller!
+
+  // Interface guarantees empty string is NULL.
+  inline const Entry *EmptyString() { return NULL; }
+
+  // Returns string of "parent" with i appended.  Pointer
+  // owned by repository
+  const Entry *Successor(const Entry *parent, IntType i) {
+    new_entry_->parent = parent;
+    new_entry_->i = i;
+
+    std::pair<typename SetType::iterator, bool> pr = set_.insert(new_entry_);
+    if (pr.second) {  // Was successfully inserted (was not there).  We need to
+                      // replace the element we inserted, which resides on the
+                      // stack, with one from the heap.
+      const Entry *ans = new_entry_;
+      new_entry_ = new Entry();
+      return ans;
+    } else {  // Was not inserted because an equivalent Entry already
+              // existed.
+      return *pr.first;
+    }
+  }
+
+  const Entry *Concatenate(const Entry *a, const Entry *b) {
+    if (a == NULL)
+      return b;
+    else if (b == NULL)
+      return a;
+    std::vector<IntType> v;
+    ConvertToVector(b, &v);
+    const Entry *ans = a;
+    for (size_t i = 0; i < v.size(); i++) ans = Successor(ans, v[i]);
+    return ans;
+  }
+  const Entry *CommonPrefix(const Entry *a, const Entry *b) {
+    std::vector<IntType> a_vec, b_vec;
+    ConvertToVector(a, &a_vec);
+    ConvertToVector(b, &b_vec);
+    const Entry *ans = NULL;
+    for (size_t i = 0;
+         i < a_vec.size() && i < b_vec.size() && a_vec[i] == b_vec[i]; i++)
+      ans = Successor(ans, a_vec[i]);
+    return ans;
+  }
+
+  // removes any elements from b that are not part of
+  // a common prefix with a.
+  void ReduceToCommonPrefix(const Entry *a, std::vector<IntType> *b) {
+    size_t a_size = Size(a), b_size = b->size();
+    while (a_size > b_size) {
+      a = a->parent;
+      a_size--;
+    }
+    if (b_size > a_size) b_size = a_size;
+    typename std::vector<IntType>::iterator b_begin = b->begin();
+    while (a_size != 0) {
+      if (a->i != *(b_begin + a_size - 1)) b_size = a_size - 1;
+      a = a->parent;
+      a_size--;
+    }
+    if (b_size != b->size()) b->resize(b_size);
+  }
+
+  // removes the first n elements of a.
+  const Entry *RemovePrefix(const Entry *a, size_t n) {
+    if (n == 0) return a;
+    std::vector<IntType> a_vec;
+    ConvertToVector(a, &a_vec);
+    assert(a_vec.size() >= n);
+    const Entry *ans = NULL;
+    for (size_t i = n; i < a_vec.size(); i++) ans = Successor(ans, a_vec[i]);
+    return ans;
+  }
+
+  // Returns true if a is a prefix of b.  If a is prefix of b,
+  // time taken is |b| - |a|.  Else, time taken is |b|.
+  bool IsPrefixOf(const Entry *a, const Entry *b) const {
+    if (a == NULL) return true;  // empty string prefix of all.
+    if (a == b) return true;
+    if (b == NULL) return false;
+    return IsPrefixOf(a, b->parent);
+  }
+
+  inline size_t Size(const Entry *entry) const {
+    size_t ans = 0;
+    while (entry != NULL) {
+      ans++;
+      entry = entry->parent;
+    }
+    return ans;
+  }
+
+  void ConvertToVector(const Entry *entry, std::vector<IntType> *out) const {
+    size_t length = Size(entry);
+    out->resize(length);
+    if (entry != NULL) {
+      typename std::vector<IntType>::reverse_iterator iter = out->rbegin();
+      while (entry != NULL) {
+        *iter = entry->i;
+        entry = entry->parent;
+        ++iter;
+      }
+    }
+  }
+
+  const Entry *ConvertFromVector(const std::vector<IntType> &vec) {
+    const Entry *e = NULL;
+    for (size_t i = 0; i < vec.size(); i++) e = Successor(e, vec[i]);
+    return e;
+  }
+
+  LatticeStringRepository() { new_entry_ = new Entry; }
+
+  void Destroy() {
+    for (typename SetType::iterator iter = set_.begin(); iter != set_.end();
+         ++iter)
+      delete *iter;
+    SetType tmp;
+    tmp.swap(set_);
+    if (new_entry_) {
+      delete new_entry_;
+      new_entry_ = NULL;
+    }
+  }
+
+  // Rebuild will rebuild this object, guaranteeing only
+  // to preserve the Entry values that are in the vector pointed
+  // to (this list does not have to be unique).  The point of
+  // this is to save memory.
+  void Rebuild(const std::vector<const Entry *> &to_keep) {
+    SetType tmp_set;
+    for (typename std::vector<const Entry *>::const_iterator iter =
+             to_keep.begin();
+         iter != to_keep.end(); ++iter)
+      RebuildHelper(*iter, &tmp_set);
+    // Now delete all elems not in tmp_set.
+    for (typename SetType::iterator iter = set_.begin(); iter != set_.end();
+         ++iter) {
+      if (tmp_set.count(*iter) == 0)
+        delete (*iter);  // delete the Entry; not needed.
+    }
+    set_.swap(tmp_set);
+  }
+
+  ~LatticeStringRepository() { Destroy(); }
+  int32 MemSize() const {
+    return set_.size() * sizeof(Entry) * 2;  // this is a lower bound
+    // on the size this structure might take.
+  }
+
+ private:
+  class EntryKey {  // Hash function object.
+   public:
+    inline size_t operator()(const Entry *entry) const {
+      size_t prime = 49109;
+      return static_cast<size_t>(entry->i) +
+             prime * reinterpret_cast<size_t>(entry->parent);
+    }
+  };
+  class EntryEqual {
+   public:
+    inline bool operator()(const Entry *e1, const Entry *e2) const {
+      return (*e1 == *e2);
+    }
+  };
+  typedef std::unordered_set<const Entry *, EntryKey, EntryEqual> SetType;
+
+  void RebuildHelper(const Entry *to_add, SetType *tmp_set) {
+    while (true) {
+      if (to_add == NULL) return;
+      typename SetType::iterator iter = tmp_set->find(to_add);
+      if (iter == tmp_set->end()) {  // not in tmp_set.
+        tmp_set->insert(to_add);
+        to_add = to_add->parent;  // and loop.
+      } else {
+        return;
+      }
+    }
+  }
+
+  KALDI_DISALLOW_COPY_AND_ASSIGN(LatticeStringRepository);
+  Entry *new_entry_;  // We always have a pre-allocated Entry ready to use,
+                      // to avoid unnecessary news and deletes.
+  SetType set_;
+};
+
+// class LatticeDeterminizer is templated on the same types that
+// CompactLatticeWeight is templated on: the base weight (Weight), typically
+// LatticeWeightTpl<float> etc. but could also be e.g. TropicalWeight, and the
+// IntType, typically int32, used for the output symbols in the compact
+// representation of strings [note: the output symbols would usually be
+// p.d.f. id's in the anticipated use of this code] It has a special requirement
+// on the Weight type: that there should be a Compare function on the weights
+// such that Compare(w1, w2) returns -1 if w1 < w2, 0 if w1 == w2, and +1 if w1
+// > w2.  This requires that there be a total order on the weights.
+
+template <class Weight, class IntType>
+class LatticeDeterminizer {
+ public:
+  // Output to Gallic acceptor (so the strings go on weights, and there is a 1-1
+  // correspondence between our states and the states in ofst.  If destroy ==
+  // true, release memory as we go (but we cannot output again).
+
+  typedef CompactLatticeWeightTpl<Weight, IntType> CompactWeight;
+  typedef ArcTpl<CompactWeight>
+      CompactArc;              // arc in compact, acceptor form of lattice
+  typedef ArcTpl<Weight> Arc;  // arc in non-compact version of lattice
+
+  // Output to standard FST with CompactWeightTpl<Weight> as its weight type
+  // (the weight stores the original output-symbol strings).  If destroy ==
+  // true, release memory as we go (but we cannot output again).
+  void Output(MutableFst<CompactArc> *ofst, bool destroy = true) {
+    assert(determinized_);
+    typedef typename Arc::StateId StateId;
+    StateId nStates = static_cast<StateId>(output_arcs_.size());
+    if (destroy) FreeMostMemory();
+    ofst->DeleteStates();
+    ofst->SetStart(kNoStateId);
+    if (nStates == 0) {
+      return;
+    }
+    for (StateId s = 0; s < nStates; s++) {
+      OutputStateId news = ofst->AddState();
+      assert(news == s);
+    }
+    ofst->SetStart(0);
+    // now process transitions.
+    for (StateId this_state = 0; this_state < nStates; this_state++) {
+      std::vector<TempArc> &this_vec(output_arcs_[this_state]);
+      typename std::vector<TempArc>::const_iterator iter = this_vec.begin(),
+                                                    end = this_vec.end();
+
+      for (; iter != end; ++iter) {
+        const TempArc &temp_arc(*iter);
+        CompactArc new_arc;
+        std::vector<Label> seq;
+        repository_.ConvertToVector(temp_arc.string, &seq);
+        CompactWeight weight(temp_arc.weight, seq);
+        if (temp_arc.nextstate == kNoStateId) {  // is really final weight.
+          ofst->SetFinal(this_state, weight);
+        } else {  // is really an arc.
+          new_arc.nextstate = temp_arc.nextstate;
+          new_arc.ilabel = temp_arc.ilabel;
+          new_arc.olabel = temp_arc.ilabel;  // acceptor.  input == output.
+          new_arc.weight = weight;           // includes string and weight.
+          ofst->AddArc(this_state, new_arc);
+        }
+      }
+      // Free up memory.  Do this inside the loop as ofst is also allocating
+      // memory
+      if (destroy) {
+        std::vector<TempArc> temp;
+        std::swap(temp, this_vec);
+      }
+    }
+    if (destroy) {
+      std::vector<std::vector<TempArc> > temp;
+      std::swap(temp, output_arcs_);
+    }
+  }
+
+  // Output to standard FST with Weight as its weight type.  We will create
+  // extra states to handle sequences of symbols on the output.  If destroy ==
+  // true, release memory as we go (but we cannot output again).
+  void Output(MutableFst<Arc> *ofst, bool destroy = true) {
+    // Outputs to standard fst.
+    OutputStateId nStates = static_cast<OutputStateId>(output_arcs_.size());
+    ofst->DeleteStates();
+    if (nStates == 0) {
+      ofst->SetStart(kNoStateId);
+      return;
+    }
+    if (destroy) FreeMostMemory();
+    // Add basic states-- but we will add extra ones to account for strings on
+    // output.
+    for (OutputStateId s = 0; s < nStates; s++) {
+      OutputStateId news = ofst->AddState();
+      assert(news == s);
+    }
+    ofst->SetStart(0);
+    for (OutputStateId this_state = 0; this_state < nStates; this_state++) {
+      std::vector<TempArc> &this_vec(output_arcs_[this_state]);
+
+      typename std::vector<TempArc>::const_iterator iter = this_vec.begin(),
+                                                    end = this_vec.end();
+      for (; iter != end; ++iter) {
+        const TempArc &temp_arc(*iter);
+        std::vector<Label> seq;
+        repository_.ConvertToVector(temp_arc.string, &seq);
+
+        if (temp_arc.nextstate == kNoStateId) {  // Really a final weight.
+          // Make a sequence of states going to a final state, with the strings
+          // as labels.  Put the weight on the first arc.
+          OutputStateId cur_state = this_state;
+          for (size_t i = 0; i < seq.size(); i++) {
+            OutputStateId next_state = ofst->AddState();
+            Arc arc;
+            arc.nextstate = next_state;
+            arc.weight = (i == 0 ? temp_arc.weight : Weight::One());
+            arc.ilabel = 0;  // epsilon.
+            arc.olabel = seq[i];
+            ofst->AddArc(cur_state, arc);
+            cur_state = next_state;
+          }
+          ofst->SetFinal(cur_state,
+                         (seq.size() == 0 ? temp_arc.weight : Weight::One()));
+        } else {  // Really an arc.
+          OutputStateId cur_state = this_state;
+          // Have to be careful with this integer comparison (i+1 < seq.size())
+          // because unsigned. i < seq.size()-1 could fail for zero-length
+          // sequences.
+          for (size_t i = 0; i + 1 < seq.size(); i++) {
+            // for all but the last element of seq, create new state.
+            OutputStateId next_state = ofst->AddState();
+            Arc arc;
+            arc.nextstate = next_state;
+            arc.weight = (i == 0 ? temp_arc.weight : Weight::One());
+            arc.ilabel = (i == 0 ? temp_arc.ilabel
+                                 : 0);  // put ilabel on first element of seq.
+            arc.olabel = seq[i];
+            ofst->AddArc(cur_state, arc);
+            cur_state = next_state;
+          }
+          // Add the final arc in the sequence.
+          Arc arc;
+          arc.nextstate = temp_arc.nextstate;
+          arc.weight = (seq.size() <= 1 ? temp_arc.weight : Weight::One());
+          arc.ilabel = (seq.size() <= 1 ? temp_arc.ilabel : 0);
+          arc.olabel = (seq.size() > 0 ? seq.back() : 0);
+          ofst->AddArc(cur_state, arc);
+        }
+      }
+      // Free up memory.  Do this inside the loop as ofst is also allocating
+      // memory
+      if (destroy) {
+        std::vector<TempArc> temp;
+        temp.swap(this_vec);
+      }
+    }
+    if (destroy) {
+      std::vector<std::vector<TempArc> > temp;
+      temp.swap(output_arcs_);
+      repository_.Destroy();
+    }
+  }
+
+  // Initializer.  After initializing the object you will typically
+  // call Determinize() and then call one of the Output functions.
+  // Note: ifst.Copy() will generally do a
+  // shallow copy.  We do it like this for memory safety, rather than
+  // keeping a reference or pointer to ifst_.
+  LatticeDeterminizer(const Fst<Arc> &ifst, DeterminizeLatticeOptions opts)
+      : num_arcs_(0),
+        num_elems_(0),
+        ifst_(ifst.Copy()),
+        opts_(opts),
+        equal_(opts_.delta),
+        determinized_(false),
+        minimal_hash_(3, hasher_, equal_),
+        initial_hash_(3, hasher_, equal_) {
+    KALDI_ASSERT(Weight::Properties() & kIdempotent);  // this algorithm won't
+    // work correctly otherwise.
+  }
+
+  // frees all except output_arcs_, which contains the important info
+  // we need to output the FST.
+  void FreeMostMemory() {
+    if (ifst_) {
+      delete ifst_;
+      ifst_ = NULL;
+    }
+    for (typename MinimalSubsetHash::iterator iter = minimal_hash_.begin();
+         iter != minimal_hash_.end(); ++iter)
+      delete iter->first;
+    {
+      MinimalSubsetHash tmp;
+      tmp.swap(minimal_hash_);
+    }
+    for (typename InitialSubsetHash::iterator iter = initial_hash_.begin();
+         iter != initial_hash_.end(); ++iter)
+      delete iter->first;
+    {
+      InitialSubsetHash tmp;
+      tmp.swap(initial_hash_);
+    }
+    {
+      std::vector<std::vector<Element> *> output_states_tmp;
+      output_states_tmp.swap(output_states_);
+    }
+    {
+      std::vector<char> tmp;
+      tmp.swap(isymbol_or_final_);
+    }
+    {
+      std::vector<OutputStateId> tmp;
+      tmp.swap(queue_);
+    }
+    {
+      std::vector<std::pair<Label, Element> > tmp;
+      tmp.swap(all_elems_tmp_);
+    }
+  }
+
+  ~LatticeDeterminizer() {
+    FreeMostMemory();  // rest is deleted by destructors.
+  }
+  void RebuildRepository() {  // rebuild the string repository,
+    // freeing stuff we don't need.. we call this when memory usage
+    // passes a supplied threshold.  We need to accumulate all the
+    // strings we need the repository to "remember", then tell it
+    // to clean the repository.
+    std::vector<StringId> needed_strings;
+    for (size_t i = 0; i < output_arcs_.size(); i++)
+      for (size_t j = 0; j < output_arcs_[i].size(); j++)
+        needed_strings.push_back(output_arcs_[i][j].string);
+
+    // the following loop covers strings present in minimal_hash_
+    // which are also accessible via output_states_.
+    for (size_t i = 0; i < output_states_.size(); i++)
+      for (size_t j = 0; j < output_states_[i]->size(); j++)
+        needed_strings.push_back((*(output_states_[i]))[j].string);
+
+    // the following loop covers strings present in initial_hash_.
+    for (typename InitialSubsetHash::const_iterator iter =
+             initial_hash_.begin();
+         iter != initial_hash_.end(); ++iter) {
+      const std::vector<Element> &vec = *(iter->first);
+      Element elem = iter->second;
+      for (size_t i = 0; i < vec.size(); i++)
+        needed_strings.push_back(vec[i].string);
+      needed_strings.push_back(elem.string);
+    }
+
+    std::sort(needed_strings.begin(), needed_strings.end());
+    needed_strings.erase(
+        std::unique(needed_strings.begin(), needed_strings.end()),
+        needed_strings.end());  // uniq the strings.
+    repository_.Rebuild(needed_strings);
+  }
+
+  bool CheckMemoryUsage() {
+    int32 repo_size = repository_.MemSize(),
+          arcs_size = num_arcs_ * sizeof(TempArc),
+          elems_size = num_elems_ * sizeof(Element),
+          total_size = repo_size + arcs_size + elems_size;
+    if (opts_.max_mem > 0 &&
+        total_size > opts_.max_mem) {  // We passed the memory threshold.
+      // This is usually due to the repository getting large, so we
+      // clean this out.
+      RebuildRepository();
+      int32 new_repo_size = repository_.MemSize(),
+            new_total_size = new_repo_size + arcs_size + elems_size;
+
+      KALDI_VLOG(2) << "Rebuilt repository in determinize-lattice: repository "
+                       "shrank from "
+                    << repo_size << " to " << new_repo_size
+                    << " bytes (approximately)";
+
+      if (new_total_size > static_cast<int32>(opts_.max_mem * 0.8)) {
+        // Rebuilding didn't help enough-- we need a margin to stop
+        // having to rebuild too often.
+        KALDI_WARN << "Failure in determinize-lattice: size exceeds maximum "
+                   << opts_.max_mem << " bytes; (repo,arcs,elems) = ("
+                   << repo_size << "," << arcs_size << "," << elems_size
+                   << "), after rebuilding, repo size was " << new_repo_size;
+        return false;
+      }
+    }
+    return true;
+  }
+
+  // Returns true on success.  Can fail for out-of-memory
+  // or max-states related reasons.
+  bool Determinize(bool *debug_ptr) {
+    assert(!determinized_);
+    // This determinizes the input fst but leaves it in the "special format"
+    // in "output_arcs_".  Must be called after Initialize().  To get the
+    // output, call one of the Output routines.
+    try {
+      InitializeDeterminization();  // some start-up tasks.
+      while (!queue_.empty()) {
+        OutputStateId out_state = queue_.back();
+        queue_.pop_back();
+        ProcessState(out_state);
+        if (debug_ptr && *debug_ptr) Debug();  // will exit.
+        if (!CheckMemoryUsage()) return false;
+      }
+      return (determinized_ = true);
+    } catch (const std::bad_alloc &) {
+      int32 repo_size = repository_.MemSize(),
+            arcs_size = num_arcs_ * sizeof(TempArc),
+            elems_size = num_elems_ * sizeof(Element),
+            total_size = repo_size + arcs_size + elems_size;
+      KALDI_WARN
+          << "Memory allocation error doing lattice determinization; using "
+          << total_size << " bytes (max = " << opts_.max_mem
+          << " (repo,arcs,elems) = (" << repo_size << "," << arcs_size << ","
+          << elems_size << ")";
+      return (determinized_ = false);
+    } catch (const std::runtime_error &) {
+      KALDI_WARN << "Caught exception doing lattice determinization";
+      return (determinized_ = false);
+    }
+  }
+
+ private:
+  typedef typename Arc::Label Label;
+  typedef typename Arc::StateId
+      StateId;  // use this when we don't know if it's input or output.
+  typedef typename Arc::StateId InputStateId;   // state in the input FST.
+  typedef typename Arc::StateId OutputStateId;  // same as above but distinguish
+                                                // states in output Fst.
+
+  typedef LatticeStringRepository<IntType> StringRepositoryType;
+  typedef const typename StringRepositoryType::Entry *StringId;
+
+  // Element of a subset [of original states]
+  struct Element {
+    StateId state;  // use StateId as this is usually InputStateId but in one
+                    // case OutputStateId.
+    StringId string;
+    Weight weight;
+    bool operator!=(const Element &other) const {
+      return (state != other.state || string != other.string ||
+              weight != other.weight);
+    }
+    // This operator is only intended to support sorting in EpsilonClosure()
+    bool operator<(const Element &other) const { return state < other.state; }
+  };
+
+  // Arcs in the format we temporarily create in this class (a representation,
+  // essentially of a Gallic Fst).
+  struct TempArc {
+    Label ilabel;
+    StringId string;  // Look it up in the StringRepository, it's a sequence of
+                      // Labels.
+    OutputStateId nextstate;  // or kNoState for final weights.
+    Weight weight;
+  };
+
+  // Hashing function used in hash of subsets.
+  // A subset is a pointer to vector<Element>.
+  // The Elements are in sorted order on state id, and without repeated states.
+  // Because the order of Elements is fixed, we can use a hashing function that
+  // is order-dependent.  However the weights are not included in the hashing
+  // function-- we hash subsets that differ only in weight to the same key. This
+  // is not optimal in terms of the O(N) performance but typically if we have a
+  // lot of determinized states that differ only in weight then the input
+  // probably was pathological in some way, or even non-determinizable.
+  //   We don't quantize the weights, in order to avoid inexactness in simple
+  //   cases.
+  // Instead we apply the delta when comparing subsets for equality, and allow a
+  // small difference.
+
+  class SubsetKey {
+   public:
+    size_t operator()(const std::vector<Element> *subset)
+        const {  // hashes only the state and string.
+      size_t hash = 0, factor = 1;
+      for (typename std::vector<Element>::const_iterator iter = subset->begin();
+           iter != subset->end(); ++iter) {
+        hash *= factor;
+        hash += iter->state + reinterpret_cast<size_t>(iter->string);
+        factor *= 23531;  // these numbers are primes.
+      }
+      return hash;
+    }
+  };
+
+  // This is the equality operator on subsets.  It checks for exact match on
+  // state-id and string, and approximate match on weights.
+  class SubsetEqual {
+   public:
+    bool operator()(const std::vector<Element> *s1,
+                    const std::vector<Element> *s2) const {
+      size_t sz = s1->size();
+      assert(sz >= 0);
+      if (sz != s2->size()) return false;
+      typename std::vector<Element>::const_iterator iter1 = s1->begin(),
+                                                    iter1_end = s1->end(),
+                                                    iter2 = s2->begin();
+      for (; iter1 < iter1_end; ++iter1, ++iter2) {
+        if (iter1->state != iter2->state || iter1->string != iter2->string ||
+            !ApproxEqual(iter1->weight, iter2->weight, delta_))
+          return false;
+      }
+      return true;
+    }
+    float delta_;
+    explicit SubsetEqual(float delta) : delta_(delta) {}
+    SubsetEqual() : delta_(kDelta) {}
+  };
+
+  // Operator that says whether two Elements have the same states.
+  // Used only for debug.
+  class SubsetEqualStates {
+   public:
+    bool operator()(const std::vector<Element> *s1,
+                    const std::vector<Element> *s2) const {
+      size_t sz = s1->size();
+      assert(sz >= 0);
+      if (sz != s2->size()) return false;
+      typename std::vector<Element>::const_iterator iter1 = s1->begin(),
+                                                    iter1_end = s1->end(),
+                                                    iter2 = s2->begin();
+      for (; iter1 < iter1_end; ++iter1, ++iter2) {
+        if (iter1->state != iter2->state) return false;
+      }
+      return true;
+    }
+  };
+
+  // Define the hash type we use to map subsets (in minimal
+  // representation) to OutputStateId.
+  typedef std::unordered_map<const std::vector<Element> *, OutputStateId,
+                             SubsetKey, SubsetEqual>
+      MinimalSubsetHash;
+
+  // Define the hash type we use to map subsets (in initial
+  // representation) to OutputStateId, together with an
+  // extra weight. [note: we interpret the Element.state in here
+  // as an OutputStateId even though it's declared as InputStateId;
+  // these types are the same anyway].
+  typedef std::unordered_map<const std::vector<Element> *, Element, SubsetKey,
+                             SubsetEqual>
+      InitialSubsetHash;
+
+  // converts the representation of the subset from canonical (all states) to
+  // minimal (only states with output symbols on arcs leaving them, and final
+  // states).  Output is not necessarily normalized, even if input_subset was.
+  void ConvertToMinimal(std::vector<Element> *subset) {
+    assert(!subset->empty());
+    typename std::vector<Element>::iterator cur_in = subset->begin(),
+                                            cur_out = subset->begin(),
+                                            end = subset->end();
+    while (cur_in != end) {
+      if (IsIsymbolOrFinal(cur_in->state)) {  // keep it...
+        *cur_out = *cur_in;
+        cur_out++;
+      }
+      cur_in++;
+    }
+    subset->resize(cur_out - subset->begin());
+  }
+
+  // Takes a minimal, normalized subset, and converts it to an OutputStateId.
+  // Involves a hash lookup, and possibly adding a new OutputStateId.
+  // If it creates a new OutputStateId, it adds it to the queue.
+  OutputStateId MinimalToStateId(const std::vector<Element> &subset) {
+    typename MinimalSubsetHash::const_iterator iter =
+        minimal_hash_.find(&subset);
+    if (iter != minimal_hash_.end())  // Found a matching subset.
+      return iter->second;
+    OutputStateId ans = static_cast<OutputStateId>(output_arcs_.size());
+    std::vector<Element> *subset_ptr = new std::vector<Element>(subset);
+    output_states_.push_back(subset_ptr);
+    num_elems_ += subset_ptr->size();
+    output_arcs_.push_back(std::vector<TempArc>());
+    minimal_hash_[subset_ptr] = ans;
+    queue_.push_back(ans);
+    return ans;
+  }
+
+  // Given a normalized initial subset of elements (i.e. before epsilon
+  // closure), compute the corresponding output-state.
+  OutputStateId InitialToStateId(const std::vector<Element> &subset_in,
+                                 Weight *remaining_weight,
+                                 StringId *common_prefix) {
+    typename InitialSubsetHash::const_iterator iter =
+        initial_hash_.find(&subset_in);
+    if (iter != initial_hash_.end()) {  // Found a matching subset.
+      const Element &elem = iter->second;
+      *remaining_weight = elem.weight;
+      *common_prefix = elem.string;
+      if (elem.weight == Weight::Zero()) KALDI_WARN << "Zero weight!";  // TEMP
+      return elem.state;
+    }
+    // else no matching subset-- have to work it out.
+    std::vector<Element> subset(subset_in);
+    // Follow through epsilons.  Will add no duplicate states.  note: after
+    // EpsilonClosure, it is the same as "canonical" subset, except not
+    // normalized (actually we never compute the normalized canonical subset,
+    // only the normalized minimal one).
+    EpsilonClosure(&subset);    // follow epsilons.
+    ConvertToMinimal(&subset);  // remove all but emitting and final states.
+
+    Element elem;  // will be used to store remaining weight and string, and
+                   // OutputStateId, in initial_hash_;
+    NormalizeSubset(&subset, &elem.weight,
+                    &elem.string);  // normalize subset; put
+    // common string and weight in "elem".  The subset is now a minimal,
+    // normalized subset.
+
+    OutputStateId ans = MinimalToStateId(subset);
+    *remaining_weight = elem.weight;
+    *common_prefix = elem.string;
+    if (elem.weight == Weight::Zero()) KALDI_WARN << "Zero weight!";  // TEMP
+
+    // Before returning "ans", add the initial subset to the hash,
+    // so that we can bypass the epsilon-closure etc., next time
+    // we process the same initial subset.
+    std::vector<Element> *initial_subset_ptr =
+        new std::vector<Element>(subset_in);
+    elem.state = ans;
+    initial_hash_[initial_subset_ptr] = elem;
+    num_elems_ += initial_subset_ptr->size();  // keep track of memory usage.
+    return ans;
+  }
+
+  // returns the Compare value (-1 if a < b, 0 if a == b, 1 if a > b) according
+  // to the ordering we defined on strings for the CompactLatticeWeightTpl.
+  // see function
+  // inline int Compare (const CompactLatticeWeightTpl<WeightType,IntType> &w1,
+  //                     const CompactLatticeWeightTpl<WeightType,IntType> &w2)
+  // in lattice-weight.h.
+  // this is the same as that, but optimized for our data structures.
+  inline int Compare(const Weight &a_w, StringId a_str, const Weight &b_w,
+                     StringId b_str) const {
+    int weight_comp = fst::Compare(a_w, b_w);
+    if (weight_comp != 0) return weight_comp;
+    // now comparing strings.
+    if (a_str == b_str) return 0;
+    std::vector<IntType> a_vec, b_vec;
+    repository_.ConvertToVector(a_str, &a_vec);
+    repository_.ConvertToVector(b_str, &b_vec);
+    // First compare their lengths.
+    int a_len = a_vec.size(), b_len = b_vec.size();
+    // use opposite order on the string lengths (c.f. Compare in
+    // lattice-weight.h)
+    if (a_len > b_len)
+      return -1;
+    else if (a_len < b_len)
+      return 1;
+    for (int i = 0; i < a_len; i++) {
+      if (a_vec[i] < b_vec[i])
+        return -1;
+      else if (a_vec[i] > b_vec[i])
+        return 1;
+    }
+    assert(
+        0);  // because we checked if a_str == b_str above, shouldn't reach here
+    return 0;
+  }
+
+  // This function computes epsilon closure of subset of states by following
+  // epsilon links. Called by InitialToStateId and Initialize. Has no side
+  // effects except on the string repository.  The "output_subset" is not
+  // necessarily normalized (in the sense of there being no common substring),
+  // unless input_subset was.
+  void EpsilonClosure(std::vector<Element> *subset) {
+    // at input, subset must have only one example of each StateId.  [will still
+    // be so at output].  This function follows input-epsilons, and augments the
+    // subset accordingly.
+
+    std::deque<Element> queue;
+    std::unordered_map<InputStateId, Element> cur_subset;
+    typedef
+        typename std::unordered_map<InputStateId, Element>::iterator MapIter;
+    typedef typename std::vector<Element>::const_iterator VecIter;
+
+    for (VecIter iter = subset->begin(); iter != subset->end(); ++iter) {
+      queue.push_back(*iter);
+      cur_subset[iter->state] = *iter;
+    }
+
+    // find whether input fst is known to be sorted on input label.
+    bool sorted =
+        ((ifst_->Properties(kILabelSorted, false) & kILabelSorted) != 0);
+    bool replaced_elems = false;  // relates to an optimization, see below.
+    int counter =
+        0;  // stops infinite loops here for non-lattice-determinizable input;
+    // useful in testing.
+    while (queue.size() != 0) {
+      Element elem = queue.front();
+      queue.pop_front();
+
+      // The next if-statement is a kind of optimization.  It's to prevent us
+      // unnecessarily repeating the processing of a state.  "cur_subset" always
+      // contains only one Element with a particular state.  The issue is that
+      // whenever we modify the Element corresponding to that state in
+      // "cur_subset", both the new (optimal) and old (less-optimal) Element
+      // will still be in "queue".  The next if-statement stops us from wasting
+      // compute by processing the old Element.
+      if (replaced_elems && cur_subset[elem.state] != elem) continue;
+      if (opts_.max_loop > 0 && counter++ > opts_.max_loop) {
+        KALDI_ERR << "Lattice determinization aborted since looped more than "
+                  << opts_.max_loop << " times during epsilon closure";
+      }
+      for (ArcIterator<Fst<Arc> > aiter(*ifst_, elem.state); !aiter.Done();
+           aiter.Next()) {
+        const Arc &arc = aiter.Value();
+        if (sorted && arc.ilabel != 0)
+          break;  // Break from the loop: due to sorting there will be no
+        // more transitions with epsilons as input labels.
+        if (arc.ilabel == 0 &&
+            arc.weight != Weight::Zero()) {  // Epsilon transition.
+          Element next_elem;
+          next_elem.state = arc.nextstate;
+          next_elem.weight = Times(elem.weight, arc.weight);
+          // now must append strings
+          if (arc.olabel == 0)
+            next_elem.string = elem.string;
+          else
+            next_elem.string = repository_.Successor(elem.string, arc.olabel);
+
+          MapIter iter = cur_subset.find(next_elem.state);
+          if (iter == cur_subset.end()) {
+            // was no such StateId: insert and add to queue.
+            cur_subset[next_elem.state] = next_elem;
+            queue.push_back(next_elem);
+          } else {
+            // was not inserted because one already there.  In normal
+            // determinization we'd add the weights.  Here, we find which one
+            // has the better weight, and keep its corresponding string.
+            int comp = Compare(next_elem.weight, next_elem.string,
+                               iter->second.weight, iter->second.string);
+            if (comp ==
+                1) {  // next_elem is better, so use its (weight, string)
+              iter->second.string = next_elem.string;
+              iter->second.weight = next_elem.weight;
+              queue.push_back(next_elem);
+              replaced_elems = true;
+            }
+            // else it is the same or worse, so use original one.
+          }
+        }
+      }
+    }
+
+    {  // copy cur_subset to subset.
+      subset->clear();
+      subset->reserve(cur_subset.size());
+      MapIter iter = cur_subset.begin(), end = cur_subset.end();
+      for (; iter != end; ++iter) subset->push_back(iter->second);
+      // sort by state ID, because the subset hash function is
+      // order-dependent(see SubsetKey)
+      std::sort(subset->begin(), subset->end());
+    }
+  }
+
+  // This function works out the final-weight of the determinized state.
+  // called by ProcessSubset.
+  // Has no side effects except on the variable repository_, and output_arcs_.
+
+  void ProcessFinal(OutputStateId output_state) {
+    const std::vector<Element> &minimal_subset =
+        *(output_states_[output_state]);
+    // processes final-weights for this subset.
+
+    // minimal_subset may be empty if the graphs is not connected/trimmed, I
+    // think, do don't check that it's nonempty.
+    bool is_final = false;
+    StringId final_string = NULL;  // = NULL to keep compiler happy.
+    Weight final_weight = Weight::Zero();
+    typename std::vector<Element>::const_iterator iter = minimal_subset.begin(),
+                                                  end = minimal_subset.end();
+    for (; iter != end; ++iter) {
+      const Element &elem = *iter;
+      Weight this_final_weight = Times(elem.weight, ifst_->Final(elem.state));
+      StringId this_final_string = elem.string;
+      if (this_final_weight != Weight::Zero() &&
+          (!is_final || Compare(this_final_weight, this_final_string,
+                                final_weight, final_string) == 1)) {  // the new
+        // (weight, string) pair is more in semiring than our current
+        // one.
+        is_final = true;
+        final_weight = this_final_weight;
+        final_string = this_final_string;
+      }
+    }
+    if (is_final) {
+      // store final weights in TempArc structure, just like a transition.
+      TempArc temp_arc;
+      temp_arc.ilabel = 0;
+      temp_arc.nextstate =
+          kNoStateId;  // special marker meaning "final weight".
+      temp_arc.string = final_string;
+      temp_arc.weight = final_weight;
+      output_arcs_[output_state].push_back(temp_arc);
+      num_arcs_++;
+    }
+  }
+
+  // NormalizeSubset normalizes the subset "elems" by
+  // removing any common string prefix (putting it in common_str),
+  // and dividing by the total weight (putting it in tot_weight).
+  void NormalizeSubset(std::vector<Element> *elems, Weight *tot_weight,
+                       StringId *common_str) {
+    if (elems->empty()) {              // just set common_str, tot_weight
+      KALDI_WARN << "[empty subset]";  // TEMP
+      // to defaults and return...
+      *common_str = repository_.EmptyString();
+      *tot_weight = Weight::Zero();
+      return;
+    }
+    size_t size = elems->size();
+    std::vector<IntType> common_prefix;
+    repository_.ConvertToVector((*elems)[0].string, &common_prefix);
+    Weight weight = (*elems)[0].weight;
+    for (size_t i = 1; i < size; i++) {
+      weight = Plus(weight, (*elems)[i].weight);
+      repository_.ReduceToCommonPrefix((*elems)[i].string, &common_prefix);
+    }
+    assert(weight != Weight::Zero());  // we made sure to ignore arcs with zero
+    // weights on them, so we shouldn't have zero here.
+    size_t prefix_len = common_prefix.size();
+    for (size_t i = 0; i < size; i++) {
+      (*elems)[i].weight = Divide((*elems)[i].weight, weight, DIVIDE_LEFT);
+      (*elems)[i].string =
+          repository_.RemovePrefix((*elems)[i].string, prefix_len);
+    }
+    *common_str = repository_.ConvertFromVector(common_prefix);
+    *tot_weight = weight;
+  }
+
+  // Take a subset of Elements that is sorted on state, and
+  // merge any Elements that have the same state (taking the best
+  // (weight, string) pair in the semiring).
+  void MakeSubsetUnique(std::vector<Element> *subset) {
+    typedef typename std::vector<Element>::iterator IterType;
+
+    // This assert is designed to fail (usually) if the subset is not sorted on
+    // state.
+    assert(subset->size() < 2 || (*subset)[0].state <= (*subset)[1].state);
+
+    IterType cur_in = subset->begin(), cur_out = cur_in, end = subset->end();
+    size_t num_out = 0;
+    // Merge elements with same state-id
+    while (cur_in != end) {  // while we have more elements to process.
+      // At this point, cur_out points to location of next place we want to put
+      // an element, cur_in points to location of next element we want to
+      // process.
+      if (cur_in != cur_out) *cur_out = *cur_in;
+      cur_in++;
+      while (cur_in != end && cur_in->state == cur_out->state) {
+        if (Compare(cur_in->weight, cur_in->string, cur_out->weight,
+                    cur_out->string) == 1) {
+          // if *cur_in > *cur_out in semiring, then take *cur_in.
+          cur_out->string = cur_in->string;
+          cur_out->weight = cur_in->weight;
+        }
+        cur_in++;
+      }
+      cur_out++;
+      num_out++;
+    }
+    subset->resize(num_out);
+  }
+
+  // ProcessTransition is called from "ProcessTransitions".  Broken out for
+  // clarity.  Processes a transition from state "state".  The set of Elements
+  // represents a set of next-states with associated weights and strings, each
+  // one arising from an arc from some state in a determinized-state; the
+  // next-states are not necessarily unique (i.e. there may be >1 entry
+  // associated with each), and any such sets of Elements have to be merged
+  // within this routine (we take the [weight, string] pair that's better in the
+  // semiring).
+  void ProcessTransition(OutputStateId state, Label ilabel,
+                         std::vector<Element> *subset) {
+    MakeSubsetUnique(subset);  // remove duplicates with the same state.
+
+    StringId common_str;
+    Weight tot_weight;
+    NormalizeSubset(subset, &tot_weight, &common_str);
+
+    OutputStateId nextstate;
+    {
+      Weight next_tot_weight;
+      StringId next_common_str;
+      nextstate = InitialToStateId(*subset, &next_tot_weight, &next_common_str);
+      common_str = repository_.Concatenate(common_str, next_common_str);
+      tot_weight = Times(tot_weight, next_tot_weight);
+    }
+
+    // Now add an arc to the next state (would have been created if necessary by
+    // InitialToStateId).
+    TempArc temp_arc;
+    temp_arc.ilabel = ilabel;
+    temp_arc.nextstate = nextstate;
+    temp_arc.string = common_str;
+    temp_arc.weight = tot_weight;
+    output_arcs_[state].push_back(temp_arc);  // record the arc.
+    num_arcs_++;
+  }
+
+  // "less than" operator for pair<Label, Element>.   Used in
+  // ProcessTransitions. Lexicographical order, which only compares the state
+  // when ordering the "Element" member of the pair.
+
+  class PairComparator {
+   public:
+    inline bool operator()(const std::pair<Label, Element> &p1,
+                           const std::pair<Label, Element> &p2) {
+      if (p1.first < p2.first) {
+        return true;
+      } else if (p1.first > p2.first) {
+        return false;
+      } else {
+        return p1.second.state < p2.second.state;
+      }
+    }
+  };
+
+  // ProcessTransitions processes emitting transitions (transitions
+  // with ilabels) out of this subset of states.
+  // Does not consider final states.  Breaks the emitting transitions up by
+  // ilabel, and creates a new transition in the determinized FST for each
+  // unique ilabel. Does this by creating a big vector of pairs <Label, Element>
+  // and then sorting them using a lexicographical ordering, and calling
+  // ProcessTransition for each range with the same ilabel. Side effects on
+  // repository, and (via ProcessTransition) on Q_, hash_, and output_arcs_.
+
+  void ProcessTransitions(OutputStateId output_state) {
+    const std::vector<Element> &minimal_subset =
+        *(output_states_[output_state]);
+    // it's possible that minimal_subset could be empty if there are
+    // unreachable parts of the graph, so don't check that it's nonempty.
+    std::vector<std::pair<Label, Element> > &all_elems(
+        all_elems_tmp_);  // use class member
+    // to avoid memory allocation/deallocation.
+    {
+      // Push back into "all_elems", elements corresponding to all
+      // non-epsilon-input transitions out of all states in "minimal_subset".
+      typename std::vector<Element>::const_iterator iter =
+                                                        minimal_subset.begin(),
+                                                    end = minimal_subset.end();
+      for (; iter != end; ++iter) {
+        const Element &elem = *iter;
+        for (ArcIterator<Fst<Arc> > aiter(*ifst_, elem.state); !aiter.Done();
+             aiter.Next()) {
+          const Arc &arc = aiter.Value();
+          if (arc.ilabel != 0 &&
+              arc.weight != Weight::Zero()) {  // Non-epsilon transition --
+                                               // ignore epsilons here.
+            std::pair<Label, Element> this_pr;
+            this_pr.first = arc.ilabel;
+            Element &next_elem(this_pr.second);
+            next_elem.state = arc.nextstate;
+            next_elem.weight = Times(elem.weight, arc.weight);
+            if (arc.olabel == 0)  // output epsilon
+              next_elem.string = elem.string;
+            else
+              next_elem.string = repository_.Successor(elem.string, arc.olabel);
+            all_elems.push_back(this_pr);
+          }
+        }
+      }
+    }
+    PairComparator pc;
+    std::sort(all_elems.begin(), all_elems.end(), pc);
+    // now sorted first on input label, then on state.
+    typedef typename std::vector<std::pair<Label, Element> >::const_iterator
+        PairIter;
+    PairIter cur = all_elems.begin(), end = all_elems.end();
+    std::vector<Element> this_subset;
+    while (cur != end) {
+      // Process ranges that share the same input symbol.
+      Label ilabel = cur->first;
+      this_subset.clear();
+      while (cur != end && cur->first == ilabel) {
+        this_subset.push_back(cur->second);
+        cur++;
+      }
+      // We now have a subset for this ilabel.
+      assert(!this_subset.empty());  // temp.
+      ProcessTransition(output_state, ilabel, &this_subset);
+    }
+    all_elems.clear();  // as it's a class variable-- want it to stay
+    // emtpy.
+  }
+
+  // ProcessState does the processing of a determinized state, i.e. it creates
+  // transitions out of it and the final-probability if any.
+  void ProcessState(OutputStateId output_state) {
+    ProcessFinal(output_state);
+    ProcessTransitions(output_state);
+  }
+
+  void Debug() {  // this function called if you send a signal
+    // SIGUSR1 to the process (and it's caught by the handler in
+    // fstdeterminizestar).  It prints out some traceback
+    // info and exits.
+
+    KALDI_WARN << "Debug function called (probably SIGUSR1 caught)";
+    // free up memory from the hash as we need a little memory
+    {
+      MinimalSubsetHash hash_tmp;
+      hash_tmp.swap(minimal_hash_);
+    }
+
+    if (output_arcs_.size() <= 2) {
+      KALDI_ERR << "Nothing to trace back";
+    }
+    size_t max_state = output_arcs_.size() - 2;  // Don't take the last
+    // one as we might be halfway into constructing it.
+
+    std::vector<OutputStateId> predecessor(max_state + 1, kNoStateId);
+    for (size_t i = 0; i < max_state; i++) {
+      for (size_t j = 0; j < output_arcs_[i].size(); j++) {
+        OutputStateId nextstate = output_arcs_[i][j].nextstate;
+        // Always find an earlier-numbered predecessor; this
+        // is always possible because of the way the algorithm
+        // works.
+        if (nextstate <= max_state && nextstate > i) predecessor[nextstate] = i;
+      }
+    }
+    std::vector<std::pair<Label, StringId> > traceback;
+    // 'traceback' is a pair of (ilabel, olabel-seq).
+    OutputStateId cur_state = max_state;  // A recently constructed state.
+
+    while (cur_state != 0 && cur_state != kNoStateId) {
+      OutputStateId last_state = predecessor[cur_state];
+      std::pair<Label, StringId> p;
+      size_t i;
+      for (i = 0; i < output_arcs_[last_state].size(); i++) {
+        if (output_arcs_[last_state][i].nextstate == cur_state) {
+          p.first = output_arcs_[last_state][i].ilabel;
+          p.second = output_arcs_[last_state][i].string;
+          traceback.push_back(p);
+          break;
+        }
+      }
+      KALDI_ASSERT(i != output_arcs_[last_state].size());  // Or fell off loop.
+      cur_state = last_state;
+    }
+    if (cur_state == kNoStateId)
+      KALDI_WARN << "Traceback did not reach start state "
+                 << "(possibly debug-code error)";
+
+    std::stringstream ss;
+    ss << "Traceback follows in format "
+       << "ilabel (olabel olabel) ilabel (olabel) ... :";
+    for (ssize_t i = traceback.size() - 1; i >= 0; i--) {
+      ss << ' ' << traceback[i].first << " ( ";
+      std::vector<Label> seq;
+      repository_.ConvertToVector(traceback[i].second, &seq);
+      for (size_t j = 0; j < seq.size(); j++) ss << seq[j] << ' ';
+      ss << ')';
+    }
+    KALDI_ERR << ss.str();
+  }
+
+  bool IsIsymbolOrFinal(InputStateId state) {  // returns true if this state
+    // of the input FST either is final or has an osymbol on an arc out of it.
+    // Uses the vector isymbol_or_final_ as a cache for this info.
+    assert(state >= 0);
+    if (isymbol_or_final_.size() <= state)
+      isymbol_or_final_.resize(state + 1, static_cast<char>(OSF_UNKNOWN));
+    if (isymbol_or_final_[state] == static_cast<char>(OSF_NO))
+      return false;
+    else if (isymbol_or_final_[state] == static_cast<char>(OSF_YES))
+      return true;
+    // else work it out...
+    isymbol_or_final_[state] = static_cast<char>(OSF_NO);
+    if (ifst_->Final(state) != Weight::Zero())
+      isymbol_or_final_[state] = static_cast<char>(OSF_YES);
+    for (ArcIterator<Fst<Arc> > aiter(*ifst_, state); !aiter.Done();
+         aiter.Next()) {
+      const Arc &arc = aiter.Value();
+      if (arc.ilabel != 0 && arc.weight != Weight::Zero()) {
+        isymbol_or_final_[state] = static_cast<char>(OSF_YES);
+        return true;
+      }
+    }
+    return IsIsymbolOrFinal(state);  // will only recurse once.
+  }
+
+  void InitializeDeterminization() {
+    if (ifst_->Properties(kExpanded, false) != 0) {  // if we know the number of
+      // states in ifst_, it might be a bit more efficient
+      // to pre-size the hashes so we're not constantly rebuilding them.
+#if !(__GNUC__ == 4 && __GNUC_MINOR__ == 0)
+      StateId num_states =
+          down_cast<const ExpandedFst<Arc> *, const Fst<Arc> >(ifst_)
+              ->NumStates();
+      minimal_hash_.rehash(num_states / 2 + 3);
+      initial_hash_.rehash(num_states / 2 + 3);
+#endif
+    }
+    InputStateId start_id = ifst_->Start();
+    if (start_id != kNoStateId) {
+      /* Insert determinized-state corresponding to the start state into hash
+         and queue.  Unlike all the other states, we don't "normalize" the
+         representation of this determinized-state before we put it into
+         minimal_hash_.  This is actually what we want, as otherwise we'd have
+         problems dealing with any extra weight and string and might have to
+         create a "super-initial" state which would make the output
+         nondeterministic.  Normalization is only needed to make the
+         determinized output more minimal anyway, it's not needed for
+         correctness. Note, we don't put anything in the initial_hash_.  The
+         initial_hash_ is only a lookaside buffer anyway, so this isn't a
+         problem-- it will get populated later if it needs to be.
+      */
+      Element elem;
+      elem.state = start_id;
+      elem.weight = Weight::One();
+      elem.string = repository_.EmptyString();  // Id of empty sequence.
+      std::vector<Element> subset;
+      subset.push_back(elem);
+      EpsilonClosure(&subset);    // follow through epsilon-inputs links
+      ConvertToMinimal(&subset);  // remove all but final states and
+      // states with input-labels on arcs out of them.
+      std::vector<Element> *subset_ptr = new std::vector<Element>(subset);
+      assert(output_arcs_.empty() && output_states_.empty());
+      // add the new state...
+      output_states_.push_back(subset_ptr);
+      output_arcs_.push_back(std::vector<TempArc>());
+      OutputStateId initial_state = 0;
+      minimal_hash_[subset_ptr] = initial_state;
+      queue_.push_back(initial_state);
+    }
+  }
+
+  KALDI_DISALLOW_COPY_AND_ASSIGN(LatticeDeterminizer);
+
+  std::vector<std::vector<Element> *>
+      output_states_;  // maps from output state to
+                       // minimal representation [normalized].
+                       // View pointers as owned in
+                       // minimal_hash_.
+  std::vector<std::vector<TempArc> >
+      output_arcs_;  // essentially an FST in our format.
+
+  int num_arcs_;   // keep track of memory usage: number of arcs in output_arcs_
+  int num_elems_;  // keep track of memory usage: number of elems in
+                   // output_states_
+
+  const Fst<Arc> *ifst_;
+  DeterminizeLatticeOptions opts_;
+  SubsetKey hasher_;  // object that computes keys-- has no data members.
+  SubsetEqual
+      equal_;  // object that compares subsets-- only data member is delta_.
+  bool determinized_;  // set to true when user called Determinize(); used to
+                       // make
+  // sure this object is used correctly.
+  MinimalSubsetHash
+      minimal_hash_;  // hash from Subset to OutputStateId.  Subset is "minimal
+                      // representation" (only include final and states and
+                      // states with nonzero ilabel on arc out of them.  Owns
+                      // the pointers in its keys.
+  InitialSubsetHash initial_hash_;  // hash from Subset to Element, which
+                                    // represents the OutputStateId together
+                                    // with an extra weight and string.  Subset
+                                    // is "initial representation".  The extra
+                                    // weight and string is needed because after
+                                    // we convert to minimal representation and
+                                    // normalize, there may be an extra weight
+                                    // and string.  Owns the pointers
+                                    // in its keys.
+  std::vector<OutputStateId>
+      queue_;  // Queue of output-states to process.  Starts with
+  // state 0, and increases and then (hopefully) decreases in length during
+  // determinization.  LIFO queue (queue discipline doesn't really matter).
+
+  std::vector<std::pair<Label, Element> >
+      all_elems_tmp_;  // temporary vector used in ProcessTransitions.
+
+  enum IsymbolOrFinal { OSF_UNKNOWN = 0, OSF_NO = 1, OSF_YES = 2 };
+
+  std::vector<char> isymbol_or_final_;  // A kind of cache; it says whether
+  // each state is (emitting or final) where emitting means it has at least one
+  // non-epsilon output arc.  Only accessed by IsIsymbolOrFinal()
+
+  LatticeStringRepository<IntType>
+      repository_;  // defines a compact and fast way of
+  // storing sequences of labels.
+};
+
+// normally Weight would be LatticeWeight<float> (which has two floats),
+// or possibly TropicalWeightTpl<float>, and IntType would be int32.
+template <class Weight, class IntType>
+bool DeterminizeLattice(const Fst<ArcTpl<Weight> > &ifst,
+                        MutableFst<ArcTpl<Weight> > *ofst,
+                        DeterminizeLatticeOptions opts, bool *debug_ptr) {
+  ofst->SetInputSymbols(ifst.InputSymbols());
+  ofst->SetOutputSymbols(ifst.OutputSymbols());
+  LatticeDeterminizer<Weight, IntType> det(ifst, opts);
+  if (!det.Determinize(debug_ptr)) return false;
+  det.Output(ofst);
+  return true;
+}
+
+// normally Weight would be LatticeWeight<float> (which has two floats),
+// or possibly TropicalWeightTpl<float>, and IntType would be int32.
+template <class Weight, class IntType>
+bool DeterminizeLattice(
+    const Fst<ArcTpl<Weight> > &ifst,
+    MutableFst<ArcTpl<CompactLatticeWeightTpl<Weight, IntType> > > *ofst,
+    DeterminizeLatticeOptions opts, bool *debug_ptr) {
+  ofst->SetInputSymbols(ifst.InputSymbols());
+  ofst->SetOutputSymbols(ifst.OutputSymbols());
+  LatticeDeterminizer<Weight, IntType> det(ifst, opts);
+  if (!det.Determinize(debug_ptr)) return false;
+  det.Output(ofst);
+  return true;
+}
+
+}  // namespace fst
+
+#endif  // KALDI_FSTEXT_DETERMINIZE_LATTICE_INL_H_
diff --git a/speechx/speechx/kaldi/fstext/determinize-lattice.h b/speechx/speechx/kaldi/fstext/determinize-lattice.h
new file mode 100644
index 00000000..4a425119
--- /dev/null
+++ b/speechx/speechx/kaldi/fstext/determinize-lattice.h
@@ -0,0 +1,144 @@
+// fstext/determinize-lattice.h
+
+// Copyright 2009-2011  Microsoft Corporation
+
+// See ../../COPYING for clarification regarding multiple authors
+//
+// Licensed under the Apache License, Version 2.0 (the "License");
+// you may not use this file except in compliance with the License.
+// You may obtain a copy of the License at
+//
+//  http://www.apache.org/licenses/LICENSE-2.0
+//
+// THIS CODE IS PROVIDED *AS IS* BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY
+// KIND, EITHER EXPRESS OR IMPLIED, INCLUDING WITHOUT LIMITATION ANY IMPLIED
+// WARRANTIES OR CONDITIONS OF TITLE, FITNESS FOR A PARTICULAR PURPOSE,
+// MERCHANTABLITY OR NON-INFRINGEMENT.
+// See the Apache 2 License for the specific language governing permissions and
+// limitations under the License.
+
+#ifndef KALDI_FSTEXT_DETERMINIZE_LATTICE_H_
+#define KALDI_FSTEXT_DETERMINIZE_LATTICE_H_
+#include <fst/fst-decl.h>
+#include <fst/fstlib.h>
+#include <algorithm>
+#include <map>
+#include <set>
+#include <vector>
+#include "fstext/lattice-weight.h"
+
+namespace fst {
+
+/// \addtogroup fst_extensions
+///  @{
+
+// For example of usage, see test-determinize-lattice.cc
+
+/*
+   DeterminizeLattice implements a special form of determinization
+   with epsilon removal, optimized for a phase of lattice generation.
+   Its input is an FST with weight-type BaseWeightType (usually a pair of
+   floats, with a lexicographical type of order, such as
+   LatticeWeightTpl<float>). Typically this would be a state-level lattice, with
+   input symbols equal to words, and output-symbols equal to p.d.f's (so like
+   the inverse of HCLG).  Imagine representing this as an acceptor of type
+   CompactLatticeWeightTpl<float>, in which the input/output symbols are words,
+   and the weights contain the original weights together with strings (with zero
+   or one symbol in them) containing the original output labels (the p.d.f.'s).
+   We determinize this using acceptor determinization with epsilon removal.
+   Remember (from lattice-weight.h) that CompactLatticeWeightTpl has a special
+   kind of semiring where we always take the string corresponding to the best
+   cost (of type BaseWeightType), and discard the other.  This corresponds to
+   taking the best output-label sequence (of p.d.f.'s) for each input-label
+   sequence (of words).  We couldn't use the Gallic weight for this, or it would
+   die as soon as it detected that the input FST was non-functional.  In our
+   case, any acyclic FST (and many cyclic ones) can be determinized. We assume
+   that there is a function Compare(const BaseWeightType &a, const
+   BaseWeightType &b) that returns (-1, 0, 1) according to whether (a < b, a ==
+   b, a > b) in the total order on the BaseWeightType... this information should
+   be the same as NaturalLess would give, but it's more efficient to do it this
+   way. You can define this for things like TropicalWeight if you need to
+   instantiate this class for that weight type.
+
+   We implement this determinization in a special way to make it efficient for
+   the types of FSTs that we will apply it to.  One issue is that if we
+   explicitly represent the strings (in CompactLatticeWeightTpl) as vectors of
+   type vector<IntType>, the algorithm takes time quadratic in the length of
+   words (in states), because propagating each arc involves copying a whole
+   vector (of integers representing p.d.f.'s).  Instead we use a hash structure
+   where each string is a pointer (Entry*), and uses a hash from (Entry*,
+   IntType), to the successor string (and a way to get the latest IntType and
+   the ancestor Entry*).  [this is the class LatticeStringRepository].
+
+   Another issue is that rather than representing a determinized-state as a
+   collection of (state, weight), we represent it in a couple of reduced forms.
+   Suppose a determinized-state is a collection of (state, weight) pairs; call
+   this the "canonical representation".  Note: these collections are always
+   normalized to remove any common weight and string part.  Define end-states as
+   the subset of states that have an arc out of them with a label on, or are
+   final.  If we represent a determinized-state a the set of just its
+   (end-state, weight) pairs, this will be a valid and more compact
+   representation, and will lead to a smaller set of determinized states (like
+   early minimization).  Call this collection of (end-state, weight) pairs the
+   "minimal representation".  As a mechanism to reduce compute, we can also
+   consider another representation. In the determinization algorithm, we start
+   off with a set of (begin-state, weight) pairs (where the "begin-states" are
+   initial or have a label on the transition into them), and the "canonical
+   representation" consists of the epsilon-closure of this set (i.e. follow
+   epsilons).  Call this set of (begin-state, weight) pairs, appropriately
+   normalized, the "initial representation".  If two initial representations are
+   the same, the "canonical representation" and hence the "minimal
+   representation" will be the same.  We can use this to reduce compute.  Note
+   that if two initial representations are different, this does not preclude the
+   other representations from being the same.
+
+*/
+
+struct DeterminizeLatticeOptions {
+  float delta;  // A small offset used to measure equality of weights.
+  int max_mem;  // If >0, determinization will fail and return false
+  // when the algorithm's (approximate) memory consumption crosses this
+  // threshold.
+  int max_loop;  // If >0, can be used to detect non-determinizable input
+  // (a case that wouldn't be caught by max_mem).
+  DeterminizeLatticeOptions() : delta(kDelta), max_mem(-1), max_loop(-1) {}
+};
+
+/**
+    This function implements the normal version of DeterminizeLattice, in which
+    the output strings are represented using sequences of arcs, where all but
+    the first one has an epsilon on the input side.  The debug_ptr argument is
+    an optional pointer to a bool that, if it becomes true while the algorithm
+    is executing, the algorithm will print a traceback and terminate (used in
+    fstdeterminizestar.cc debug non-terminating determinization).  More
+    efficient if ifst is arc-sorted on input label.  If the number of arcs gets
+    more than max_states, it will throw std::runtime_error (otherwise this code
+    does not use exceptions).  This is mainly useful for debug.  */
+template <class Weight, class IntType>
+bool DeterminizeLattice(
+    const Fst<ArcTpl<Weight> > &ifst, MutableFst<ArcTpl<Weight> > *ofst,
+    DeterminizeLatticeOptions opts = DeterminizeLatticeOptions(),
+    bool *debug_ptr = NULL);
+
+/*  This is a version of DeterminizeLattice with a slightly more "natural"
+   output format, where the output sequences are encoded using the
+   CompactLatticeArcTpl template (i.e. the sequences of output symbols are
+   represented directly as strings) More efficient if ifst is arc-sorted on
+   input label. If the #arcs gets more than max_arcs, it will throw
+   std::runtime_error (otherwise this code does not use exceptions).  This is
+   mainly useful for debug.
+*/
+template <class Weight, class IntType>
+bool DeterminizeLattice(
+    const Fst<ArcTpl<Weight> > &ifst,
+    MutableFst<ArcTpl<CompactLatticeWeightTpl<Weight, IntType> > > *ofst,
+    DeterminizeLatticeOptions opts = DeterminizeLatticeOptions(),
+    bool *debug_ptr = NULL);
+
+/// @} end "addtogroup fst_extensions"
+
+}  // end namespace fst
+
+#include "fstext/determinize-lattice-inl.h"
+
+#endif  // KALDI_FSTEXT_DETERMINIZE_LATTICE_H_
diff --git a/speechx/speechx/kaldi/fstext/determinize-star-inl.h b/speechx/speechx/kaldi/fstext/determinize-star-inl.h
new file mode 100644
index 00000000..b5b6d2fa
--- /dev/null
+++ b/speechx/speechx/kaldi/fstext/determinize-star-inl.h
@@ -0,0 +1,1204 @@
+// fstext/determinize-star-inl.h
+
+// Copyright 2009-2011  Microsoft Corporation;  Jan Silovsky
+//           2015 Hainan Xu
+
+// See ../../COPYING for clarification regarding multiple authors
+//
+// Licensed under the Apache License, Version 2.0 (the "License");
+// you may not use this file except in compliance with the License.
+// You may obtain a copy of the License at
+//
+//  http://www.apache.org/licenses/LICENSE-2.0
+//
+// THIS CODE IS PROVIDED *AS IS* BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY
+// KIND, EITHER EXPRESS OR IMPLIED, INCLUDING WITHOUT LIMITATION ANY IMPLIED
+// WARRANTIES OR CONDITIONS OF TITLE, FITNESS FOR A PARTICULAR PURPOSE,
+// MERCHANTABLITY OR NON-INFRINGEMENT.
+// See the Apache 2 License for the specific language governing permissions and
+// limitations under the License.
+
+#ifndef KALDI_FSTEXT_DETERMINIZE_STAR_INL_H_
+#define KALDI_FSTEXT_DETERMINIZE_STAR_INL_H_
+// Do not include this file directly.  It is included by determinize-star.h
+
+#include <algorithm>
+#include <climits>
+#include <deque>
+#include <limits>
+#include <string>
+#include <unordered_map>
+#include <utility>
+#include <vector>
+using std::unordered_map;
+
+#include "base/kaldi-error.h"
+
+namespace fst {
+
+// This class maps back and forth from/to integer id's to sequences of strings.
+// used in determinization algorithm.
+
+template <class Label, class StringId>
+class StringRepository {
+  // Label and StringId are both integer types, possibly the same.
+  // This is a utility that maps back and forth between a vector<Label> and
+  // StringId representation of sequences of Labels.  It is to save memory, and
+  // to save compute. We treat sequences of length zero and one separately, for
+  // efficiency.
+
+ public:
+  class VectorKey {  // Hash function object.
+   public:
+    size_t operator()(const std::vector<Label> *vec) const {
+      assert(vec != NULL);
+      size_t hash = 0, factor = 1;
+      for (typename std::vector<Label>::const_iterator it = vec->begin();
+           it != vec->end(); it++) {
+        hash += factor * (*it);
+        factor *= 103333;  // just an arbitrary prime number.
+      }
+      return hash;
+    }
+  };
+  class VectorEqual {  // Equality-operator function object.
+   public:
+    size_t operator()(const std::vector<Label> *vec1,
+                      const std::vector<Label> *vec2) const {
+      return (*vec1 == *vec2);
+    }
+  };
+
+  typedef unordered_map<const std::vector<Label> *, StringId, VectorKey,
+                        VectorEqual>
+      MapType;
+
+  StringId IdOfEmpty() { return no_symbol; }
+
+  StringId IdOfLabel(Label l) {
+    if (l >= 0 && l <= (Label)single_symbol_range) {
+      return l + single_symbol_start;
+    } else {
+      // l is out of the allowed range so we have to treat it as a sequence of
+      // length one.  Should be v. rare.
+      std::vector<Label> v;
+      v.push_back(l);
+      return IdOfSeqInternal(v);
+    }
+  }
+
+  StringId IdOfSeq(
+      const std::vector<Label> &v) {  // also works for sizes 0 and 1.
+    size_t sz = v.size();
+    if (sz == 0)
+      return no_symbol;
+    else if (v.size() == 1)
+      return IdOfLabel(v[0]);
+    else
+      return IdOfSeqInternal(v);
+  }
+
+  inline bool IsEmptyString(StringId id) { return id == no_symbol; }
+  void SeqOfId(StringId id, std::vector<Label> *v) {
+    if (id == no_symbol) {
+      v->clear();
+    } else if (id >= single_symbol_start) {
+      v->resize(1);
+      (*v)[0] = id - single_symbol_start;
+    } else {
+      assert(static_cast<size_t>(id) < vec_.size());
+      *v = *(vec_[id]);
+    }
+  }
+  StringId RemovePrefix(StringId id, size_t prefix_len) {
+    if (prefix_len == 0) {
+      return id;
+    } else {
+      std::vector<Label> v;
+      SeqOfId(id, &v);
+      size_t sz = v.size();
+      assert(sz >= prefix_len);
+      std::vector<Label> v_noprefix(sz - prefix_len);
+      for (size_t i = 0; i < sz - prefix_len; i++)
+        v_noprefix[i] = v[i + prefix_len];
+      return IdOfSeq(v_noprefix);
+    }
+  }
+
+  StringRepository() {
+    // The following are really just constants but don't want to complicate
+    // compilation so make them class variables.  Due to the brokenness of
+    // <limits>, they can't be accessed as constants.
+    string_end = (std::numeric_limits<StringId>::max() / 2) -
+                 1;  // all hash values must be <= this.
+    no_symbol = (std::numeric_limits<StringId>::max() /
+                 2);  // reserved for empty sequence.
+    single_symbol_start = (std::numeric_limits<StringId>::max() / 2) + 1;
+    single_symbol_range =
+        std::numeric_limits<StringId>::max() - single_symbol_start;
+  }
+  void Destroy() {
+    for (typename std::vector<std::vector<Label> *>::iterator iter =
+             vec_.begin();
+         iter != vec_.end(); ++iter)
+      delete *iter;
+    std::vector<std::vector<Label> *> tmp_vec;
+    tmp_vec.swap(vec_);
+    MapType tmp_map;
+    tmp_map.swap(map_);
+  }
+  ~StringRepository() { Destroy(); }
+
+ private:
+  KALDI_DISALLOW_COPY_AND_ASSIGN(StringRepository);
+
+  StringId IdOfSeqInternal(const std::vector<Label> &v) {
+    typename MapType::iterator iter = map_.find(&v);
+    if (iter != map_.end()) {
+      return iter->second;
+    } else {  // must add it to map.
+      StringId this_id = (StringId)vec_.size();
+      std::vector<Label> *v_new = new std::vector<Label>(v);
+      vec_.push_back(v_new);
+      map_[v_new] = this_id;
+      assert(this_id < string_end);  // or we used up the labels.
+      return this_id;
+    }
+  }
+
+  std::vector<std::vector<Label> *> vec_;
+  MapType map_;
+
+  static const StringId string_start =
+      (StringId)0;      // This must not change.  It's assumed.
+  StringId string_end;  // = (numeric_limits<StringId>::max() / 2) - 1; // all
+                        // hash values must be <= this.
+  StringId no_symbol;   // = (numeric_limits<StringId>::max() / 2); // reserved
+                        // for empty sequence.
+  StringId
+      single_symbol_start;  // =  (numeric_limits<StringId>::max() / 2) + 1;
+  StringId single_symbol_range;  // =  numeric_limits<StringId>::max() -
+                                 // single_symbol_start;
+};
+
+template <class F>
+class DeterminizerStar {
+  typedef typename F::Arc Arc;
+
+ public:
+  // Output to Gallic acceptor (so the strings go on weights, and there is a 1-1
+  // correspondence between our states and the states in ofst.  If destroy ==
+  // true, release memory as we go (but we cannot output again).
+  void Output(MutableFst<GallicArc<Arc> > *ofst, bool destroy = true);
+
+  // Output to standard FST.  We will create extra states to handle sequences of
+  // symbols on the output.  If destroy == true, release memory as we go (but we
+  // cannot output again).
+
+  void Output(MutableFst<Arc> *ofst, bool destroy = true);
+
+  // Initializer.  After initializing the object you will typically call
+  // Determinize() and then one of the Output functions.
+  DeterminizerStar(const Fst<Arc> &ifst, float delta = kDelta,
+                   int max_states = -1, bool allow_partial = false)
+      : ifst_(ifst.Copy()),
+        delta_(delta),
+        max_states_(max_states),
+        determinized_(false),
+        allow_partial_(allow_partial),
+        is_partial_(false),
+        equal_(delta),
+        hash_(ifst.Properties(kExpanded, false)
+                  ? down_cast<const ExpandedFst<Arc> *, const Fst<Arc> >(&ifst)
+                                ->NumStates() /
+                            2 +
+                        3
+                  : 20,
+              hasher_, equal_),
+        epsilon_closure_(ifst_, max_states, &repository_, delta) {}
+
+  void Determinize(bool *debug_ptr) {
+    assert(!determinized_);
+    // This determinizes the input fst but leaves it in the "special format"
+    // in "output_arcs_".
+    InputStateId start_id = ifst_->Start();
+    if (start_id == kNoStateId) {
+      determinized_ = true;
+      return;  // Nothing to do.
+    } else {   // Insert start state into hash and queue.
+      Element elem;
+      elem.state = start_id;
+      elem.weight = Weight::One();
+      elem.string = repository_.IdOfEmpty();  // Id of empty sequence.
+      std::vector<Element> vec;
+      vec.push_back(elem);
+      OutputStateId cur_id = SubsetToStateId(vec);
+      assert(cur_id == 0 && "Do not call Determinize twice.");
+    }
+    while (!Q_.empty()) {
+      std::pair<std::vector<Element> *, OutputStateId> cur_pair = Q_.front();
+      Q_.pop_front();
+      ProcessSubset(cur_pair);
+      if (debug_ptr && *debug_ptr) Debug();  // will exit.
+      if (max_states_ > 0 && output_arcs_.size() > max_states_) {
+        if (allow_partial_ == false) {
+          KALDI_ERR << "Determinization aborted since passed " << max_states_
+                    << " states";
+        } else {
+          KALDI_WARN << "Determinization terminated since passed "
+                     << max_states_
+                     << " states, partial results will be generated";
+          is_partial_ = true;
+          break;
+        }
+      }
+    }
+    determinized_ = true;
+  }
+
+  bool IsPartial() { return is_partial_; }
+
+  // frees all except output_arcs_, which contains the important info
+  // we need to output.
+  void FreeMostMemory() {
+    if (ifst_) {
+      delete ifst_;
+      ifst_ = NULL;
+    }
+    for (typename SubsetHash::iterator iter = hash_.begin();
+         iter != hash_.end(); ++iter)
+      delete iter->first;
+    SubsetHash tmp;
+    tmp.swap(hash_);
+  }
+
+  ~DeterminizerStar() { FreeMostMemory(); }
+
+ private:
+  typedef typename Arc::Label Label;
+  typedef typename Arc::Weight Weight;
+  typedef typename Arc::StateId InputStateId;
+  typedef typename Arc::StateId
+      OutputStateId;  // same as above but distinguish states in output Fst.
+  typedef typename Arc::Label StringId;  // Id type used in the StringRepository
+  typedef StringRepository<Label, StringId> StringRepositoryType;
+
+  // Element of a subset [of original states]
+
+  struct Element {
+    InputStateId state;
+    StringId string;
+    Weight weight;
+    bool operator!=(const Element &other) const {
+      return (state != other.state || string != other.string ||
+              weight != other.weight);
+    }
+  };
+
+  // Arcs in the format we temporarily create in this class (a representation,
+  // essentially of a Gallic Fst).
+  struct TempArc {
+    Label ilabel;
+    StringId ostring;  // Look it up in the StringRepository, it's a sequence of
+                       // Labels.
+    OutputStateId nextstate;  // or kNoState for final weights.
+    Weight weight;
+  };
+
+  // Hashing function used in hash of subsets.
+  // A subset is a pointer to vector<Element>.
+  // The Elements are in sorted order on state id, and without repeated states.
+  // Because the order of Elements is fixed, we can use a hashing function that
+  // is order-dependent.  However the weights are not included in the hashing
+  // function-- we hash subsets that differ only in weight to the same key. This
+  // is not optimal in terms of the O(N) performance but typically if we have a
+  // lot of determinized states that differ only in weight then the input
+  // probably was pathological in some way, or even non-determinizable.
+  //   We don't quantize the weights, in order to avoid inexactness in simple
+  //   cases.
+  // Instead we apply the delta when comparing subsets for equality, and allow a
+  // small difference.
+
+  class SubsetKey {
+   public:
+    size_t operator()(const std::vector<Element> *subset)
+        const {  // hashes only the state and string.
+      size_t hash = 0, factor = 1;
+      for (typename std::vector<Element>::const_iterator iter = subset->begin();
+           iter != subset->end(); ++iter) {
+        hash *= factor;
+        hash += iter->state + 103333 * iter->string;
+        factor *= 23531;  // these numbers are primes.
+      }
+      return hash;
+    }
+  };
+
+  // This is the equality operator on subsets.  It checks for exact match on
+  // state-id and string, and approximate match on weights.
+  class SubsetEqual {
+   public:
+    bool operator()(const std::vector<Element> *s1,
+                    const std::vector<Element> *s2) const {
+      size_t sz = s1->size();
+      assert(sz >= 0);
+      if (sz != s2->size()) return false;
+      typename std::vector<Element>::const_iterator iter1 = s1->begin(),
+                                                    iter1_end = s1->end(),
+                                                    iter2 = s2->begin();
+      for (; iter1 < iter1_end; ++iter1, ++iter2) {
+        if (iter1->state != iter2->state || iter1->string != iter2->string ||
+            !ApproxEqual(iter1->weight, iter2->weight, delta_))
+          return false;
+      }
+      return true;
+    }
+    float delta_;
+    explicit SubsetEqual(float delta) : delta_(delta) {}
+    SubsetEqual() : delta_(kDelta) {}
+  };
+
+  // Operator that says whether two Elements have the same states.
+  // Used only for debug.
+  class SubsetEqualStates {
+   public:
+    bool operator()(const std::vector<Element> *s1,
+                    const std::vector<Element> *s2) const {
+      size_t sz = s1->size();
+      assert(sz >= 0);
+      if (sz != s2->size()) return false;
+      typename std::vector<Element>::const_iterator iter1 = s1->begin(),
+                                                    iter1_end = s1->end(),
+                                                    iter2 = s2->begin();
+      for (; iter1 < iter1_end; ++iter1, ++iter2) {
+        if (iter1->state != iter2->state) return false;
+      }
+      return true;
+    }
+  };
+
+  // Define the hash type we use to store subsets.
+  typedef unordered_map<const std::vector<Element> *, OutputStateId, SubsetKey,
+                        SubsetEqual>
+      SubsetHash;
+
+  class EpsilonClosure {
+   public:
+    EpsilonClosure(const Fst<Arc> *ifst, int max_states,
+                   StringRepository<Label, StringId> *repository, float delta)
+        : ifst_(ifst),
+          max_states_(max_states),
+          repository_(repository),
+          delta_(delta) {}
+
+    // This function computes epsilon closure of subset of states by following
+    // epsilon links. Called by ProcessSubset. Has no side effects except on the
+    // repository.
+    void GetEpsilonClosure(const std::vector<Element> &input_subset,
+                           std::vector<Element> *output_subset);
+
+   private:
+    struct EpsilonClosureInfo {
+      EpsilonClosureInfo() {}
+      EpsilonClosureInfo(const Element &e, const Weight &w, bool i)
+          : element(e), weight_to_process(w), in_queue(i) {}
+      // the weight in the Element struct is the total current weight
+      // that has been processed already
+      Element element;
+      // this stores the weight that we haven't processed (propagated)
+      Weight weight_to_process;
+      // whether "this" struct is in the queue
+      // we store the info here so that we don't have to look it up every time
+      bool in_queue;
+      bool operator<(const EpsilonClosureInfo &other) const {
+        return this->element.state < other.element.state;
+      }
+    };
+
+    // to further speed up EpsilonClosure() computation, we have 2 queues
+    // the 2nd queue is used when we first iterate over the input set -
+    // if queue_2_.empty() then we directly set output_set equal to input_set
+    // and return immediately
+    // Since Epsilon arcs are relatively rare, this way we could efficiently
+    // detect the epsilon-free case, without having to waste our computation
+    // e.g. allocating the EpsilonClosureInfo structure; this also lets us do a
+    // level-by-level traversal, which could avoid some (unfortunately not all)
+    // duplicate computation if epsilons form a DAG that is not a tree
+    //
+    // We put the queues here for better efficiency for memory allocation
+    std::deque<typename Arc::StateId> queue_;
+    std::vector<Element> queue_2_;
+
+    // the following 2 structures together form our *virtual "map"*
+    // basically we need a map from state_id to EpsilonClosureInfo that operates
+    // in O(1) time, while still takes relatively small mem, and this does it
+    // well for efficiency we don't clear id_to_index_ of its outdated
+    // information As a result each time we do a look-up, we need to check if
+    // (ecinfo_[id_to_index_[id]].element.state == id) Yet this is still faster
+    // than using a std::map<StateId, EpsilonClosureInfo>
+    std::vector<int> id_to_index_;
+    // unlike id_to_index_, we clear the content of ecinfo_ each time we call
+    // EpsilonClosure(). This needed because we need an efficient way to
+    // traverse the virtual map - it is just too costly to traverse the
+    // id_to_index_ vector.
+    std::vector<EpsilonClosureInfo> ecinfo_;
+
+    // Add one element (elem) into cur_subset
+    // it also adds the necessary stuff to queue_, set the correct weight
+    void AddOneElement(const Element &elem, const Weight &unprocessed_weight);
+
+    // Sub-routine that we call in EpsilonClosure()
+    // It takes the current "unprocessed_weight" and propagate it to the
+    // states accessible from elem.state by an epsilon arc
+    // and add the results to cur_subset.
+    // save_to_queue_2 is set true when we iterate over the initial subset
+    // - then we save it to queue_2 s.t. if it's empty, we directly return
+    // the input set
+    void ExpandOneElement(const Element &elem, bool sorted,
+                          const Weight &unprocessed_weight,
+                          bool save_to_queue_2 = false);
+
+    // no pointers below would take the ownership
+    const Fst<Arc> *ifst_;
+    int max_states_;
+    StringRepository<Label, StringId> *repository_;
+    float delta_;
+  };
+
+  // This function works out the final-weight of the determinized state.
+  // called by ProcessSubset.
+  // Has no side effects except on the variable repository_, and output_arcs_.
+
+  void ProcessFinal(const std::vector<Element> &closed_subset,
+                    OutputStateId state) {
+    // processes final-weights for this subset.
+    bool is_final = false;
+    StringId final_string = 0;  // = 0 to keep compiler happy.
+    Weight final_weight =
+        Weight::One();  // This value will never be accessed, and
+    // we just set it to avoid spurious compiler warnings.  We avoid setting it
+    // to Zero() because floating-point infinities can sometimes generate
+    // interrupts and slow things down.
+    typename std::vector<Element>::const_iterator iter = closed_subset.begin(),
+                                                  end = closed_subset.end();
+    for (; iter != end; ++iter) {
+      const Element &elem = *iter;
+      Weight this_final_weight = ifst_->Final(elem.state);
+      if (this_final_weight != Weight::Zero()) {
+        if (!is_final) {  // first final-weight
+          final_string = elem.string;
+          final_weight = Times(elem.weight, this_final_weight);
+          is_final = true;
+        } else {  // already have one.
+          if (final_string != elem.string) {
+            KALDI_ERR << "FST was not functional -> not determinizable";
+          }
+          final_weight =
+              Plus(final_weight, Times(elem.weight, this_final_weight));
+        }
+      }
+    }
+    if (is_final) {
+      // store final weights in TempArc structure, just like a transition.
+      TempArc temp_arc;
+      temp_arc.ilabel = 0;
+      temp_arc.nextstate =
+          kNoStateId;  // special marker meaning "final weight".
+      temp_arc.ostring = final_string;
+      temp_arc.weight = final_weight;
+      output_arcs_[state].push_back(temp_arc);
+    }
+  }
+
+  // ProcessTransition is called from "ProcessTransitions".  Broken out for
+  // clarity.  Has side effects on output_arcs_, and (via SubsetToStateId), Q_
+  // and hash_.
+  void ProcessTransition(OutputStateId state, Label ilabel,
+                         std::vector<Element> *subset);
+
+  // "less than" operator for pair<Label, Element>.   Used in
+  // ProcessTransitions. Lexicographical order, with comparing the state only
+  // for "Element".
+
+  class PairComparator {
+   public:
+    inline bool operator()(const std::pair<Label, Element> &p1,
+                           const std::pair<Label, Element> &p2) {
+      if (p1.first < p2.first) {
+        return true;
+      } else if (p1.first > p2.first) {
+        return false;
+      } else {
+        return p1.second.state < p2.second.state;
+      }
+    }
+  };
+
+  // ProcessTransitions handles transitions out of this subset of states.
+  // Ignores epsilon transitions (epsilon closure already handled that).
+  // Does not consider final states.  Breaks the transitions up by ilabel,
+  // and creates a new transition in determinized FST, for each ilabel.
+  // Does this by creating a big vector of pairs <Label, Element> and then
+  // sorting them using a lexicographical ordering, and calling
+  // ProcessTransition for each range with the same ilabel. Side effects on
+  // repository, and (via ProcessTransition) on Q_, hash_, and output_arcs_.
+  void ProcessTransitions(const std::vector<Element> &closed_subset,
+                          OutputStateId state) {
+    std::vector<std::pair<Label, Element> > all_elems;
+    {  // Push back into "all_elems", elements corresponding to all
+       // non-epsilon-input transitions
+      // out of all states in "closed_subset".
+      typename std::vector<Element>::const_iterator iter =
+                                                        closed_subset.begin(),
+                                                    end = closed_subset.end();
+      for (; iter != end; ++iter) {
+        const Element &elem = *iter;
+        for (ArcIterator<Fst<Arc> > aiter(*ifst_, elem.state); !aiter.Done();
+             aiter.Next()) {
+          const Arc &arc = aiter.Value();
+          if (arc.ilabel !=
+              0) {  // Non-epsilon transition -- ignore epsilons here.
+            std::pair<Label, Element> this_pr;
+            this_pr.first = arc.ilabel;
+            Element &next_elem(this_pr.second);
+            next_elem.state = arc.nextstate;
+            next_elem.weight = Times(elem.weight, arc.weight);
+            if (arc.olabel == 0) {  // output epsilon-- this is simple case so
+                                    // handle separately for efficiency
+              next_elem.string = elem.string;
+            } else {
+              std::vector<Label> seq;
+              repository_.SeqOfId(elem.string, &seq);
+              seq.push_back(arc.olabel);
+              next_elem.string = repository_.IdOfSeq(seq);
+            }
+            all_elems.push_back(this_pr);
+          }
+        }
+      }
+    }
+    PairComparator pc;
+    std::sort(all_elems.begin(), all_elems.end(), pc);
+    // now sorted first on input label, then on state.
+    typedef typename std::vector<std::pair<Label, Element> >::const_iterator
+        PairIter;
+    PairIter cur = all_elems.begin(), end = all_elems.end();
+    std::vector<Element> this_subset;
+    while (cur != end) {
+      // Process ranges that share the same input symbol.
+      Label ilabel = cur->first;
+      this_subset.clear();
+      while (cur != end && cur->first == ilabel) {
+        this_subset.push_back(cur->second);
+        cur++;
+      }
+      // We now have a subset for this ilabel.
+      ProcessTransition(state, ilabel, &this_subset);
+    }
+  }
+
+  // SubsetToStateId converts a subset (vector of Elements) to a StateId in the
+  // output fst.  This is a hash lookup; if no such state exists, it adds a new
+  // state to the hash and adds a new pair to the queue. Side effects on hash_
+  // and Q_, and on output_arcs_ [just affects the size].
+  OutputStateId SubsetToStateId(
+      const std::vector<Element> &subset) {  // may add the subset to the queue.
+    typedef typename SubsetHash::iterator IterType;
+    IterType iter = hash_.find(&subset);
+    if (iter == hash_.end()) {  // was not there.
+      std::vector<Element> *new_subset = new std::vector<Element>(subset);
+      OutputStateId new_state_id = (OutputStateId)output_arcs_.size();
+      bool ans =
+          hash_
+              .insert(std::pair<const std::vector<Element> *, OutputStateId>(
+                  new_subset, new_state_id))
+              .second;
+      assert(ans);
+      output_arcs_.push_back(std::vector<TempArc>());
+      if (allow_partial_ == false) {
+        // If --allow-partial is not requested, we do the old way.
+        Q_.push_front(std::pair<std::vector<Element> *, OutputStateId>(
+            new_subset, new_state_id));
+      } else {
+        // If --allow-partial is requested, we do breadth first search. This
+        // ensures that when we return partial results, we return the states
+        // that are reachable by the fewest steps from the start state.
+        Q_.push_back(std::pair<std::vector<Element> *, OutputStateId>(
+            new_subset, new_state_id));
+      }
+      return new_state_id;
+    } else {
+      return iter->second;  // the OutputStateId.
+    }
+  }
+
+  // ProcessSubset does the processing of a determinized state, i.e. it creates
+  // transitions out of it and adds new determinized states to the queue if
+  // necessary. The first stage is "EpsilonClosure" (follow epsilons to get a
+  // possibly larger set of (states, weights)).  After that we ignore epsilons.
+  // We process the final-weight of the state, and then handle transitions out
+  // (this may add more determinized states to the queue).
+  void ProcessSubset(
+      const std::pair<std::vector<Element> *, OutputStateId> &pair) {
+    const std::vector<Element> *subset = pair.first;
+    OutputStateId state = pair.second;
+
+    std::vector<Element> closed_subset;  // subset after epsilon closure.
+    epsilon_closure_.GetEpsilonClosure(*subset, &closed_subset);
+
+    // Now follow non-epsilon arcs [and also process final states]
+    ProcessFinal(closed_subset, state);
+
+    // Now handle transitions out of these states.
+    ProcessTransitions(closed_subset, state);
+  }
+
+  void Debug();
+
+  KALDI_DISALLOW_COPY_AND_ASSIGN(DeterminizerStar);
+  std::deque<std::pair<std::vector<Element> *, OutputStateId> >
+      Q_;  // queue of subsets to be processed.
+
+  std::vector<std::vector<TempArc> >
+      output_arcs_;  // essentially an FST in our format.
+
+  const Fst<Arc> *ifst_;
+  float delta_;
+  int max_states_;
+  bool determinized_;   // used to check usage.
+  bool allow_partial_;  // output paritial results or not
+  bool is_partial_;     // if we get partial results or not
+  SubsetKey hasher_;    // object that computes keys-- has no data members.
+  SubsetEqual
+      equal_;  // object that compares subsets-- only data member is delta_.
+  SubsetHash hash_;  // hash from Subset to StateId in final Fst.
+
+  StringRepository<Label, StringId>
+      repository_;  // associate integer id's with sequences of labels.
+  EpsilonClosure epsilon_closure_;
+};
+
+template <class F>
+bool DeterminizeStar(F &ifst,  // NOLINT
+                     MutableFst<typename F::Arc> *ofst, float delta,
+                     bool *debug_ptr, int max_states, bool allow_partial) {
+  ofst->SetOutputSymbols(ifst.OutputSymbols());
+  ofst->SetInputSymbols(ifst.InputSymbols());
+  DeterminizerStar<F> det(ifst, delta, max_states, allow_partial);
+  det.Determinize(debug_ptr);
+  det.Output(ofst);
+  return det.IsPartial();
+}
+
+template <class F>
+bool DeterminizeStar(F &ifst,  // NOLINT
+                     MutableFst<GallicArc<typename F::Arc> > *ofst, float delta,
+                     bool *debug_ptr, int max_states, bool allow_partial) {
+  ofst->SetOutputSymbols(ifst.InputSymbols());
+  ofst->SetInputSymbols(ifst.InputSymbols());
+  DeterminizerStar<F> det(ifst, delta, max_states, allow_partial);
+  det.Determinize(debug_ptr);
+  det.Output(ofst);
+  return det.IsPartial();
+}
+
+template <class F>
+void DeterminizerStar<F>::EpsilonClosure::GetEpsilonClosure(
+    const std::vector<Element> &input_subset,
+    std::vector<Element> *output_subset) {
+  ecinfo_.resize(0);
+  size_t size = input_subset.size();
+  // find whether input fst is known to be sorted in input label.
+  bool sorted =
+      ((ifst_->Properties(kILabelSorted, false) & kILabelSorted) != 0);
+
+  // size is still the input_subset.size()
+  for (size_t i = 0; i < size; i++) {
+    ExpandOneElement(input_subset[i], sorted, input_subset[i].weight, true);
+  }
+
+  size_t s = queue_2_.size();
+  if (s == 0) {
+    *output_subset = input_subset;
+    return;
+  } else {
+    // queue_2 not empty. Need to create the vector<info>
+    for (size_t i = 0; i < size; i++) {
+      // the weight has not been processed yet,
+      // so put all of them in the "weight_to_process"
+      ecinfo_.push_back(
+          EpsilonClosureInfo(input_subset[i], input_subset[i].weight, false));
+      ecinfo_.back().element.weight = Weight::Zero();  // clear the weight
+
+      if (id_to_index_.size() < input_subset[i].state + 1) {
+        id_to_index_.resize(2 * input_subset[i].state + 1, -1);
+      }
+      id_to_index_[input_subset[i].state] = ecinfo_.size() - 1;
+    }
+  }
+
+  {
+    Element elem;
+    elem.weight = Weight::Zero();
+    for (size_t i = 0; i < s; i++) {
+      elem.state = queue_2_[i].state;
+      elem.string = queue_2_[i].string;
+      AddOneElement(elem, queue_2_[i].weight);
+    }
+    queue_2_.resize(0);
+  }
+
+  int counter = 0;  // relates to max-states option, used for test.
+  while (!queue_.empty()) {
+    InputStateId id = queue_.front();
+
+    // no need to check validity of the index
+    // since anything in the queue we are sure they're in the "virtual set"
+    int index = id_to_index_[id];
+    EpsilonClosureInfo &info = ecinfo_[index];
+    Element &elem = info.element;
+    Weight unprocessed_weight = info.weight_to_process;
+
+    elem.weight = Plus(elem.weight, unprocessed_weight);
+    info.weight_to_process = Weight::Zero();
+
+    info.in_queue = false;
+    queue_.pop_front();
+
+    if (max_states_ > 0 && counter++ > max_states_) {
+      KALDI_ERR << "Determinization aborted since looped more than "
+                << max_states_ << " times during epsilon closure";
+    }
+
+    // generally we need to be careful about iterator-invalidation problem
+    // here we pass a reference (elem), which could be an issue.
+    // In the beginning of ExpandOneElement, we make a copy of elem.string
+    // to avoid that issue
+    ExpandOneElement(elem, sorted, unprocessed_weight);
+  }
+
+  {
+    // this sorting is based on StateId
+    sort(ecinfo_.begin(), ecinfo_.end());
+
+    output_subset->clear();
+
+    size = ecinfo_.size();
+    output_subset->reserve(size);
+    for (size_t i = 0; i < size; i++) {
+      EpsilonClosureInfo &info = ecinfo_[i];
+      if (info.weight_to_process != Weight::Zero()) {
+        info.element.weight = Plus(info.element.weight, info.weight_to_process);
+      }
+      output_subset->push_back(info.element);
+    }
+  }
+}
+
+template <class F>
+void DeterminizerStar<F>::EpsilonClosure::AddOneElement(
+    const Element &elem, const Weight &unprocessed_weight) {
+  // first we try to find the element info in the ecinfo_ vector
+  int index = -1;
+  if (elem.state < id_to_index_.size()) {
+    index = id_to_index_[elem.state];
+  }
+  if (index != -1) {
+    if (index >= ecinfo_.size()) {
+      index = -1;
+    } else if (ecinfo_[index].element.state != elem.state) {
+      // since ecinfo_ might store outdated information, we need to check
+      index = -1;
+    }
+  }
+
+  if (index == -1) {
+    // was no such StateId: insert and add to queue.
+    ecinfo_.push_back(EpsilonClosureInfo(elem, unprocessed_weight, true));
+    size_t size = id_to_index_.size();
+    if (size < elem.state + 1) {
+      // double the size to reduce memory operations
+      id_to_index_.resize(2 * elem.state + 1, -1);
+    }
+    id_to_index_[elem.state] = ecinfo_.size() - 1;
+    queue_.push_back(elem.state);
+
+  } else {  // one is already there.  Add weights.
+    EpsilonClosureInfo &info = ecinfo_[index];
+    if (info.element.string != elem.string) {
+      // Non-functional FST.
+      std::ostringstream ss;
+      ss << "FST was not functional -> not determinizable.";
+      {  // Print some debugging information.  Can be helpful to debug
+        // the inputs when FSTs are mysteriously non-functional.
+        std::vector<Label> tmp_seq;
+        repository_->SeqOfId(info.element.string, &tmp_seq);
+        ss << "\nFirst string:";
+        for (size_t i = 0; i < tmp_seq.size(); i++) ss << ' ' << tmp_seq[i];
+        ss << "\nSecond string:";
+        repository_->SeqOfId(elem.string, &tmp_seq);
+        for (size_t i = 0; i < tmp_seq.size(); i++) ss << ' ' << tmp_seq[i];
+      }
+      KALDI_ERR << ss.str();
+    }
+
+    info.weight_to_process = Plus(info.weight_to_process, unprocessed_weight);
+
+    if (!info.in_queue) {
+      // this is because the code in "else" below: the
+      // iter->second.weight_to_process might not be Zero()
+      Weight weight = Plus(info.element.weight, info.weight_to_process);
+
+      // What is done below is, we propagate the weight (by adding them
+      // to the queue only when the change is big enough;
+      // otherwise we just store the weight, until before returning
+      // we add the element.weight and weight_to_process together
+      if (!ApproxEqual(weight, info.element.weight, delta_)) {
+        // add extra part of weight to queue.
+        info.in_queue = true;
+        queue_.push_back(elem.state);
+      }
+    }
+  }
+}
+
+template <class F>
+void DeterminizerStar<F>::EpsilonClosure::ExpandOneElement(
+    const Element &elem, bool sorted, const Weight &unprocessed_weight,
+    bool save_to_queue_2) {
+  StringId str =
+      elem.string;  // copy it here because there is an iterator-
+                    // - invalidation problem (it really happens for some FSTs)
+
+  // now we are going to propagate the "unprocessed_weight"
+  for (ArcIterator<Fst<Arc> > aiter(*ifst_, elem.state); !aiter.Done();
+       aiter.Next()) {
+    const Arc &arc = aiter.Value();
+    if (sorted && arc.ilabel > 0) {
+      break;
+      // Break from the loop: due to sorting there will be no
+      // more transitions with epsilons as input labels.
+    }
+    if (arc.ilabel != 0) {
+      continue;  // we only process epsilons here
+    }
+    Element next_elem;
+    next_elem.state = arc.nextstate;
+    next_elem.weight = Weight::Zero();
+    Weight next_unprocessed_weight = Times(unprocessed_weight, arc.weight);
+
+    // now must append strings
+    if (arc.olabel == 0) {
+      next_elem.string = str;
+    } else {
+      std::vector<Label> seq;
+      repository_->SeqOfId(str, &seq);
+      if (arc.olabel != 0) seq.push_back(arc.olabel);
+      next_elem.string = repository_->IdOfSeq(seq);
+    }
+    if (save_to_queue_2) {
+      next_elem.weight = next_unprocessed_weight;
+      queue_2_.push_back(next_elem);
+    } else {
+      AddOneElement(next_elem, next_unprocessed_weight);
+    }
+  }
+}
+
+template <class F>
+void DeterminizerStar<F>::Output(MutableFst<GallicArc<Arc> > *ofst,
+                                 bool destroy) {
+  assert(determinized_);
+  if (destroy) determinized_ = false;
+  typedef GallicWeight<Label, Weight> ThisGallicWeight;
+  typedef typename Arc::StateId StateId;
+  if (destroy) FreeMostMemory();
+  StateId nStates = static_cast<StateId>(output_arcs_.size());
+  ofst->DeleteStates();
+  ofst->SetStart(kNoStateId);
+  if (nStates == 0) {
+    return;
+  }
+  for (StateId s = 0; s < nStates; s++) {
+    OutputStateId news = ofst->AddState();
+    assert(news == s);
+  }
+  ofst->SetStart(0);
+  // now process transitions.
+  for (StateId this_state = 0; this_state < nStates; this_state++) {
+    std::vector<TempArc> &this_vec(output_arcs_[this_state]);
+    typename std::vector<TempArc>::const_iterator iter = this_vec.begin(),
+                                                  end = this_vec.end();
+    for (; iter != end; ++iter) {
+      const TempArc &temp_arc(*iter);
+      GallicArc<Arc> new_arc;
+      std::vector<Label> seq;
+      repository_.SeqOfId(temp_arc.ostring, &seq);
+      StringWeight<Label, STRING_LEFT> string_weight;
+      for (size_t i = 0; i < seq.size(); i++) string_weight.PushBack(seq[i]);
+      ThisGallicWeight gallic_weight(string_weight, temp_arc.weight);
+
+      if (temp_arc.nextstate == kNoStateId) {  // is really final weight.
+        ofst->SetFinal(this_state, gallic_weight);
+      } else {  // is really an arc.
+        new_arc.nextstate = temp_arc.nextstate;
+        new_arc.ilabel = temp_arc.ilabel;
+        new_arc.olabel = temp_arc.ilabel;  // acceptor.  input == output.
+        new_arc.weight = gallic_weight;    // includes string and weight.
+        ofst->AddArc(this_state, new_arc);
+      }
+    }
+    // Free up memory.  Do this inside the loop as ofst is also allocating
+    // memory
+    if (destroy) {
+      std::vector<TempArc> temp;
+      temp.swap(this_vec);
+    }
+  }
+  if (destroy) {
+    std::vector<std::vector<TempArc> > temp;
+    temp.swap(output_arcs_);
+  }
+}
+
+template <class F>
+void DeterminizerStar<F>::Output(MutableFst<Arc> *ofst, bool destroy) {
+  assert(determinized_);
+  if (destroy) determinized_ = false;
+  // Outputs to standard fst.
+  OutputStateId num_states = static_cast<OutputStateId>(output_arcs_.size());
+  if (destroy) FreeMostMemory();
+  ofst->DeleteStates();
+  if (num_states == 0) {
+    ofst->SetStart(kNoStateId);
+    return;
+  }
+  // Add basic states-- but will add extra ones to account for strings on
+  // output.
+  for (OutputStateId s = 0; s < num_states; s++) {
+    OutputStateId news = ofst->AddState();
+    assert(news == s);
+  }
+  ofst->SetStart(0);
+  for (OutputStateId this_state = 0; this_state < num_states; this_state++) {
+    std::vector<TempArc> &this_vec(output_arcs_[this_state]);
+
+    typename std::vector<TempArc>::const_iterator iter = this_vec.begin(),
+                                                  end = this_vec.end();
+    for (; iter != end; ++iter) {
+      const TempArc &temp_arc(*iter);
+      std::vector<Label> seq;
+      repository_.SeqOfId(temp_arc.ostring, &seq);
+      if (temp_arc.nextstate == kNoStateId) {  // Really a final weight.
+        // Make a sequence of states going to a final state, with the strings as
+        // labels. Put the weight on the first arc.
+        OutputStateId cur_state = this_state;
+        for (size_t i = 0; i < seq.size(); i++) {
+          OutputStateId next_state = ofst->AddState();
+          Arc arc;
+          arc.nextstate = next_state;
+          arc.weight = (i == 0 ? temp_arc.weight : Weight::One());
+          arc.ilabel = 0;  // epsilon.
+          arc.olabel = seq[i];
+          ofst->AddArc(cur_state, arc);
+          cur_state = next_state;
+        }
+        ofst->SetFinal(cur_state,
+                       (seq.size() == 0 ? temp_arc.weight : Weight::One()));
+      } else {  // Really an arc.
+        OutputStateId cur_state = this_state;
+        // Have to be careful with this integer comparison (i+1 < seq.size())
+        // because unsigned. i < seq.size()-1 could fail for zero-length
+        // sequences.
+        for (size_t i = 0; i + 1 < seq.size(); i++) {
+          // for all but the last element of seq, create new state.
+          OutputStateId next_state = ofst->AddState();
+          Arc arc;
+          arc.nextstate = next_state;
+          arc.weight = (i == 0 ? temp_arc.weight : Weight::One());
+          arc.ilabel = (i == 0 ? temp_arc.ilabel
+                               : 0);  // put ilabel on first element of seq.
+          arc.olabel = seq[i];
+          ofst->AddArc(cur_state, arc);
+          cur_state = next_state;
+        }
+        // Add the final arc in the sequence.
+        Arc arc;
+        arc.nextstate = temp_arc.nextstate;
+        arc.weight = (seq.size() <= 1 ? temp_arc.weight : Weight::One());
+        arc.ilabel = (seq.size() <= 1 ? temp_arc.ilabel : 0);
+        arc.olabel = (seq.size() > 0 ? seq.back() : 0);
+        ofst->AddArc(cur_state, arc);
+      }
+    }
+    // Free up memory.  Do this inside the loop as ofst is also allocating
+    // memory
+    if (destroy) {
+      std::vector<TempArc> temp;
+      temp.swap(this_vec);
+    }
+  }
+  if (destroy) {
+    std::vector<std::vector<TempArc> > temp;
+    temp.swap(output_arcs_);
+    repository_.Destroy();
+  }
+}
+
+template <class F>
+void DeterminizerStar<F>::ProcessTransition(OutputStateId state, Label ilabel,
+                                            std::vector<Element> *subset) {
+  // At input, "subset" may contain duplicates for a given dest state (but in
+  // sorted order).  This function removes duplicates from "subset", normalizes
+  // it, and adds a transition to the dest. state (possibly affecting Q_ and
+  // hash_, if state did not exist).
+
+  typedef typename std::vector<Element>::iterator IterType;
+  {  // This block makes the subset have one unique Element per state, adding
+     // the weights.
+    IterType cur_in = subset->begin(), cur_out = cur_in, end = subset->end();
+    size_t num_out = 0;
+    // Merge elements with same state-id
+    while (cur_in != end) {  // while we have more elements to process.
+      // At this point, cur_out points to location of next place we want to put
+      // an element, cur_in points to location of next element we want to
+      // process.
+      if (cur_in != cur_out) *cur_out = *cur_in;
+      cur_in++;
+      while (cur_in != end &&
+             cur_in->state == cur_out->state) {  // merge elements.
+        if (cur_in->string != cur_out->string) {
+          KALDI_ERR << "FST was not functional -> not determinizable";
+        }
+        cur_out->weight = Plus(cur_out->weight, cur_in->weight);
+        cur_in++;
+      }
+      cur_out++;
+      num_out++;
+    }
+    subset->resize(num_out);
+  }
+
+  StringId common_str;
+  Weight tot_weight;
+  {  // This block computes common_str and tot_weight (essentially: the common
+     // divisor)
+    // and removes them from the elements.
+    std::vector<Label> seq;
+
+    IterType begin = subset->begin(), iter, end = subset->end();
+    {  // This block computes "seq", which is the common prefix, and
+       // "common_str",
+      // which is the StringId version of "seq".
+      std::vector<Label> tmp_seq;
+      for (iter = begin; iter != end; ++iter) {
+        if (iter == begin) {
+          repository_.SeqOfId(iter->string, &seq);
+        } else {
+          repository_.SeqOfId(iter->string, &tmp_seq);
+          if (tmp_seq.size() < seq.size())
+            seq.resize(tmp_seq.size());  // size of shortest one.
+          for (size_t i = 0; i < seq.size();
+               i++)  // seq.size() is the shorter one at this point.
+            if (tmp_seq[i] != seq[i]) seq.resize(i);
+        }
+        if (seq.size() == 0) break;  // will not get any prefix.
+      }
+      common_str = repository_.IdOfSeq(seq);
+    }
+
+    {  // This block computes "tot_weight".
+      iter = begin;
+      tot_weight = iter->weight;
+      for (++iter; iter != end; ++iter)
+        tot_weight = Plus(tot_weight, iter->weight);
+    }
+
+    // Now divide out common stuff from elements.
+    size_t prefix_len = seq.size();
+    for (iter = begin; iter != end; ++iter) {
+      iter->weight = Divide(iter->weight, tot_weight);
+      iter->string = repository_.RemovePrefix(iter->string, prefix_len);
+    }
+  }
+
+  // Now add an arc to the state that the subset represents.
+  // We may create a new state id for this (in SubsetToStateId).
+  TempArc temp_arc;
+  temp_arc.ilabel = ilabel;
+  temp_arc.nextstate =
+      SubsetToStateId(*subset);  // may or may not really add the subset.
+  temp_arc.ostring = common_str;
+  temp_arc.weight = tot_weight;
+  output_arcs_[state].push_back(temp_arc);  // record the arc.
+}
+
+template <class F>
+void DeterminizerStar<F>::Debug() {
+  // this function called if you send a signal
+  // SIGUSR1 to the process (and it's caught by the handler in
+  // fstdeterminizestar).  It prints out some traceback
+  // info and exits.
+
+  KALDI_WARN << "Debug function called (probably SIGUSR1 caught)";
+  // free up memory from the hash as we need a little memory
+  {
+    SubsetHash hash_tmp;
+    std::swap(hash_tmp, hash_);
+  }
+
+  if (output_arcs_.size() <= 2) {
+    KALDI_ERR << "Nothing to trace back";
+  }
+  size_t max_state = output_arcs_.size() - 2;  // don't take the last
+  // one as we might be halfway into constructing it.
+
+  std::vector<OutputStateId> predecessor(max_state + 1, kNoStateId);
+  for (size_t i = 0; i < max_state; i++) {
+    for (size_t j = 0; j < output_arcs_[i].size(); j++) {
+      OutputStateId nextstate = output_arcs_[i][j].nextstate;
+      // Always find an earlier-numbered predecessor; this
+      // is always possible because of the way the algorithm
+      // works.
+      if (nextstate <= max_state && nextstate > i) predecessor[nextstate] = i;
+    }
+  }
+  std::vector<std::pair<Label, StringId> > traceback;
+  // 'traceback' is a pair of (ilabel, olabel-seq).
+  OutputStateId cur_state = max_state;  // A recently constructed state.
+
+  while (cur_state != 0 && cur_state != kNoStateId) {
+    OutputStateId last_state = predecessor[cur_state];
+    std::pair<Label, StringId> p;
+    size_t i;
+    for (i = 0; i < output_arcs_[last_state].size(); i++) {
+      if (output_arcs_[last_state][i].nextstate == cur_state) {
+        p.first = output_arcs_[last_state][i].ilabel;
+        p.second = output_arcs_[last_state][i].ostring;
+        traceback.push_back(p);
+        break;
+      }
+    }
+    KALDI_ASSERT(i != output_arcs_[last_state].size());  // Or fell off loop.
+    cur_state = last_state;
+  }
+  if (cur_state == kNoStateId)
+    KALDI_WARN << "Traceback did not reach start state "
+               << "(possibly debug-code error)";
+
+  std::stringstream ss;
+  ss << "Traceback follows in format "
+     << "ilabel (olabel olabel) ilabel (olabel) ... :";
+  for (ssize_t i = traceback.size() - 1; i >= 0; i--) {
+    ss << ' ' << traceback[i].first << " ( ";
+    std::vector<Label> seq;
+    repository_.SeqOfId(traceback[i].second, &seq);
+    for (size_t j = 0; j < seq.size(); j++) ss << seq[j] << ' ';
+    ss << ')';
+  }
+  KALDI_ERR << ss.str();
+}
+
+}  // namespace fst
+
+#endif  // KALDI_FSTEXT_DETERMINIZE_STAR_INL_H_
diff --git a/speechx/speechx/kaldi/fstext/determinize-star.h b/speechx/speechx/kaldi/fstext/determinize-star.h
new file mode 100644
index 00000000..ec4b2369
--- /dev/null
+++ b/speechx/speechx/kaldi/fstext/determinize-star.h
@@ -0,0 +1,116 @@
+// fstext/determinize-star.h
+
+// Copyright 2009-2011  Microsoft Corporation
+//                2014  Guoguo Chen
+//                2015  Hainan Xu
+
+// See ../../COPYING for clarification regarding multiple authors
+//
+// Licensed under the Apache License, Version 2.0 (the "License");
+// you may not use this file except in compliance with the License.
+// You may obtain a copy of the License at
+//
+//  http://www.apache.org/licenses/LICENSE-2.0
+//
+// THIS CODE IS PROVIDED *AS IS* BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY
+// KIND, EITHER EXPRESS OR IMPLIED, INCLUDING WITHOUT LIMITATION ANY IMPLIED
+// WARRANTIES OR CONDITIONS OF TITLE, FITNESS FOR A PARTICULAR PURPOSE,
+// MERCHANTABLITY OR NON-INFRINGEMENT.
+// See the Apache 2 License for the specific language governing permissions and
+// limitations under the License.
+
+#ifndef KALDI_FSTEXT_DETERMINIZE_STAR_H_
+#define KALDI_FSTEXT_DETERMINIZE_STAR_H_
+#include <fst/fst-decl.h>
+#include <fst/fstlib.h>
+#include <algorithm>
+#include <map>
+#include <set>
+#include <stdexcept>  // this algorithm uses exceptions
+#include <vector>
+
+namespace fst {
+
+/// \addtogroup fst_extensions
+///  @{
+
+// For example of usage, see test-determinize-star.cc
+
+/*
+   DeterminizeStar implements determinization with epsilon removal, which we
+   distinguish with a star.
+
+   We define a determinized* FST as one in which no state has more than one
+   transition with the same input-label.  Epsilon input labels are not allowed
+   except starting from states that have exactly one arc exiting them (and are
+   not final).  [In the normal definition of determinized, epsilon-input labels
+   are not allowed at all, whereas in Mohri's definition, epsilons are treated
+   as ordinary symbols].  The determinized* definition is intended to simulate
+   the effect of allowing strings of output symbols at each state.
+
+   The algorithm implemented here takes an Fst<Arc>, and a pointer to a
+   MutableFst<Arc> where it puts its output.  The weight type is assumed to be a
+   float-weight.  It does epsilon removal and determinization.
+   This algorithm may fail if the input has epsilon cycles under
+   certain circumstances (i.e. the semiring is non-idempotent, e.g. the log
+   semiring, or there are negative cost epsilon cycles).
+
+   This implementation is much less fancy than the one in fst/determinize.h, and
+   does not have an "on-demand" version.
+
+   The algorithm is a fairly normal determinization algorithm.  We keep in
+   memory the subsets of states, together with their leftover strings and their
+   weights.  The only difference is we detect input epsilon transitions and
+   treat them "specially".
+*/
+
+// This algorithm will be slightly faster if you sort the input fst on input
+// label.
+
+/**
+    This function implements the normal version of DeterminizeStar, in which the
+    output strings are represented using sequences of arcs, where all but the
+    first one has an epsilon on the input side.  The debug_ptr argument is an
+    optional pointer to a bool that, if it becomes true while the algorithm is
+    executing, the algorithm will print a traceback and terminate (used in
+    fstdeterminizestar.cc debug non-terminating determinization).
+    If max_states is positive, it will stop determinization and throw an
+    exception as soon as the max-states is reached. This can be useful in test.
+    If allow_partial is true, the algorithm will output partial results when the
+    specified max_states is reached (when larger than zero), instead of throwing
+    out an error.
+
+    Caution, the return status is un-intuitive: this function will return false
+   if determinization completed normally, and true if it was stopped early by
+    reaching the 'max-states' limit, and a partial FST was generated.
+*/
+template <class F>
+bool DeterminizeStar(F &ifst, MutableFst<typename F::Arc> *ofst,  // NOLINT
+                     float delta = kDelta, bool *debug_ptr = NULL,
+                     int max_states = -1, bool allow_partial = false);
+
+/*  This is a version of DeterminizeStar with a slightly more "natural" output
+   format, where the output sequences are encoded using the GallicArc (i.e. the
+   output symbols are strings. If max_states is positive, it will stop
+   determinization and throw an exception as soon as the max-states is reached.
+   This can be useful in test. If allow_partial is true, the algorithm will
+   output partial results when the specified max_states is reached (when larger
+   than zero), instead of throwing out an error.
+
+    Caution, the return status is un-intuitive: this function will return false
+   if determinization completed normally, and true if it was stopped early by
+    reaching the 'max-states' limit, and a partial FST was generated.
+*/
+template <class F>
+bool DeterminizeStar(F &ifst,  // NOLINT
+                     MutableFst<GallicArc<typename F::Arc> > *ofst,
+                     float delta = kDelta, bool *debug_ptr = NULL,
+                     int max_states = -1, bool allow_partial = false);
+
+/// @} end "addtogroup fst_extensions"
+
+}  // end namespace fst
+
+#include "fstext/determinize-star-inl.h"
+
+#endif  // KALDI_FSTEXT_DETERMINIZE_STAR_H_
diff --git a/speechx/speechx/kaldi/fstext/fstext-lib.h b/speechx/speechx/kaldi/fstext/fstext-lib.h
new file mode 100644
index 00000000..fa27488b
--- /dev/null
+++ b/speechx/speechx/kaldi/fstext/fstext-lib.h
@@ -0,0 +1,34 @@
+// fstext/fstext-lib.h
+
+// Copyright 2009-2012  Microsoft Corporation  Johns Hopkins University (author:
+// Daniel Povey)
+
+// See ../../COPYING for clarification regarding multiple authors
+//
+// Licensed under the Apache License, Version 2.0 (the "License");
+// you may not use this file except in compliance with the License.
+// You may obtain a copy of the License at
+//
+//  http://www.apache.org/licenses/LICENSE-2.0
+//
+// THIS CODE IS PROVIDED *AS IS* BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY
+// KIND, EITHER EXPRESS OR IMPLIED, INCLUDING WITHOUT LIMITATION ANY IMPLIED
+// WARRANTIES OR CONDITIONS OF TITLE, FITNESS FOR A PARTICULAR PURPOSE,
+// MERCHANTABLITY OR NON-INFRINGEMENT.
+// See the Apache 2 License for the specific language governing permissions and
+// limitations under the License.
+
+#ifndef KALDI_FSTEXT_FSTEXT_LIB_H_
+#define KALDI_FSTEXT_FSTEXT_LIB_H_
+
+#include "fst/fstlib.h"
+#include "fstext/determinize-lattice.h"
+#include "fstext/determinize-star.h"
+#include "fstext/fstext-utils.h"
+#include "fstext/kaldi-fst-io.h"
+#include "fstext/lattice-utils.h"
+#include "fstext/lattice-weight.h"
+#include "fstext/pre-determinize.h"
+#include "fstext/table-matcher.h"
+
+#endif  // KALDI_FSTEXT_FSTEXT_LIB_H_
diff --git a/speechx/speechx/kaldi/fstext/fstext-utils-inl.h b/speechx/speechx/kaldi/fstext/fstext-utils-inl.h
new file mode 100644
index 00000000..d888bf98
--- /dev/null
+++ b/speechx/speechx/kaldi/fstext/fstext-utils-inl.h
@@ -0,0 +1,1265 @@
+// fstext/fstext-utils-inl.h
+
+// Copyright 2009-2012  Microsoft Corporation  Johns Hopkins University (Author:
+// Daniel Povey)
+//                2014  Telepoint Global Hosting Service, LLC. (Author: David
+//                Snyder)
+
+// See ../../COPYING for clarification regarding multiple authors
+//
+// Licensed under the Apache License, Version 2.0 (the "License");
+// you may not use this file except in compliance with the License.
+// You may obtain a copy of the License at
+//
+//  http://www.apache.org/licenses/LICENSE-2.0
+//
+// THIS CODE IS PROVIDED *AS IS* BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY
+// KIND, EITHER EXPRESS OR IMPLIED, INCLUDING WITHOUT LIMITATION ANY IMPLIED
+// WARRANTIES OR CONDITIONS OF TITLE, FITNESS FOR A PARTICULAR PURPOSE,
+// MERCHANTABLITY OR NON-INFRINGEMENT.
+// See the Apache 2 License for the specific language governing permissions and
+// limitations under the License.
+
+#ifndef KALDI_FSTEXT_FSTEXT_UTILS_INL_H_
+#define KALDI_FSTEXT_FSTEXT_UTILS_INL_H_
+
+#include <algorithm>
+#include <cstring>
+#include <map>
+#include <set>
+#include <sstream>
+#include <string>
+#include <unordered_map>
+#include <unordered_set>
+#include <utility>
+#include <vector>
+
+#include "base/kaldi-common.h"
+#include "fstext/determinize-star.h"
+#include "fstext/pre-determinize.h"
+#include "util/const-integer-set.h"
+#include "util/kaldi-io.h"
+#include "util/stl-utils.h"
+#include "util/text-utils.h"
+
+namespace fst {
+
+template <class Arc>
+typename Arc::Label HighestNumberedOutputSymbol(const Fst<Arc> &fst) {
+  typename Arc::Label ans = 0;
+  for (StateIterator<Fst<Arc> > siter(fst); !siter.Done(); siter.Next()) {
+    typename Arc::StateId s = siter.Value();
+    for (ArcIterator<Fst<Arc> > aiter(fst, s); !aiter.Done(); aiter.Next()) {
+      const Arc &arc = aiter.Value();
+      ans = std::max(ans, arc.olabel);
+    }
+  }
+  return ans;
+}
+
+template <class Arc>
+typename Arc::Label HighestNumberedInputSymbol(const Fst<Arc> &fst) {
+  typename Arc::Label ans = 0;
+  for (StateIterator<Fst<Arc> > siter(fst); !siter.Done(); siter.Next()) {
+    typename Arc::StateId s = siter.Value();
+    for (ArcIterator<Fst<Arc> > aiter(fst, s); !aiter.Done(); aiter.Next()) {
+      const Arc &arc = aiter.Value();
+      ans = std::max(ans, arc.ilabel);
+    }
+  }
+  return ans;
+}
+
+template <class Arc>
+typename Arc::StateId NumArcs(const ExpandedFst<Arc> &fst) {
+  typedef typename Arc::StateId StateId;
+  StateId num_arcs = 0;
+  for (StateId s = 0; s < fst.NumStates(); s++) num_arcs += fst.NumArcs(s);
+  return num_arcs;
+}
+
+template <class Arc, class I>
+void GetOutputSymbols(const Fst<Arc> &fst, bool include_eps,
+                      std::vector<I> *symbols) {
+  KALDI_ASSERT_IS_INTEGER_TYPE(I);
+  std::set<I> all_syms;
+  for (StateIterator<Fst<Arc> > siter(fst); !siter.Done(); siter.Next()) {
+    typename Arc::StateId s = siter.Value();
+    for (ArcIterator<Fst<Arc> > aiter(fst, s); !aiter.Done(); aiter.Next()) {
+      const Arc &arc = aiter.Value();
+      all_syms.insert(arc.olabel);
+    }
+  }
+
+  // Remove epsilon, if instructed.
+  if (!include_eps && !all_syms.empty() && *all_syms.begin() == 0)
+    all_syms.erase(0);
+  KALDI_ASSERT(symbols != NULL);
+  kaldi::CopySetToVector(all_syms, symbols);
+}
+
+template <class Arc, class I>
+void GetInputSymbols(const Fst<Arc> &fst, bool include_eps,
+                     std::vector<I> *symbols) {
+  KALDI_ASSERT_IS_INTEGER_TYPE(I);
+  unordered_set<I> all_syms;
+  for (StateIterator<Fst<Arc> > siter(fst); !siter.Done(); siter.Next()) {
+    typename Arc::StateId s = siter.Value();
+    for (ArcIterator<Fst<Arc> > aiter(fst, s); !aiter.Done(); aiter.Next()) {
+      const Arc &arc = aiter.Value();
+      all_syms.insert(arc.ilabel);
+    }
+  }
+  // Remove epsilon, if instructed.
+  if (!include_eps && all_syms.count(0) != 0) all_syms.erase(0);
+  KALDI_ASSERT(symbols != NULL);
+  kaldi::CopySetToVector(all_syms, symbols);
+  std::sort(symbols->begin(), symbols->end());
+}
+
+template <class Arc, class I>
+class RemoveSomeInputSymbolsMapper {
+ public:
+  Arc operator()(const Arc &arc_in) {
+    Arc ans = arc_in;
+    if (to_remove_set_.count(ans.ilabel) != 0)
+      ans.ilabel = 0;  // remove this symbol
+    return ans;
+  }
+  MapFinalAction FinalAction() { return MAP_NO_SUPERFINAL; }
+  MapSymbolsAction InputSymbolsAction() { return MAP_CLEAR_SYMBOLS; }
+  MapSymbolsAction OutputSymbolsAction() { return MAP_COPY_SYMBOLS; }
+  uint64 Properties(uint64 props) const {
+    // remove the following as we don't know now if any of them are true.
+    uint64 to_remove = kAcceptor | kNotAcceptor | kIDeterministic |
+                       kNonIDeterministic | kNoEpsilons | kNoIEpsilons |
+                       kILabelSorted | kNotILabelSorted;
+    return props & ~to_remove;
+  }
+  explicit RemoveSomeInputSymbolsMapper(const std::vector<I> &to_remove)
+      : to_remove_set_(to_remove) {
+    KALDI_ASSERT_IS_INTEGER_TYPE(I);
+    assert(to_remove_set_.count(0) == 0);  // makes no sense to remove epsilon.
+  }
+
+ private:
+  kaldi::ConstIntegerSet<I> to_remove_set_;
+};
+
+template <class Arc, class I>
+using LookaheadFst = ArcMapFst<Arc, Arc, RemoveSomeInputSymbolsMapper<Arc, I> >;
+
+// Lookahead composition is used for optimized online
+// composition of FSTs during decoding. See
+// nnet3/nnet3-latgen-faster-lookahead.cc. For details of compose filters
+// see DefaultLookAhead in fst/compose.h
+template <class Arc, class I>
+LookaheadFst<Arc, I> *LookaheadComposeFst(const Fst<Arc> &ifst1,
+                                          const Fst<Arc> &ifst2,
+                                          const std::vector<I> &to_remove) {
+  fst::CacheOptions cache_opts(true, 1 << 25LL);
+  fst::CacheOptions cache_opts_map(true, 0);
+  fst::ArcMapFstOptions arcmap_opts(cache_opts);
+  RemoveSomeInputSymbolsMapper<Arc, I> mapper(to_remove);
+  return new LookaheadFst<Arc, I>(ComposeFst<Arc>(ifst1, ifst2, cache_opts),
+                                  mapper, arcmap_opts);
+}
+
+template <class Arc, class I>
+void RemoveSomeInputSymbols(const std::vector<I> &to_remove,
+                            MutableFst<Arc> *fst) {
+  KALDI_ASSERT_IS_INTEGER_TYPE(I);
+  RemoveSomeInputSymbolsMapper<Arc, I> mapper(to_remove);
+  Map(fst, mapper);
+}
+
+template <class Arc, class I>
+class MapInputSymbolsMapper {
+ public:
+  Arc operator()(const Arc &arc_in) {
+    Arc ans = arc_in;
+    if (ans.ilabel > 0 && ans.ilabel < static_cast<typename Arc::Label>(
+                                           (*symbol_mapping_).size()))
+      ans.ilabel = (*symbol_mapping_)[ans.ilabel];
+    return ans;
+  }
+  MapFinalAction FinalAction() const { return MAP_NO_SUPERFINAL; }
+  MapSymbolsAction InputSymbolsAction() const { return MAP_CLEAR_SYMBOLS; }
+  MapSymbolsAction OutputSymbolsAction() const { return MAP_COPY_SYMBOLS; }
+  uint64 Properties(uint64 props) const {  // Not tested.
+    bool remove_epsilons =
+        (symbol_mapping_->size() > 0 && (*symbol_mapping_)[0] != 0);
+    bool add_epsilons = (symbol_mapping_->size() > 1 &&
+                         *std::min_element(symbol_mapping_->begin() + 1,
+                                           symbol_mapping_->end()) == 0);
+
+    // remove the following as we don't know now if any of them are true.
+    uint64 props_to_remove = kAcceptor | kNotAcceptor | kIDeterministic |
+                             kNonIDeterministic | kILabelSorted |
+                             kNotILabelSorted;
+    if (remove_epsilons) props_to_remove |= kEpsilons | kIEpsilons;
+    if (add_epsilons) props_to_remove |= kNoEpsilons | kNoIEpsilons;
+    uint64 props_to_add = 0;
+    if (remove_epsilons && !add_epsilons)
+      props_to_add |= kNoEpsilons | kNoIEpsilons;
+    return (props & ~props_to_remove) | props_to_add;
+  }
+  // initialize with copy = false only if the "to_remove" argument will not be
+  // deleted in the lifetime of this object.
+  MapInputSymbolsMapper(const std::vector<I> &to_remove, bool copy) {
+    KALDI_ASSERT_IS_INTEGER_TYPE(I);
+    if (copy)
+      symbol_mapping_ = new std::vector<I>(to_remove);
+    else
+      symbol_mapping_ = &to_remove;
+    owned = copy;
+  }
+  ~MapInputSymbolsMapper() {
+    if (owned && symbol_mapping_ != NULL) delete symbol_mapping_;
+  }
+
+ private:
+  bool owned;
+  const std::vector<I> *symbol_mapping_;
+};
+
+template <class Arc, class I>
+void MapInputSymbols(const std::vector<I> &symbol_mapping,
+                     MutableFst<Arc> *fst) {
+  KALDI_ASSERT_IS_INTEGER_TYPE(I);
+  // false == don't copy the "symbol_mapping", retain pointer--
+  // safe since short-lived object.
+  MapInputSymbolsMapper<Arc, I> mapper(symbol_mapping, false);
+  Map(fst, mapper);
+}
+
+template <class Arc, class I>
+bool GetLinearSymbolSequence(const Fst<Arc> &fst, std::vector<I> *isymbols_out,
+                             std::vector<I> *osymbols_out,
+                             typename Arc::Weight *tot_weight_out) {
+  typedef typename Arc::StateId StateId;
+  typedef typename Arc::Weight Weight;
+
+  Weight tot_weight = Weight::One();
+  std::vector<I> ilabel_seq;
+  std::vector<I> olabel_seq;
+
+  StateId cur_state = fst.Start();
+  if (cur_state == kNoStateId) {  // empty sequence.
+    if (isymbols_out != NULL) isymbols_out->clear();
+    if (osymbols_out != NULL) osymbols_out->clear();
+    if (tot_weight_out != NULL) *tot_weight_out = Weight::Zero();
+    return true;
+  }
+  while (1) {
+    Weight w = fst.Final(cur_state);
+    if (w != Weight::Zero()) {  // is final..
+      tot_weight = Times(w, tot_weight);
+      if (fst.NumArcs(cur_state) != 0) return false;
+      if (isymbols_out != NULL) *isymbols_out = ilabel_seq;
+      if (osymbols_out != NULL) *osymbols_out = olabel_seq;
+      if (tot_weight_out != NULL) *tot_weight_out = tot_weight;
+      return true;
+    } else {
+      if (fst.NumArcs(cur_state) != 1) return false;
+
+      ArcIterator<Fst<Arc> > iter(fst, cur_state);  // get the only arc.
+      const Arc &arc = iter.Value();
+      tot_weight = Times(arc.weight, tot_weight);
+      if (arc.ilabel != 0) ilabel_seq.push_back(arc.ilabel);
+      if (arc.olabel != 0) olabel_seq.push_back(arc.olabel);
+      cur_state = arc.nextstate;
+    }
+  }
+}
+
+// see fstext-utils.h for comment.
+template <class Arc>
+void ConvertNbestToVector(const Fst<Arc> &fst,
+                          std::vector<VectorFst<Arc> > *fsts_out) {
+  typedef typename Arc::Weight Weight;
+  typedef typename Arc::StateId StateId;
+  fsts_out->clear();
+  StateId start_state = fst.Start();
+  if (start_state == kNoStateId) return;  // No output.
+  size_t n_arcs = fst.NumArcs(start_state);
+  bool start_is_final = (fst.Final(start_state) != Weight::Zero());
+  fsts_out->reserve(n_arcs + (start_is_final ? 1 : 0));
+
+  if (start_is_final) {
+    fsts_out->resize(fsts_out->size() + 1);
+    StateId start_state_out = fsts_out->back().AddState();
+    fsts_out->back().SetFinal(start_state_out, fst.Final(start_state));
+  }
+
+  for (ArcIterator<Fst<Arc> > start_aiter(fst, start_state);
+       !start_aiter.Done(); start_aiter.Next()) {
+    fsts_out->resize(fsts_out->size() + 1);
+    VectorFst<Arc> &ofst = fsts_out->back();
+    const Arc &first_arc = start_aiter.Value();
+    StateId cur_state = start_state, cur_ostate = ofst.AddState();
+    ofst.SetStart(cur_ostate);
+    StateId next_ostate = ofst.AddState();
+    ofst.AddArc(cur_ostate, Arc(first_arc.ilabel, first_arc.olabel,
+                                first_arc.weight, next_ostate));
+    cur_state = first_arc.nextstate;
+    cur_ostate = next_ostate;
+    while (1) {
+      size_t this_n_arcs = fst.NumArcs(cur_state);
+      KALDI_ASSERT(this_n_arcs <= 1);  // or it violates our assumptions
+                                       // about the input.
+      if (this_n_arcs == 1) {
+        KALDI_ASSERT(fst.Final(cur_state) == Weight::Zero());
+        // or problem with ShortestPath.
+        ArcIterator<Fst<Arc> > aiter(fst, cur_state);
+        const Arc &arc = aiter.Value();
+        next_ostate = ofst.AddState();
+        ofst.AddArc(cur_ostate,
+                    Arc(arc.ilabel, arc.olabel, arc.weight, next_ostate));
+        cur_state = arc.nextstate;
+        cur_ostate = next_ostate;
+      } else {
+        KALDI_ASSERT(fst.Final(cur_state) != Weight::Zero());
+        // or problem with ShortestPath.
+        ofst.SetFinal(cur_ostate, fst.Final(cur_state));
+        break;
+      }
+    }
+  }
+}
+
+// see fstext-utils.sh for comment.
+template <class Arc>
+void NbestAsFsts(const Fst<Arc> &fst, size_t n,
+                 std::vector<VectorFst<Arc> > *fsts_out) {
+  KALDI_ASSERT(n > 0);
+  KALDI_ASSERT(fsts_out != NULL);
+  VectorFst<Arc> nbest_fst;
+  ShortestPath(fst, &nbest_fst, n);
+  ConvertNbestToVector(nbest_fst, fsts_out);
+}
+
+template <class Arc, class I>
+void MakeLinearAcceptorWithAlternatives(
+    const std::vector<std::vector<I> > &labels, MutableFst<Arc> *ofst) {
+  typedef typename Arc::StateId StateId;
+  typedef typename Arc::Weight Weight;
+
+  ofst->DeleteStates();
+  StateId cur_state = ofst->AddState();
+  ofst->SetStart(cur_state);
+  for (size_t i = 0; i < labels.size(); i++) {
+    KALDI_ASSERT(labels[i].size() != 0);
+    StateId next_state = ofst->AddState();
+    for (size_t j = 0; j < labels[i].size(); j++) {
+      Arc arc(labels[i][j], labels[i][j], Weight::One(), next_state);
+      ofst->AddArc(cur_state, arc);
+    }
+    cur_state = next_state;
+  }
+  ofst->SetFinal(cur_state, Weight::One());
+}
+
+template <class Arc, class I>
+void MakeLinearAcceptor(const std::vector<I> &labels, MutableFst<Arc> *ofst) {
+  typedef typename Arc::StateId StateId;
+  typedef typename Arc::Weight Weight;
+
+  ofst->DeleteStates();
+  StateId cur_state = ofst->AddState();
+  ofst->SetStart(cur_state);
+  for (size_t i = 0; i < labels.size(); i++) {
+    StateId next_state = ofst->AddState();
+    Arc arc(labels[i], labels[i], Weight::One(), next_state);
+    ofst->AddArc(cur_state, arc);
+    cur_state = next_state;
+  }
+  ofst->SetFinal(cur_state, Weight::One());
+}
+
+template <class I>
+void GetSymbols(const SymbolTable &symtab, bool include_eps,
+                std::vector<I> *syms_out) {
+  KALDI_ASSERT(syms_out != NULL);
+  syms_out->clear();
+  for (SymbolTableIterator iter(symtab); !iter.Done(); iter.Next()) {
+    if (include_eps || iter.Value() != 0) {
+      syms_out->push_back(iter.Value());
+      KALDI_ASSERT(syms_out->back() ==
+                   iter.Value());  // an integer-range thing.
+    }
+  }
+}
+
+template <class Arc>
+void SafeDeterminizeWrapper(MutableFst<Arc> *ifst, MutableFst<Arc> *ofst,
+                            float delta) {
+  typename Arc::Label highest_sym = HighestNumberedInputSymbol(*ifst);
+  std::vector<typename Arc::Label> extra_syms;
+  PreDeterminize(ifst, (typename Arc::Label)(highest_sym + 1), &extra_syms);
+  DeterminizeStar(*ifst, ofst, delta);
+  RemoveSomeInputSymbols(extra_syms, ofst);  // remove the extra symbols.
+}
+
+template <class Arc>
+void SafeDeterminizeMinimizeWrapper(MutableFst<Arc> *ifst, VectorFst<Arc> *ofst,
+                                    float delta) {
+  typename Arc::Label highest_sym = HighestNumberedInputSymbol(*ifst);
+  std::vector<typename Arc::Label> extra_syms;
+  PreDeterminize(ifst, (typename Arc::Label)(highest_sym + 1), &extra_syms);
+  DeterminizeStar(*ifst, ofst, delta);
+  RemoveSomeInputSymbols(extra_syms, ofst);  // remove the extra symbols.
+  RemoveEpsLocal(ofst);  // this is "safe" and will never hurt.
+  MinimizeEncoded(ofst, delta);
+}
+
+inline void DeterminizeStarInLog(VectorFst<StdArc> *fst, float delta,
+                                 bool *debug_ptr, int max_states) {
+  // DeterminizeStarInLog determinizes 'fst' in the log semiring, using
+  // the DeterminizeStar algorithm (which also removes epsilons).
+
+  ArcSort(fst, ILabelCompare<StdArc>());  // helps DeterminizeStar to be faster.
+  VectorFst<LogArc> *fst_log =
+      new VectorFst<LogArc>;  // Want to determinize in log semiring.
+  Cast(*fst, fst_log);
+  VectorFst<StdArc> tmp;
+  *fst = tmp;  // make fst empty to free up memory. [actually may make no
+               // difference..]
+  VectorFst<LogArc> *fst_det_log = new VectorFst<LogArc>;
+  DeterminizeStar(*fst_log, fst_det_log, delta, debug_ptr, max_states);
+  Cast(*fst_det_log, fst);
+  delete fst_log;
+  delete fst_det_log;
+}
+
+inline void DeterminizeInLog(VectorFst<StdArc> *fst) {
+  // DeterminizeInLog determinizes 'fst' in the log semiring.
+
+  ArcSort(fst, ILabelCompare<StdArc>());  // helps DeterminizeStar to be faster.
+  VectorFst<LogArc> *fst_log =
+      new VectorFst<LogArc>;  // Want to determinize in log semiring.
+  Cast(*fst, fst_log);
+  VectorFst<StdArc> tmp;
+  *fst = tmp;  // make fst empty to free up memory. [actually may make no
+               // difference..]
+  VectorFst<LogArc> *fst_det_log = new VectorFst<LogArc>;
+  Determinize(*fst_log, fst_det_log);
+  Cast(*fst_det_log, fst);
+  delete fst_log;
+  delete fst_det_log;
+}
+
+// make it inline to avoid having to put it in a .cc file.
+// destructive algorithm (changes ifst as well as ofst).
+inline void SafeDeterminizeMinimizeWrapperInLog(VectorFst<StdArc> *ifst,
+                                                VectorFst<StdArc> *ofst,
+                                                float delta) {
+  VectorFst<LogArc> *ifst_log =
+      new VectorFst<LogArc>;  // Want to determinize in log semiring.
+  Cast(*ifst, ifst_log);
+  VectorFst<LogArc> *ofst_log = new VectorFst<LogArc>;
+  SafeDeterminizeWrapper(ifst_log, ofst_log, delta);
+  Cast(*ofst_log, ofst);
+  delete ifst_log;
+  delete ofst_log;
+  RemoveEpsLocal(ofst);  // this is "safe" and will never hurt.  Do this in
+                         // tropical, which is important.
+  MinimizeEncoded(ofst, delta);  // Non-deterministic minimization will fail in
+                                 // log semiring so do it with StdARc.
+}
+
+inline void SafeDeterminizeWrapperInLog(VectorFst<StdArc> *ifst,
+                                        VectorFst<StdArc> *ofst, float delta) {
+  VectorFst<LogArc> *ifst_log =
+      new VectorFst<LogArc>;  // Want to determinize in log semiring.
+  Cast(*ifst, ifst_log);
+  VectorFst<LogArc> *ofst_log = new VectorFst<LogArc>;
+  SafeDeterminizeWrapper(ifst_log, ofst_log, delta);
+  Cast(*ofst_log, ofst);
+  delete ifst_log;
+  delete ofst_log;
+}
+
+template <class Arc>
+void RemoveWeights(MutableFst<Arc> *ifst) {
+  typedef typename Arc::StateId StateId;
+  typedef typename Arc::Weight Weight;
+
+  for (StateIterator<MutableFst<Arc> > siter(*ifst); !siter.Done();
+       siter.Next()) {
+    StateId s = siter.Value();
+    for (MutableArcIterator<MutableFst<Arc> > aiter(ifst, s); !aiter.Done();
+         aiter.Next()) {
+      Arc arc(aiter.Value());
+      arc.weight = Weight::One();
+      aiter.SetValue(arc);
+    }
+    if (ifst->Final(s) != Weight::Zero()) ifst->SetFinal(s, Weight::One());
+  }
+  ifst->SetProperties(kUnweighted, kUnweighted);
+}
+
+// Used in PrecedingInputSymbolsAreSame (non-functor version), and
+// similar routines.
+template <class T>
+struct IdentityFunction {
+  typedef T Arg;
+  typedef T Result;
+  T operator()(const T &t) const { return t; }
+};
+
+template <class Arc>
+bool PrecedingInputSymbolsAreSame(bool start_is_epsilon, const Fst<Arc> &fst) {
+  IdentityFunction<typename Arc::Label> f;
+  return PrecedingInputSymbolsAreSameClass(start_is_epsilon, fst, f);
+}
+
+template <class Arc, class F>  // F is functor type from labels to classes.
+bool PrecedingInputSymbolsAreSameClass(bool start_is_epsilon,
+                                       const Fst<Arc> &fst, const F &f) {
+  typedef typename F::Result ClassType;
+  typedef typename Arc::StateId StateId;
+  std::vector<ClassType> classes;
+  ClassType noClass = f(kNoLabel);
+
+  if (start_is_epsilon) {
+    StateId start_state = fst.Start();
+    if (start_state < 0 || start_state == kNoStateId)
+      return true;  // empty fst-- doesn't matter.
+    classes.resize(start_state + 1, noClass);
+    classes[start_state] = 0;
+  }
+
+  for (StateIterator<Fst<Arc> > siter(fst); !siter.Done(); siter.Next()) {
+    StateId s = siter.Value();
+    for (ArcIterator<Fst<Arc> > aiter(fst, s); !aiter.Done(); aiter.Next()) {
+      const Arc &arc = aiter.Value();
+      if (classes.size() <= arc.nextstate)
+        classes.resize(arc.nextstate + 1, noClass);
+      if (classes[arc.nextstate] == noClass)
+        classes[arc.nextstate] = f(arc.ilabel);
+      else if (classes[arc.nextstate] != f(arc.ilabel))
+        return false;
+    }
+  }
+  return true;
+}
+
+template <class Arc>
+bool FollowingInputSymbolsAreSame(bool end_is_epsilon, const Fst<Arc> &fst) {
+  IdentityFunction<typename Arc::Label> f;
+  return FollowingInputSymbolsAreSameClass(end_is_epsilon, fst, f);
+}
+
+template <class Arc, class F>
+bool FollowingInputSymbolsAreSameClass(bool end_is_epsilon, const Fst<Arc> &fst,
+                                       const F &f) {
+  typedef typename Arc::StateId StateId;
+  typedef typename Arc::Weight Weight;
+  typedef typename F::Result ClassType;
+  const ClassType noClass = f(kNoLabel), epsClass = f(0);
+  for (StateIterator<Fst<Arc> > siter(fst); !siter.Done(); siter.Next()) {
+    StateId s = siter.Value();
+    ClassType c = noClass;
+    for (ArcIterator<Fst<Arc> > aiter(fst, s); !aiter.Done(); aiter.Next()) {
+      const Arc &arc = aiter.Value();
+      if (c == noClass)
+        c = f(arc.ilabel);
+      else if (c != f(arc.ilabel))
+        return false;
+    }
+    if (end_is_epsilon && c != noClass && c != epsClass &&
+        fst.Final(s) != Weight::Zero())
+      return false;
+  }
+  return true;
+}
+
+template <class Arc>
+void MakePrecedingInputSymbolsSame(bool start_is_epsilon,
+                                   MutableFst<Arc> *fst) {
+  IdentityFunction<typename Arc::Label> f;
+  MakePrecedingInputSymbolsSameClass(start_is_epsilon, fst, f);
+}
+
+template <class Arc, class F>
+void MakePrecedingInputSymbolsSameClass(bool start_is_epsilon,
+                                        MutableFst<Arc> *fst, const F &f) {
+  typedef typename F::Result ClassType;
+  typedef typename Arc::StateId StateId;
+  typedef typename Arc::Weight Weight;
+  std::vector<ClassType> classes;
+  ClassType noClass = f(kNoLabel);
+  ClassType epsClass = f(0);
+  if (start_is_epsilon) {  // treat having-start-state as epsilon in-transition.
+    StateId start_state = fst->Start();
+    if (start_state < 0 || start_state == kNoStateId)  // empty FST.
+      return;
+    classes.resize(start_state + 1, noClass);
+    classes[start_state] = epsClass;
+  }
+
+  // Find bad states (states with multiple input-symbols into them).
+  std::set<StateId> bad_states;  // states that we need to change.
+  for (StateIterator<Fst<Arc> > siter(*fst); !siter.Done(); siter.Next()) {
+    StateId s = siter.Value();
+    for (ArcIterator<Fst<Arc> > aiter(*fst, s); !aiter.Done(); aiter.Next()) {
+      const Arc &arc = aiter.Value();
+      if (classes.size() <= static_cast<size_t>(arc.nextstate))
+        classes.resize(arc.nextstate + 1, noClass);
+      if (classes[arc.nextstate] == noClass)
+        classes[arc.nextstate] = f(arc.ilabel);
+      else if (classes[arc.nextstate] != f(arc.ilabel))
+        bad_states.insert(arc.nextstate);
+    }
+  }
+  if (bad_states.empty()) return;  // Nothing to do.
+  kaldi::ConstIntegerSet<StateId> bad_states_ciset(
+      bad_states);  // faster lookup.
+
+  // Work out list of arcs we have to change as (state, arc-offset).
+  // Can't do the actual changes in this pass, since we have to add new
+  // states which invalidates the iterators.
+  std::vector<std::pair<StateId, size_t> > arcs_to_change;
+  for (StateIterator<Fst<Arc> > siter(*fst); !siter.Done(); siter.Next()) {
+    StateId s = siter.Value();
+    for (ArcIterator<Fst<Arc> > aiter(*fst, s); !aiter.Done(); aiter.Next()) {
+      const Arc &arc = aiter.Value();
+      if (arc.ilabel != 0 && bad_states_ciset.count(arc.nextstate) != 0)
+        arcs_to_change.push_back(std::make_pair(s, aiter.Position()));
+    }
+  }
+  KALDI_ASSERT(!arcs_to_change.empty());  // since !bad_states.empty().
+
+  std::map<std::pair<StateId, ClassType>, StateId> state_map;
+  // state_map is a map from (bad-state, input-symbol-class) to dummy-state.
+
+  for (size_t i = 0; i < arcs_to_change.size(); i++) {
+    StateId s = arcs_to_change[i].first;
+    ArcIterator<MutableFst<Arc> > aiter(*fst, s);
+    aiter.Seek(arcs_to_change[i].second);
+    Arc arc = aiter.Value();
+
+    // Transition is non-eps transition to "bad" state.  Introduce new state (or
+    // find existing one).
+    std::pair<StateId, ClassType> p(arc.nextstate, f(arc.ilabel));
+    if (state_map.count(p) == 0) {
+      StateId newstate = state_map[p] = fst->AddState();
+      fst->AddArc(newstate, Arc(0, 0, Weight::One(), arc.nextstate));
+    }
+    StateId dst_state = state_map[p];
+    arc.nextstate = dst_state;
+
+    // Initialize the MutableArcIterator only now, as the call to NewState()
+    // may have invalidated the first arc iterator.
+    MutableArcIterator<MutableFst<Arc> > maiter(fst, s);
+    maiter.Seek(arcs_to_change[i].second);
+    maiter.SetValue(arc);
+  }
+}
+
+template <class Arc>
+void MakeFollowingInputSymbolsSame(bool end_is_epsilon, MutableFst<Arc> *fst) {
+  IdentityFunction<typename Arc::Label> f;
+  MakeFollowingInputSymbolsSameClass(end_is_epsilon, fst, f);
+}
+
+template <class Arc, class F>
+void MakeFollowingInputSymbolsSameClass(bool end_is_epsilon,
+                                        MutableFst<Arc> *fst, const F &f) {
+  typedef typename Arc::StateId StateId;
+  typedef typename Arc::Weight Weight;
+  typedef typename F::Result ClassType;
+  std::vector<StateId> bad_states;
+  ClassType noClass = f(kNoLabel);
+  ClassType epsClass = f(0);
+  for (StateIterator<Fst<Arc> > siter(*fst); !siter.Done(); siter.Next()) {
+    StateId s = siter.Value();
+    ClassType c = noClass;
+    bool bad = false;
+    for (ArcIterator<Fst<Arc> > aiter(*fst, s); !aiter.Done(); aiter.Next()) {
+      const Arc &arc = aiter.Value();
+      if (c == noClass) {
+        c = f(arc.ilabel);
+      } else if (c != f(arc.ilabel)) {
+        bad = true;
+        break;
+      }
+    }
+    if (end_is_epsilon && c != noClass && c != epsClass &&
+        fst->Final(s) != Weight::Zero())
+      bad = true;
+    if (bad) bad_states.push_back(s);
+  }
+  std::vector<Arc> my_arcs;
+  for (size_t i = 0; i < bad_states.size(); i++) {
+    StateId s = bad_states[i];
+    my_arcs.clear();
+    for (ArcIterator<MutableFst<Arc> > aiter(*fst, s); !aiter.Done();
+         aiter.Next())
+      my_arcs.push_back(aiter.Value());
+
+    for (size_t j = 0; j < my_arcs.size(); j++) {
+      Arc &arc = my_arcs[j];
+      if (arc.ilabel != 0) {
+        StateId newstate = fst->AddState();
+        // Create a new state for each non-eps arc in original FST, out of each
+        // bad state. Not as optimal as it could be, but does avoid some
+        // complicated weight-pushing issues in which, to maintain
+        // stochasticity, we would have to know which semiring we want to
+        // maintain stochasticity in.
+        fst->AddArc(newstate, Arc(arc.ilabel, 0, Weight::One(), arc.nextstate));
+        MutableArcIterator<MutableFst<Arc> > maiter(fst, s);
+        maiter.Seek(j);
+        maiter.SetValue(Arc(0, arc.olabel, arc.weight, newstate));
+      }
+    }
+  }
+}
+
+template <class Arc>
+VectorFst<Arc> *MakeLoopFst(const std::vector<const ExpandedFst<Arc> *> &fsts) {
+  typedef typename Arc::Weight Weight;
+  typedef typename Arc::StateId StateId;
+  typedef typename Arc::Label Label;
+
+  VectorFst<Arc> *ans = new VectorFst<Arc>;
+  StateId loop_state = ans->AddState();  // = 0.
+  ans->SetStart(loop_state);
+  ans->SetFinal(loop_state, Weight::One());
+
+  // "cache" is used as an optimization when some of the pointers in "fsts"
+  // may have the same value.
+  unordered_map<const ExpandedFst<Arc> *, Arc> cache;
+
+  for (Label i = 0; i < static_cast<Label>(fsts.size()); i++) {
+    const ExpandedFst<Arc> *fst = fsts[i];
+    if (fst == NULL) continue;
+    {  // optimization with cache: helpful if some members of "fsts" may
+      // contain the same pointer value (e.g. in GetHTransducer).
+      typename unordered_map<const ExpandedFst<Arc> *, Arc>::iterator iter =
+          cache.find(fst);
+      if (iter != cache.end()) {
+        Arc arc = iter->second;
+        arc.olabel = i;
+        ans->AddArc(0, arc);
+        continue;
+      }
+    }
+
+    KALDI_ASSERT(fst->Properties(kAcceptor, true) ==
+                 kAcceptor);  // expect acceptor.
+
+    StateId fst_num_states = fst->NumStates();
+    StateId fst_start_state = fst->Start();
+
+    if (fst_start_state == kNoStateId) continue;  // empty fst.
+
+    bool share_start_state =
+        fst->Properties(kInitialAcyclic, true) == kInitialAcyclic &&
+        fst->NumArcs(fst_start_state) == 1 &&
+        fst->Final(fst_start_state) == Weight::Zero();
+
+    std::vector<StateId> state_map(fst_num_states);  // fst state -> ans state
+    for (StateId s = 0; s < fst_num_states; s++) {
+      if (s == fst_start_state && share_start_state)
+        state_map[s] = loop_state;
+      else
+        state_map[s] = ans->AddState();
+    }
+    if (!share_start_state) {
+      Arc arc(0, i, Weight::One(), state_map[fst_start_state]);
+      cache[fst] = arc;
+      ans->AddArc(0, arc);
+    }
+    for (StateId s = 0; s < fst_num_states; s++) {
+      // Add arcs out of state s.
+      for (ArcIterator<ExpandedFst<Arc> > aiter(*fst, s); !aiter.Done();
+           aiter.Next()) {
+        const Arc &arc = aiter.Value();
+        Label olabel = (s == fst_start_state && share_start_state ? i : 0);
+        Arc newarc(arc.ilabel, olabel, arc.weight, state_map[arc.nextstate]);
+        ans->AddArc(state_map[s], newarc);
+        if (s == fst_start_state && share_start_state) cache[fst] = newarc;
+      }
+      if (fst->Final(s) != Weight::Zero()) {
+        KALDI_ASSERT(!(s == fst_start_state && share_start_state));
+        ans->AddArc(state_map[s], Arc(0, 0, fst->Final(s), loop_state));
+      }
+    }
+  }
+  return ans;
+}
+
+template <class Arc>
+void ClearSymbols(bool clear_input, bool clear_output, MutableFst<Arc> *fst) {
+  for (StateIterator<MutableFst<Arc> > siter(*fst); !siter.Done();
+       siter.Next()) {
+    typename Arc::StateId s = siter.Value();
+    for (MutableArcIterator<MutableFst<Arc> > aiter(fst, s); !aiter.Done();
+         aiter.Next()) {
+      Arc arc = aiter.Value();
+      bool change = false;
+      if (clear_input && arc.ilabel != 0) {
+        arc.ilabel = 0;
+        change = true;
+      }
+      if (clear_output && arc.olabel != 0) {
+        arc.olabel = 0;
+        change = true;
+      }
+      if (change) {
+        aiter.SetValue(arc);
+      }
+    }
+  }
+}
+
+template <class Arc>
+void ApplyProbabilityScale(float scale, MutableFst<Arc> *fst) {
+  typedef typename Arc::Weight Weight;
+  typedef typename Arc::StateId StateId;
+  for (StateIterator<MutableFst<Arc> > siter(*fst); !siter.Done();
+       siter.Next()) {
+    StateId s = siter.Value();
+    for (MutableArcIterator<MutableFst<Arc> > aiter(fst, s); !aiter.Done();
+         aiter.Next()) {
+      Arc arc = aiter.Value();
+      arc.weight = Weight(arc.weight.Value() * scale);
+      aiter.SetValue(arc);
+    }
+    if (fst->Final(s) != Weight::Zero())
+      fst->SetFinal(s, Weight(fst->Final(s).Value() * scale));
+  }
+}
+
+// return arc-offset of self-loop with ilabel (or -1 if none exists).
+// if more than one such self-loop, pick first one.
+template <class Arc>
+ssize_t FindSelfLoopWithILabel(const Fst<Arc> &fst, typename Arc::StateId s) {
+  for (ArcIterator<Fst<Arc> > aiter(fst, s); !aiter.Done(); aiter.Next())
+    if (aiter.Value().nextstate == s && aiter.Value().ilabel != 0)
+      return static_cast<ssize_t>(aiter.Position());
+  return static_cast<ssize_t>(-1);
+}
+
+template <class Arc>
+bool EqualAlign(const Fst<Arc> &ifst, typename Arc::StateId length,
+                int rand_seed, MutableFst<Arc> *ofst, int num_retries) {
+  srand(rand_seed);
+  KALDI_ASSERT(ofst->NumStates() == 0);  // make sure ofst empty.
+  // make sure all states can reach final-state (or this algorithm may enter
+  // infinite loop.
+  KALDI_ASSERT(ifst.Properties(kCoAccessible, true) == kCoAccessible);
+
+  typedef typename Arc::StateId StateId;
+  typedef typename Arc::Weight Weight;
+
+  if (ifst.Start() == kNoStateId) {
+    KALDI_WARN << "Empty input fst.";
+    return false;
+  }
+  // First select path through ifst.
+  std::vector<StateId> path;
+  std::vector<size_t> arc_offsets;  // arc taken out of each state.
+  std::vector<int> nof_ilabels;
+
+  StateId num_ilabels = 0;
+  int retry_no = 0;
+
+  // Under normal circumstances, this will be one-pass-only process
+  // Multiple tries might be needed in special cases, typically when
+  // the number of frames is close to number of transitions from
+  // the start node to the final node. It usually happens for really
+  // short utterances
+  do {
+    num_ilabels = 0;
+    arc_offsets.clear();
+    path.clear();
+    path.push_back(ifst.Start());
+
+    while (1) {
+      // Select either an arc or final-prob.
+      StateId s = path.back();
+      size_t num_arcs = ifst.NumArcs(s);
+      size_t num_arcs_tot = num_arcs;
+      if (ifst.Final(s) != Weight::Zero()) num_arcs_tot++;
+      // kaldi::RandInt is a bit like Rand(), but gets around situations
+      // where RAND_MAX is very small.
+      // Change this to Rand() % num_arcs_tot if compile issues arise
+      size_t arc_offset =
+          static_cast<size_t>(kaldi::RandInt(0, num_arcs_tot - 1));
+
+      if (arc_offset < num_arcs) {  // an actual arc.
+        ArcIterator<Fst<Arc> > aiter(ifst, s);
+        aiter.Seek(arc_offset);
+        const Arc &arc = aiter.Value();
+        if (arc.nextstate == s) {
+          continue;  // don't take this self-loop arc
+        } else {
+          arc_offsets.push_back(arc_offset);
+          path.push_back(arc.nextstate);
+          if (arc.ilabel != 0) num_ilabels++;
+        }
+      } else {
+        break;  // Chose final-prob.
+      }
+    }
+
+    nof_ilabels.push_back(num_ilabels);
+  } while ((++retry_no < num_retries) && (num_ilabels > length));
+
+  if (num_ilabels > length) {
+    std::stringstream ilabel_vec;
+    std::copy(nof_ilabels.begin(), nof_ilabels.end(),
+              std::ostream_iterator<int>(ilabel_vec, ","));
+    std::string s = ilabel_vec.str();
+    s.erase(s.end() - 1);
+    KALDI_WARN << "EqualAlign: the randomly constructed paths lengths: " << s;
+    KALDI_WARN << "EqualAlign: utterance has too few frames " << length
+               << " to align.";
+    return false;  // can't make it shorter by adding self-loops!.
+  }
+
+  StateId num_self_loops = 0;
+  std::vector<ssize_t> self_loop_offsets(path.size());
+  for (size_t i = 0; i < path.size(); i++)
+    if ((self_loop_offsets[i] = FindSelfLoopWithILabel(ifst, path[i])) !=
+        static_cast<ssize_t>(-1))
+      num_self_loops++;
+
+  if (num_self_loops == 0 && num_ilabels < length) {
+    KALDI_WARN << "No self-loops on chosen path; cannot match length.";
+    return false;  // no self-loops to make it longer.
+  }
+
+  StateId num_extra = length - num_ilabels;  // Number of self-loops we need.
+
+  StateId min_num_loops = 0;
+  if (num_extra != 0)
+    min_num_loops = num_extra / num_self_loops;  // prevent div by zero.
+  StateId num_with_one_more_loop = num_extra - (min_num_loops * num_self_loops);
+  KALDI_ASSERT(num_with_one_more_loop < num_self_loops || num_self_loops == 0);
+
+  ofst->AddState();
+  ofst->SetStart(0);
+  StateId cur_state = 0;
+  StateId counter = 0;  // tell us when we should stop adding one more loop.
+  for (size_t i = 0; i < path.size(); i++) {
+    // First, add any self-loops that are necessary.
+    StateId num_loops = 0;
+    if (self_loop_offsets[i] != static_cast<ssize_t>(-1)) {
+      num_loops = min_num_loops + (counter < num_with_one_more_loop ? 1 : 0);
+      counter++;
+    }
+    for (StateId j = 0; j < num_loops; j++) {
+      ArcIterator<Fst<Arc> > aiter(ifst, path[i]);
+      aiter.Seek(self_loop_offsets[i]);
+      Arc arc = aiter.Value();
+      KALDI_ASSERT(arc.nextstate == path[i] &&
+                   arc.ilabel != 0);  // make sure self-loop with ilabel.
+      StateId next_state = ofst->AddState();
+      ofst->AddArc(cur_state,
+                   Arc(arc.ilabel, arc.olabel, arc.weight, next_state));
+      cur_state = next_state;
+    }
+    if (i + 1 < path.size()) {  // add forward transition.
+      ArcIterator<Fst<Arc> > aiter(ifst, path[i]);
+      aiter.Seek(arc_offsets[i]);
+      Arc arc = aiter.Value();
+      KALDI_ASSERT(arc.nextstate == path[i + 1]);
+      StateId next_state = ofst->AddState();
+      ofst->AddArc(cur_state,
+                   Arc(arc.ilabel, arc.olabel, arc.weight, next_state));
+      cur_state = next_state;
+    } else {  // add final-prob.
+      Weight weight = ifst.Final(path[i]);
+      KALDI_ASSERT(weight != Weight::Zero());
+      ofst->SetFinal(cur_state, weight);
+    }
+  }
+  return true;
+}
+
+// This function identifies two types of useless arcs:
+// those where arc A and arc B both go from state X to
+// state Y with the same input symbol (remove the one
+// with smaller probability, or an arbitrary one if they
+// are the same); and those where A is an arc from state X
+// to state X, with epsilon input symbol [remove A].
+// Only works for tropical (not log) semiring as it uses
+// NaturalLess.
+template <class Arc>
+void RemoveUselessArcs(MutableFst<Arc> *fst) {
+  typedef typename Arc::Label Label;
+  typedef typename Arc::StateId StateId;
+  typedef typename Arc::Weight Weight;
+  NaturalLess<Weight> nl;
+  StateId non_coacc_state = kNoStateId;
+  size_t num_arcs_removed = 0, tot_arcs = 0;
+  for (StateIterator<MutableFst<Arc> > siter(*fst); !siter.Done();
+       siter.Next()) {
+    std::vector<size_t> arcs_to_delete;
+    std::vector<Arc> arcs;
+    // pair2arclist lets us look up the arcs
+    std::map<std::pair<Label, StateId>, std::vector<size_t> > pair2arclist;
+    StateId state = siter.Value();
+    for (ArcIterator<MutableFst<Arc> > aiter(*fst, state); !aiter.Done();
+         aiter.Next()) {
+      size_t pos = arcs.size();
+      const Arc &arc = aiter.Value();
+      arcs.push_back(arc);
+      pair2arclist[std::make_pair(arc.ilabel, arc.nextstate)].push_back(pos);
+    }
+    typename std::map<std::pair<Label, StateId>, std::vector<size_t> >::iterator
+        iter = pair2arclist.begin(),
+        end = pair2arclist.end();
+    for (; iter != end; ++iter) {
+      const std::vector<size_t> &poslist = iter->second;
+      if (poslist.size() > 1) {  // >1 arc with same ilabel, dest-state
+        size_t best_pos = poslist[0];
+        Weight best_weight = arcs[best_pos].weight;
+        for (size_t j = 1; j < poslist.size(); j++) {
+          size_t pos = poslist[j];
+          Weight this_weight = arcs[pos].weight;
+          if (nl(this_weight,
+                 best_weight)) {  // NaturalLess seems to be somehow
+            // "backwards".
+            best_weight = this_weight;  // found a better one.
+            best_pos = pos;
+          }
+        }
+        for (size_t j = 0; j < poslist.size(); j++)
+          if (poslist[j] != best_pos) arcs_to_delete.push_back(poslist[j]);
+      } else {
+        KALDI_ASSERT(poslist.size() == 1);
+        size_t pos = poslist[0];
+        Arc &arc = arcs[pos];
+        if (arc.ilabel == 0 && arc.nextstate == state)
+          arcs_to_delete.push_back(pos);
+      }
+    }
+    tot_arcs += arcs.size();
+    if (arcs_to_delete.size() != 0) {
+      num_arcs_removed += arcs_to_delete.size();
+      if (non_coacc_state == kNoStateId) non_coacc_state = fst->AddState();
+      MutableArcIterator<MutableFst<Arc> > maiter(fst, state);
+      for (size_t j = 0; j < arcs_to_delete.size(); j++) {
+        size_t pos = arcs_to_delete[j];
+        maiter.Seek(pos);
+        arcs[pos].nextstate = non_coacc_state;
+        maiter.SetValue(arcs[pos]);
+      }
+    }
+  }
+  if (non_coacc_state != kNoStateId) Connect(fst);
+  KALDI_VLOG(1) << "removed " << num_arcs_removed << " of " << tot_arcs
+                << "arcs.";
+}
+
+template <class Arc>
+void PhiCompose(const Fst<Arc> &fst1, const Fst<Arc> &fst2,
+                typename Arc::Label phi_label, MutableFst<Arc> *ofst) {
+  KALDI_ASSERT(phi_label !=
+               kNoLabel);  // just use regular compose in this case.
+  typedef Fst<Arc> F;
+  typedef PhiMatcher<SortedMatcher<F> > PM;
+  CacheOptions base_opts;
+  base_opts.gc_limit = 0;  // Cache only the last state for fastest copy.
+  // ComposeFstImplOptions templated on matcher for fst1, matcher for fst2.
+  // The matcher for fst1 doesn't matter; we'll use fst2's matcher.
+  ComposeFstImplOptions<SortedMatcher<F>, PM> impl_opts(base_opts);
+
+  // the false below is something called phi_loop which is something I don't
+  // fully understand, but I don't think we want it.
+
+  // These pointers are taken ownership of, by ComposeFst.
+  PM *phi_matcher = new PM(fst2, MATCH_INPUT, phi_label, false);
+  SortedMatcher<F> *sorted_matcher =
+      new SortedMatcher<F>(fst1, MATCH_NONE);  // tell it
+  // not to use this matcher, as this would mean we would
+  // not follow phi transitions.
+  impl_opts.matcher1 = sorted_matcher;
+  impl_opts.matcher2 = phi_matcher;
+  *ofst = ComposeFst<Arc>(fst1, fst2, impl_opts);
+  Connect(ofst);
+}
+
+template <class Arc>
+void PropagateFinalInternal(typename Arc::Label phi_label,
+                            typename Arc::StateId s, MutableFst<Arc> *fst) {
+  typedef typename Arc::Weight Weight;
+  if (fst->Final(s) == Weight::Zero()) {
+    // search for phi transition.  We assume there
+    // is just one-- phi nondeterminism is not allowed
+    // anyway.
+    int num_phis = 0;
+    for (ArcIterator<Fst<Arc> > aiter(*fst, s); !aiter.Done(); aiter.Next()) {
+      const Arc &arc = aiter.Value();
+      if (arc.ilabel == phi_label) {
+        num_phis++;
+        if (arc.nextstate == s) continue;  // don't expect
+        // phi loops but ignore them anyway.
+
+        // If this recurses infinitely, it means there
+        // are loops of phi transitions, which there should
+        // not be in a normal backoff LM.  We could make this
+        // routine work for this case, but currently there is
+        // no need.
+        PropagateFinalInternal(phi_label, arc.nextstate, fst);
+        if (fst->Final(arc.nextstate) != Weight::Zero())
+          fst->SetFinal(s, Times(fst->Final(arc.nextstate), arc.weight));
+      }
+      KALDI_ASSERT(num_phis <= 1 && "Phi nondeterminism found");
+    }
+  }
+}
+
+template <class Arc>
+void PropagateFinal(typename Arc::Label phi_label, MutableFst<Arc> *fst) {
+  typedef typename Arc::StateId StateId;
+  if (fst->Properties(kIEpsilons, true))  // just warn.
+    KALDI_WARN << "PropagateFinal: this may not work as desired "
+                  "since your FST has input epsilons.";
+  StateId num_states = fst->NumStates();
+  for (StateId s = 0; s < num_states; s++)
+    PropagateFinalInternal(phi_label, s, fst);
+}
+
+template <class Arc>
+void RhoCompose(const Fst<Arc> &fst1, const Fst<Arc> &fst2,
+                typename Arc::Label rho_label, MutableFst<Arc> *ofst) {
+  KALDI_ASSERT(rho_label !=
+               kNoLabel);  // just use regular compose in this case.
+  typedef Fst<Arc> F;
+  typedef RhoMatcher<SortedMatcher<F> > RM;
+  CacheOptions base_opts;
+  base_opts.gc_limit = 0;  // Cache only the last state for fastest copy.
+  // ComposeFstImplOptions templated on matcher for fst1, matcher for fst2.
+  // The matcher for fst1 doesn't matter; we'll use fst2's matcher.
+  ComposeFstImplOptions<SortedMatcher<F>, RM> impl_opts(base_opts);
+
+  // the false below is something called rho_loop which is something I don't
+  // fully understand, but I don't think we want it.
+
+  // These pointers are taken ownership of, by ComposeFst.
+  RM *rho_matcher = new RM(fst2, MATCH_INPUT, rho_label);
+  SortedMatcher<F> *sorted_matcher =
+      new SortedMatcher<F>(fst1, MATCH_NONE);  // tell it
+  // not to use this matcher, as this would mean we would
+  // not follow rho transitions.
+  impl_opts.matcher1 = sorted_matcher;
+  impl_opts.matcher2 = rho_matcher;
+  *ofst = ComposeFst<Arc>(fst1, fst2, impl_opts);
+  Connect(ofst);
+}
+
+// Declare an override of the template below.
+template <>
+inline bool IsStochasticFst(const Fst<LogArc> &fst, float delta,
+                            LogArc::Weight *min_sum, LogArc::Weight *max_sum);
+
+// Will override this for LogArc where NaturalLess will not work.
+template <class Arc>
+inline bool IsStochasticFst(const Fst<Arc> &fst, float delta,
+                            typename Arc::Weight *min_sum,
+                            typename Arc::Weight *max_sum) {
+  typedef typename Arc::StateId StateId;
+  typedef typename Arc::Weight Weight;
+  NaturalLess<Weight> nl;
+  bool first_time = true;
+  bool ans = true;
+  if (min_sum) *min_sum = Arc::Weight::One();
+  if (max_sum) *max_sum = Arc::Weight::One();
+  for (StateIterator<Fst<Arc> > siter(fst); !siter.Done(); siter.Next()) {
+    StateId s = siter.Value();
+    Weight sum = fst.Final(s);
+    for (ArcIterator<Fst<Arc> > aiter(fst, s); !aiter.Done(); aiter.Next()) {
+      const Arc &arc = aiter.Value();
+      sum = Plus(sum, arc.weight);
+    }
+    if (!ApproxEqual(Weight::One(), sum, delta)) ans = false;
+    if (first_time) {
+      first_time = false;
+      if (max_sum) *max_sum = sum;
+      if (min_sum) *min_sum = sum;
+    } else {
+      if (max_sum && nl(*max_sum, sum)) *max_sum = sum;
+      if (min_sum && nl(sum, *min_sum)) *min_sum = sum;
+    }
+  }
+  if (first_time) {  // just avoid NaNs if FST was empty.
+    if (max_sum) *max_sum = Weight::One();
+    if (min_sum) *min_sum = Weight::One();
+  }
+  return ans;
+}
+
+// Overriding template for LogArc as NaturalLess does not work there.
+template <>
+inline bool IsStochasticFst(const Fst<LogArc> &fst, float delta,
+                            LogArc::Weight *min_sum, LogArc::Weight *max_sum) {
+  typedef LogArc Arc;
+  typedef Arc::StateId StateId;
+  typedef Arc::Weight Weight;
+  bool first_time = true;
+  bool ans = true;
+  if (min_sum) *min_sum = LogArc::Weight::One();
+  if (max_sum) *max_sum = LogArc::Weight::One();
+  for (StateIterator<Fst<Arc> > siter(fst); !siter.Done(); siter.Next()) {
+    StateId s = siter.Value();
+    Weight sum = fst.Final(s);
+    for (ArcIterator<Fst<Arc> > aiter(fst, s); !aiter.Done(); aiter.Next()) {
+      const Arc &arc = aiter.Value();
+      sum = Plus(sum, arc.weight);
+    }
+    if (!ApproxEqual(Weight::One(), sum, delta)) ans = false;
+    if (first_time) {
+      first_time = false;
+      if (max_sum) *max_sum = sum;
+      if (min_sum) *min_sum = sum;
+    } else {
+      // note that max and min are reversed from their normal
+      // meanings here (max and min w.r.t. the underlying probabilities).
+      if (max_sum && sum.Value() < max_sum->Value()) *max_sum = sum;
+      if (min_sum && sum.Value() > min_sum->Value()) *min_sum = sum;
+    }
+  }
+  if (first_time) {  // just avoid NaNs if FST was empty.
+    if (max_sum) *max_sum = Weight::One();
+    if (min_sum) *min_sum = Weight::One();
+  }
+  return ans;
+}
+
+// Tests whether a tropical FST is stochastic in the log
+// semiring. (casts it and does the check.)
+// This function deals with the generic fst.
+// This version currently supports ConstFst<StdArc> or VectorFst<StdArc>.
+// Otherwise, it will be died with an error.
+inline bool IsStochasticFstInLog(const Fst<StdArc> &fst, float delta,
+                                 StdArc::Weight *min_sum,
+                                 StdArc::Weight *max_sum) {
+  bool ans = false;
+  LogArc::Weight log_min = LogArc::Weight::One(),
+                 log_max = LogArc::Weight::Zero();
+  if (fst.Type() == "const") {
+    ConstFst<LogArc> logfst;
+    Cast(dynamic_cast<const ConstFst<StdArc> &>(fst), &logfst);
+    ans = IsStochasticFst(logfst, delta, &log_min, &log_max);
+  } else if (fst.Type() == "vector") {
+    VectorFst<LogArc> logfst;
+    Cast(dynamic_cast<const VectorFst<StdArc> &>(fst), &logfst);
+    ans = IsStochasticFst(logfst, delta, &log_min, &log_max);
+  } else {
+    KALDI_ERR << "This version currently supports ConstFst<StdArc> "
+              << "or VectorFst<StdArc>";
+  }
+  if (min_sum) *min_sum = StdArc::Weight(log_min.Value());
+  if (max_sum) *max_sum = StdArc::Weight(log_max.Value());
+  return ans;
+}
+
+}  // namespace fst.
+
+#endif  // KALDI_FSTEXT_FSTEXT_UTILS_INL_H_
diff --git a/speechx/speechx/kaldi/fstext/fstext-utils.h b/speechx/speechx/kaldi/fstext/fstext-utils.h
new file mode 100644
index 00000000..b0aed022
--- /dev/null
+++ b/speechx/speechx/kaldi/fstext/fstext-utils.h
@@ -0,0 +1,386 @@
+// fstext/fstext-utils.h
+
+// Copyright 2009-2011  Microsoft Corporation
+//           2012-2013  Johns Hopkins University (Author: Daniel Povey)
+//                2013  Guoguo Chen
+//                2014  Telepoint Global Hosting Service, LLC. (Author: David
+//                Snyder)
+
+// See ../../COPYING for clarification regarding multiple authors
+//
+// Licensed under the Apache License, Version 2.0 (the "License");
+// you may not use this file except in compliance with the License.
+// You may obtain a copy of the License at
+//
+//  http://www.apache.org/licenses/LICENSE-2.0
+//
+// THIS CODE IS PROVIDED *AS IS* BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY
+// KIND, EITHER EXPRESS OR IMPLIED, INCLUDING WITHOUT LIMITATION ANY IMPLIED
+// WARRANTIES OR CONDITIONS OF TITLE, FITNESS FOR A PARTICULAR PURPOSE,
+// MERCHANTABLITY OR NON-INFRINGEMENT.
+// See the Apache 2 License for the specific language governing permissions and
+// limitations under the License.
+
+#ifndef KALDI_FSTEXT_FSTEXT_UTILS_H_
+#define KALDI_FSTEXT_FSTEXT_UTILS_H_
+
+#include <fst/fst-decl.h>
+#include <fst/fstlib.h>
+
+#include <algorithm>
+#include <map>
+#include <set>
+#include <vector>
+
+#include "fstext/determinize-star.h"
+#include "fstext/remove-eps-local.h"
+#include "base/kaldi-common.h"  // for error reporting macros.
+#include "util/text-utils.h"  // for SplitStringToVector
+#include "fst/script/print-impl.h"
+
+namespace fst {
+
+/// Returns the highest numbered output symbol id of the FST (or zero
+/// for an empty FST.
+template <class Arc>
+typename Arc::Label HighestNumberedOutputSymbol(const Fst<Arc> &fst);
+
+/// Returns the highest numbered input symbol id of the FST (or zero
+/// for an empty FST.
+template <class Arc>
+typename Arc::Label HighestNumberedInputSymbol(const Fst<Arc> &fst);
+
+/// Returns the total number of arcs in an FST.
+template <class Arc>
+typename Arc::StateId NumArcs(const ExpandedFst<Arc> &fst);
+
+/// GetInputSymbols gets the list of symbols on the input of fst
+/// (including epsilon, if include_eps == true), as a sorted, unique
+/// list.
+template <class Arc, class I>
+void GetInputSymbols(const Fst<Arc> &fst, bool include_eps,
+                     std::vector<I> *symbols);
+
+/// GetOutputSymbols gets the list of symbols on the output of fst
+/// (including epsilon, if include_eps == true)
+template <class Arc, class I>
+void GetOutputSymbols(const Fst<Arc> &fst, bool include_eps,
+                      std::vector<I> *symbols);
+
+/// ClearSymbols sets all the symbols on the input and/or
+/// output side of the FST to zero, as specified.
+/// It does not alter the symbol tables.
+template <class Arc>
+void ClearSymbols(bool clear_input, bool clear_output, MutableFst<Arc> *fst);
+
+template <class I>
+void GetSymbols(const SymbolTable &symtab, bool include_eps,
+                std::vector<I> *syms_out);
+
+inline void DeterminizeStarInLog(VectorFst<StdArc> *fst, float delta = kDelta,
+                                 bool *debug_ptr = NULL, int max_states = -1);
+
+// e.g. of using this function: PushInLog<REWEIGHT_TO_INITIAL>(fst,
+// kPushWeights|kPushLabels);
+
+template <ReweightType rtype>  // == REWEIGHT_TO_{INITIAL, FINAL}
+void PushInLog(VectorFst<StdArc> *fst, uint32 ptype, float delta = kDelta) {
+  // PushInLog pushes the FST
+  // and returns a new pushed FST (labels and weights pushed to the left).
+  VectorFst<LogArc> *fst_log =
+      new VectorFst<LogArc>;  // Want to determinize in log semiring.
+  Cast(*fst, fst_log);
+  VectorFst<StdArc> tmp;
+  *fst = tmp;  // free up memory.
+  VectorFst<LogArc> *fst_pushed_log = new VectorFst<LogArc>;
+  Push<LogArc, rtype>(*fst_log, fst_pushed_log, ptype, delta);
+  Cast(*fst_pushed_log, fst);
+  delete fst_log;
+  delete fst_pushed_log;
+}
+
+// Minimizes after encoding; applicable to all FSTs.  It is like what you get
+// from the Minimize() function, except it will not push the weights, or the
+// symbols.  This is better for our recipes, as we avoid ever pushing the
+// weights.  However, it will only minimize optimally if your graphs are such
+// that the symbols are as far to the left as they can go, and the weights
+// in combinable paths are the same... hard to formalize this, but it's
+// something that is satisified by our normal FSTs.
+template <class Arc>
+void MinimizeEncoded(VectorFst<Arc> *fst, float delta = kDelta) {
+  Map(fst, QuantizeMapper<Arc>(delta));
+  EncodeMapper<Arc> encoder(kEncodeLabels | kEncodeWeights, ENCODE);
+  Encode(fst, &encoder);
+  internal::AcceptorMinimize(fst);
+  Decode(fst, encoder);
+}
+
+/// GetLinearSymbolSequence gets the symbol sequence from a linear FST.
+/// If the FST is not just a linear sequence, it returns false.   If it is
+/// a linear sequence (including the empty FST), it returns true.  In this
+/// case it outputs the symbol
+/// sequences as "isymbols_out" and "osymbols_out" (removing epsilons), and
+/// the total weight as "tot_weight". The total weight will be Weight::Zero()
+/// if the FST is empty.  If any of the output pointers are NULL, it does not
+/// create that output.
+template <class Arc, class I>
+bool GetLinearSymbolSequence(const Fst<Arc> &fst, std::vector<I> *isymbols_out,
+                             std::vector<I> *osymbols_out,
+                             typename Arc::Weight *tot_weight_out);
+
+/// This function converts an FST with a special structure, which is
+/// output by the OpenFst functions ShortestPath and RandGen, and converts
+/// them into a std::vector of separate FSTs.  This special structure is that
+/// the only state that has more than one (arcs-out or final-prob) is the
+/// start state.  fsts_out is resized to the appropriate size.
+template <class Arc>
+void ConvertNbestToVector(const Fst<Arc> &fst,
+                          std::vector<VectorFst<Arc> > *fsts_out);
+
+/// Takes the n-shortest-paths (using ShortestPath), but outputs
+/// the result as a vector of up to n fsts.  This function will
+/// size the "fsts_out" vector to however many paths it got
+/// (which will not exceed n).  n must be >= 1.
+template <class Arc>
+void NbestAsFsts(const Fst<Arc> &fst, size_t n,
+                 std::vector<VectorFst<Arc> > *fsts_out);
+
+/// Creates unweighted linear acceptor from symbol sequence.
+template <class Arc, class I>
+void MakeLinearAcceptor(const std::vector<I> &labels, MutableFst<Arc> *ofst);
+
+/// Creates an unweighted acceptor with a linear structure, with alternatives
+/// at each position.  Epsilon is treated like a normal symbol here.
+/// Each position in "labels" must have at least one alternative.
+template <class Arc, class I>
+void MakeLinearAcceptorWithAlternatives(
+    const std::vector<std::vector<I> > &labels, MutableFst<Arc> *ofst);
+
+/// Does PreDeterminize and DeterminizeStar and then removes the disambiguation
+/// symbols. This is a form of determinization that will never blow up. Note
+/// that ifst is non-const and can be considered to be destroyed by this
+/// operation.
+/// Does not do epsilon removal (RemoveEpsLocal)-- this is so it's safe to cast
+/// to log and do this, and maintain equivalence in tropical.
+
+template <class Arc>
+void SafeDeterminizeWrapper(MutableFst<Arc> *ifst, MutableFst<Arc> *ofst,
+                            float delta = kDelta);
+
+/// SafeDeterminizeMinimizeWapper is as SafeDeterminizeWrapper except that it
+/// also minimizes (encoded minimization, which is safe).  This algorithm will
+/// destroy "ifst".
+template <class Arc>
+void SafeDeterminizeMinimizeWrapper(MutableFst<Arc> *ifst, VectorFst<Arc> *ofst,
+                                    float delta = kDelta);
+
+/// SafeDeterminizeMinimizeWapperInLog is as SafeDeterminizeMinimizeWrapper
+/// except it first casts tothe log semiring.
+void SafeDeterminizeMinimizeWrapperInLog(VectorFst<StdArc> *ifst,
+                                         VectorFst<StdArc> *ofst,
+                                         float delta = kDelta);
+
+/// RemoveSomeInputSymbols removes any symbol that appears in "to_remove", from
+/// the input side of the FST, replacing them with epsilon.
+template <class Arc, class I>
+void RemoveSomeInputSymbols(const std::vector<I> &to_remove,
+                            MutableFst<Arc> *fst);
+
+// MapInputSymbols will replace any input symbol i that is between 0 and
+// symbol_map.size()-1, with symbol_map[i].  It removes the input symbol
+// table of the FST.
+template <class Arc, class I>
+void MapInputSymbols(const std::vector<I> &symbol_map, MutableFst<Arc> *fst);
+
+template <class Arc>
+void RemoveWeights(MutableFst<Arc> *fst);
+
+/// Returns true if and only if the FST is such that the input symbols
+/// on arcs entering any given state all have the same value.
+/// if "start_is_epsilon", treat start-state as an epsilon input arc
+/// [i.e. ensure only epsilon can enter start-state].
+template <class Arc>
+bool PrecedingInputSymbolsAreSame(bool start_is_epsilon, const Fst<Arc> &fst);
+
+/// This is as PrecedingInputSymbolsAreSame, but with a functor f that maps
+/// labels to classes. The function tests whether the symbols preceding any
+/// given state are in the same class. Formally, f is of a type F that has an
+/// operator of type F::Result F::operator() (F::Arg a) const; where F::Result
+/// is an integer type and F::Arc can be constructed from Arc::Label. this must
+/// apply to valid labels and also to kNoLabel (so we can have a marker for the
+/// invalid labels.
+template <class Arc, class F>
+bool PrecedingInputSymbolsAreSameClass(bool start_is_epsilon,
+                                       const Fst<Arc> &fst, const F &f);
+
+/// Returns true if and only if the FST is such that the input symbols
+/// on arcs exiting any given state all have the same value.
+/// If end_is_epsilon, treat end-state as an epsilon output arc [i.e. ensure
+/// end-states cannot have non-epsilon output transitions.]
+template <class Arc>
+bool FollowingInputSymbolsAreSame(bool end_is_epsilon, const Fst<Arc> &fst);
+
+template <class Arc, class F>
+bool FollowingInputSymbolsAreSameClass(bool end_is_epsilon, const Fst<Arc> &fst,
+                                       const F &f);
+
+/// MakePrecedingInputSymbolsSame ensures that all arcs entering any given fst
+/// state have the same input symbol.  It does this by detecting states
+/// that have differing input symbols going in, and inserting, for each of
+/// the preceding arcs with non-epsilon input symbol, a new dummy state that
+/// has an epsilon link to the fst state.
+/// If "start_is_epsilon", ensure that start-state can have only epsilon-links
+/// into it.
+template <class Arc>
+void MakePrecedingInputSymbolsSame(bool start_is_epsilon, MutableFst<Arc> *fst);
+
+/// As MakePrecedingInputSymbolsSame, but takes a functor object that maps
+/// labels to classes.
+template <class Arc, class F>
+void MakePrecedingInputSymbolsSameClass(bool start_is_epsilon,
+                                        MutableFst<Arc> *fst, const F &f);
+
+/// MakeFollowingInputSymbolsSame ensures that all arcs exiting any given fst
+/// state have the same input symbol.  It does this by detecting states that
+/// have differing input symbols on arcs that exit it, and inserting, for each
+/// of the following arcs with non-epsilon input symbol, a new dummy state that
+/// has an input-epsilon link from the fst state.  The output symbol and weight
+/// stay on the link to the dummy state (in order to keep the FST
+/// output-deterministic and stochastic, if it already was). If end_is_epsilon,
+/// treat "being a final-state" like having an epsilon output link.
+template <class Arc>
+void MakeFollowingInputSymbolsSame(bool end_is_epsilon, MutableFst<Arc> *fst);
+
+/// As MakeFollowingInputSymbolsSame, but takes a functor object that maps
+/// labels to classes.
+template <class Arc, class F>
+void MakeFollowingInputSymbolsSameClass(bool end_is_epsilon,
+                                        MutableFst<Arc> *fst, const F &f);
+
+/// MakeLoopFst creates an FST that has a state that is both initial and
+/// final (weight == Weight::One()), and for each non-NULL pointer fsts[i],
+/// it has an arc out whose output-symbol is i and which goes to a
+/// sub-graph whose input language is equivalent to fsts[i], where the
+/// final-state becomes a transition to the loop-state.  Each fst in "fsts"
+/// should be an acceptor.  The fst MakeLoopFst returns is output-deterministic,
+/// but not output-epsilon free necessarily, and arcs are sorted on output
+/// label. Note: if some of the pointers in the input vector "fsts" have the
+/// same value, "MakeLoopFst" uses this to speed up the computation.
+
+/// Formally: suppose I is the set of indexes i such that fsts[i] != NULL.
+/// Let L[i] be the language that the acceptor fsts[i] accepts.
+/// Let the language K be the set of input-output pairs i:l such
+/// that i in I and l in L[i].  Then the FST returned by MakeLoopFst
+/// accepts the language K*, where * is the Kleene closure (CLOSURE_STAR)
+/// of K.
+
+/// We could have implemented this via a combination of "project",
+/// "concat", "union" and "closure".  But that FST would have been
+/// less well optimized and would have a lot of final-states.
+
+template <class Arc>
+VectorFst<Arc> *MakeLoopFst(const std::vector<const ExpandedFst<Arc> *> &fsts);
+
+/// ApplyProbabilityScale is applicable to FSTs in the log or tropical semiring.
+/// It multiplies the arc and final weights by "scale" [this is not the Mul
+/// operation of the semiring, it's actual multiplication, which is equivalent
+/// to taking a power in the semiring].
+template <class Arc>
+void ApplyProbabilityScale(float scale, MutableFst<Arc> *fst);
+
+/// EqualAlign is similar to RandGen, but it generates a sequence with exactly
+/// "length" input symbols.  It returns true on success, false on failure
+/// (failure is partly random but should never happen in practice for normal
+/// speech models.) It generates a random path through the input FST, finds out
+/// which subset of the states it visits along the way have self-loops with
+/// inupt symbols on them, and outputs a path with exactly enough self-loops to
+/// have the requested number of input symbols. Note that EqualAlign does not
+/// use the probabilities on the FST.  It just uses equal probabilities in the
+/// first stage of selection (since the output will anyway not be a truly random
+/// sample from the FST). The input fst "ifst" must be connected or this may
+/// enter an infinite loop.
+template <class Arc>
+bool EqualAlign(const Fst<Arc> &ifst, typename Arc::StateId length,
+                int rand_seed, MutableFst<Arc> *ofst, int num_retries = 10);
+
+// RemoveUselessArcs removes arcs such that there is no input symbol
+// sequence for which the best path through the FST would contain
+// those arcs [for these purposes, epsilon is not treated as a real symbol].
+// This is mainly geared towards decoding-graph FSTs which may contain
+// transitions that have less likely words on them that would never be
+// taken.  We do not claim that this algorithm removes all such arcs;
+// it just does the best job it can.
+// Only works for tropical (not log) semiring as it uses
+// NaturalLess.
+template <class Arc>
+void RemoveUselessArcs(MutableFst<Arc> *fst);
+
+// PhiCompose is a version of composition where
+// the right hand FST (fst2) is treated as a backoff
+// LM, with the phi symbol (e.g. #0) treated as a
+// "failure transition", only taken when we don't
+// have a match for the requested symbol.
+template <class Arc>
+void PhiCompose(const Fst<Arc> &fst1, const Fst<Arc> &fst2,
+                typename Arc::Label phi_label, MutableFst<Arc> *fst);
+
+// PropagateFinal propagates final-probs through
+// "phi" transitions (note that here, phi_label may
+// be epsilon if you want).  If you have a backoff LM
+// with special symbols ("phi") on the backoff arcs
+// instead of epsilon, you may use PhiCompose to compose
+// with it, but this won't do the right thing w.r.t.
+// final probabilities.  You should first call PropagateFinal
+// on the FST with phi's i it (fst2 in PhiCompose above),
+// to fix this.  If a state does not have a final-prob,
+// but has a phi transition, it makes the state's final-prob
+// (phi-prob * final-prob-of-dest-state), and does this
+// recursively i.e. follows phi transitions on the dest state
+// first.  It behaves as if there were a super-final state
+// with a special symbol leading to it, from each currently
+// final state.  Note that this may not behave as desired
+// if there are epsilons in your FST; it might be better
+// to remove those before calling this function.
+
+template <class Arc>
+void PropagateFinal(typename Arc::Label phi_label, MutableFst<Arc> *fst);
+
+// PhiCompose is a version of composition where
+// the right hand FST (fst2) has speciall "rho transitions"
+// which are taken whenever no normal transition matches; these
+// transitions will be rewritten with whatever symbol was on
+// the first FST.
+template <class Arc>
+void RhoCompose(const Fst<Arc> &fst1, const Fst<Arc> &fst2,
+                typename Arc::Label rho_label, MutableFst<Arc> *fst);
+
+/** This function returns true if, in the semiring of the FST, the sum (within
+    the semiring) of all the arcs out of each state in the FST is one, to within
+    delta.  After MakeStochasticFst, this should be true (for a connected FST).
+
+    @param fst [in] the FST that we are testing.
+    @param delta [in] the tolerance to within which we test equality to 1.
+    @param min_sum [out] if non, NULL, contents will be set to the minimum sum
+   of weights.
+    @param max_sum [out] if non, NULL, contents will be set to the maximum sum
+   of weights.
+    @return Returns true if the FST is stochastic, and false otherwise.
+*/
+
+template <class Arc>
+bool IsStochasticFst(const Fst<Arc> &fst,
+                     float delta = kDelta,  // kDelta = 1.0/1024.0 by default.
+                     typename Arc::Weight *min_sum = NULL,
+                     typename Arc::Weight *max_sum = NULL);
+
+// IsStochasticFstInLog makes sure it's stochastic after casting to log.
+inline bool IsStochasticFstInLog(
+    const Fst<StdArc> &fst,
+    float delta = kDelta,  // kDelta = 1.0/1024.0 by default.
+    StdArc::Weight *min_sum = NULL, StdArc::Weight *max_sum = NULL);
+
+}  // end namespace fst
+
+#include "fstext/fstext-utils-inl.h"
+
+#endif  // KALDI_FSTEXT_FSTEXT_UTILS_H_
diff --git a/speechx/speechx/kaldi/fstext/kaldi-fst-io-inl.h b/speechx/speechx/kaldi/fstext/kaldi-fst-io-inl.h
new file mode 100644
index 00000000..a80505dd
--- /dev/null
+++ b/speechx/speechx/kaldi/fstext/kaldi-fst-io-inl.h
@@ -0,0 +1,208 @@
+// fstext/kaldi-fst-io-inl.h
+
+// Copyright 2009-2011  Microsoft Corporation
+//           2012-2015  Johns Hopkins University (Author: Daniel Povey)
+//                2013  Guoguo Chen
+
+// See ../../COPYING for clarification regarding multiple authors
+//
+// Licensed under the Apache License, Version 2.0 (the "License");
+// you may not use this file except in compliance with the License.
+// You may obtain a copy of the License at
+//
+//  http://www.apache.org/licenses/LICENSE-2.0
+//
+// THIS CODE IS PROVIDED *AS IS* BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY
+// KIND, EITHER EXPRESS OR IMPLIED, INCLUDING WITHOUT LIMITATION ANY IMPLIED
+// WARRANTIES OR CONDITIONS OF TITLE, FITNESS FOR A PARTICULAR PURPOSE,
+// MERCHANTABLITY OR NON-INFRINGEMENT.
+// See the Apache 2 License for the specific language governing permissions and
+// limitations under the License.
+
+#ifndef KALDI_FSTEXT_KALDI_FST_IO_INL_H_
+#define KALDI_FSTEXT_KALDI_FST_IO_INL_H_
+
+#include <string>
+#include <vector>
+
+#include "util/text-utils.h"
+
+namespace fst {
+
+template <class Arc>
+void WriteFstKaldi(std::ostream &os, bool binary, const VectorFst<Arc> &t) {
+  bool ok;
+  if (binary) {
+    // Binary-mode writing.
+    ok = t.Write(os, FstWriteOptions());
+  } else {
+    // Text-mode output.  Note: we expect that t.InputSymbols() and
+    // t.OutputSymbols() would always return NULL.  The corresponding input
+    // routine would not work if the FST actually had symbols attached.  Write a
+    // newline to start the FST; in a table, the first line of the FST will
+    // appear on its own line.
+    os << '\n';
+    bool acceptor = false, write_one = false;
+    FstPrinter<Arc> printer(t, t.InputSymbols(), t.OutputSymbols(), NULL,
+                            acceptor, write_one, "\t");
+    printer.Print(&os, "<unknown>");
+    if (os.fail()) KALDI_ERR << "Stream failure detected writing FST to stream";
+    // Write another newline as a terminating character.  The read routine will
+    // detect this [this is a Kaldi mechanism, not something in the original
+    // OpenFst code].
+    os << '\n';
+    ok = os.good();
+  }
+  if (!ok) {
+    KALDI_ERR << "Error writing FST to stream";
+  }
+}
+
+// Utility function used in ReadFstKaldi
+template <class W>
+inline bool StrToWeight(const std::string &s, bool allow_zero, W *w) {
+  std::istringstream strm(s);
+  strm >> *w;
+  if (strm.fail() || (!allow_zero && *w == W::Zero())) {
+    return false;
+  }
+  return true;
+}
+
+template <class Arc>
+void ReadFstKaldi(std::istream &is, bool binary, VectorFst<Arc> *fst) {
+  typedef typename Arc::Weight Weight;
+  typedef typename Arc::StateId StateId;
+  if (binary) {
+    // We don't have access to the filename here, so write [unknown].
+    VectorFst<Arc> *ans =
+        VectorFst<Arc>::Read(is, fst::FstReadOptions(std::string("[unknown]")));
+    if (ans == NULL) {
+      KALDI_ERR << "Error reading FST from stream.";
+    }
+    *fst = *ans;  // shallow copy.
+    delete ans;
+  } else {
+    // Consume the \r on Windows, the \n that the text-form FST format starts
+    // with, and any extra spaces that might have got in there somehow.
+    while (std::isspace(is.peek()) && is.peek() != '\n') is.get();
+    if (is.peek() == '\n') {
+      is.get();  // consume the newline.
+    } else {     // saw spaces but no newline.. this is not expected.
+      KALDI_ERR << "Reading FST: unexpected sequence of spaces "
+                << " at file position " << is.tellg();
+    }
+    using kaldi::ConvertStringToInteger;
+    using kaldi::SplitStringToIntegers;
+    using std::string;
+    using std::vector;
+    fst->DeleteStates();
+    string line;
+    size_t nline = 0;
+    string separator = FLAGS_fst_field_separator + "\r\n";
+    while (std::getline(is, line)) {
+      nline++;
+      vector<string> col;
+      // on Windows we'll write in text and read in binary mode.
+      kaldi::SplitStringToVector(line, separator.c_str(), true, &col);
+      if (col.size() == 0) break;  // Empty line is a signal to stop, in our
+      // archive format.
+      if (col.size() > 5) {
+        KALDI_ERR << "Bad line in FST: " << line;
+      }
+      StateId s;
+      if (!ConvertStringToInteger(col[0], &s)) {
+        KALDI_ERR << "Bad line in FST: " << line;
+      }
+      while (s >= fst->NumStates()) fst->AddState();
+      if (nline == 1) fst->SetStart(s);
+
+      bool ok = true;
+      Arc arc;
+      Weight w;
+      StateId d = s;
+      switch (col.size()) {
+        case 1:
+          fst->SetFinal(s, Weight::One());
+          break;
+        case 2:
+          if (!StrToWeight(col[1], true, &w))
+            ok = false;
+          else
+            fst->SetFinal(s, w);
+          break;
+        case 3:  // 3 columns not ok for Lattice format; it's not an acceptor.
+          ok = false;
+          break;
+        case 4:
+          ok = ConvertStringToInteger(col[1], &arc.nextstate) &&
+               ConvertStringToInteger(col[2], &arc.ilabel) &&
+               ConvertStringToInteger(col[3], &arc.olabel);
+          if (ok) {
+            d = arc.nextstate;
+            arc.weight = Weight::One();
+            fst->AddArc(s, arc);
+          }
+          break;
+        case 5:
+          ok = ConvertStringToInteger(col[1], &arc.nextstate) &&
+               ConvertStringToInteger(col[2], &arc.ilabel) &&
+               ConvertStringToInteger(col[3], &arc.olabel) &&
+               StrToWeight(col[4], false, &arc.weight);
+          if (ok) {
+            d = arc.nextstate;
+            fst->AddArc(s, arc);
+          }
+          break;
+        default:
+          ok = false;
+      }
+      while (d >= fst->NumStates()) fst->AddState();
+      if (!ok) KALDI_ERR << "Bad line in FST: " << line;
+    }
+  }
+}
+
+template <class Arc>  // static
+bool VectorFstTplHolder<Arc>::Write(std::ostream &os, bool binary, const T &t) {
+  try {
+    WriteFstKaldi(os, binary, t);
+    return true;
+  } catch (...) {
+    return false;
+  }
+}
+
+template <class Arc>  // static
+bool VectorFstTplHolder<Arc>::Read(std::istream &is) {
+  Clear();
+  int c = is.peek();
+  if (c == -1) {
+    KALDI_WARN << "End of stream detected reading Fst";
+    return false;
+  } else if (isspace(c)) {  // The text form of the FST begins
+    // with space (normally, '\n'), so this means it's text (the binary form
+    // cannot begin with space because it starts with the FST Type() which is
+    // not space).
+    try {
+      t_ = new VectorFst<Arc>();
+      ReadFstKaldi(is, false, t_);
+    } catch (...) {
+      Clear();
+      return false;
+    }
+  } else {  // reading a binary FST.
+    try {
+      t_ = new VectorFst<Arc>();
+      ReadFstKaldi(is, true, t_);
+    } catch (...) {
+      Clear();
+      return false;
+    }
+  }
+  return true;
+}
+
+}  // namespace fst.
+
+#endif  // KALDI_FSTEXT_KALDI_FST_IO_INL_H_
diff --git a/speechx/speechx/kaldi/fstext/kaldi-fst-io.cc b/speechx/speechx/kaldi/fstext/kaldi-fst-io.cc
new file mode 100644
index 00000000..bd919a62
--- /dev/null
+++ b/speechx/speechx/kaldi/fstext/kaldi-fst-io.cc
@@ -0,0 +1,148 @@
+// fstext/kaldi-fst-io.cc
+
+// Copyright 2009-2011  Microsoft Corporation
+//           2012-2015  Johns Hopkins University (Author: Daniel Povey)
+//                2013  Guoguo Chen
+
+// See ../../COPYING for clarification regarding multiple authors
+//
+// Licensed under the Apache License, Version 2.0 (the "License");
+// you may not use this file except in compliance with the License.
+// You may obtain a copy of the License at
+//
+//  http://www.apache.org/licenses/LICENSE-2.0
+//
+// THIS CODE IS PROVIDED *AS IS* BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY
+// KIND, EITHER EXPRESS OR IMPLIED, INCLUDING WITHOUT LIMITATION ANY IMPLIED
+// WARRANTIES OR CONDITIONS OF TITLE, FITNESS FOR A PARTICULAR PURPOSE,
+// MERCHANTABLITY OR NON-INFRINGEMENT.
+// See the Apache 2 License for the specific language governing permissions and
+// limitations under the License.
+
+#include "fstext/kaldi-fst-io.h"
+
+#include <string>
+
+#include "base/kaldi-error.h"
+#include "base/kaldi-math.h"
+#include "util/kaldi-io.h"
+
+namespace fst {
+
+VectorFst<StdArc> *ReadFstKaldi(std::string rxfilename) {
+  if (rxfilename == "") rxfilename = "-";  // interpret "" as stdin,
+  // for compatibility with OpenFst conventions.
+  kaldi::Input ki(rxfilename);
+  fst::FstHeader hdr;
+  if (!hdr.Read(ki.Stream(), rxfilename))
+    KALDI_ERR << "Reading FST: error reading FST header from "
+              << kaldi::PrintableRxfilename(rxfilename);
+  FstReadOptions ropts("<unspecified>", &hdr);
+  VectorFst<StdArc> *fst = VectorFst<StdArc>::Read(ki.Stream(), ropts);
+  if (!fst)
+    KALDI_ERR << "Could not read fst from "
+              << kaldi::PrintableRxfilename(rxfilename);
+  return fst;
+}
+
+// Register const fst to load it automatically. Other types like
+// olabel_lookahead or ngram or compact_fst should be registered
+// through OpenFst registration API.
+static fst::FstRegisterer<VectorFst<StdArc>> VectorFst_StdArc_registerer;
+static fst::FstRegisterer<ConstFst<StdArc>> ConstFst_StdArc_registerer;
+
+Fst<StdArc> *ReadFstKaldiGeneric(std::string rxfilename, bool throw_on_err) {
+  if (rxfilename == "") rxfilename = "-";  // interpret "" as stdin,
+  // for compatibility with OpenFst conventions.
+  kaldi::Input ki(rxfilename);
+  fst::FstHeader hdr;
+  // Read FstHeader which contains the type of FST
+  if (!hdr.Read(ki.Stream(), rxfilename)) {
+    if (throw_on_err) {
+      KALDI_ERR << "Reading FST: error reading FST header from "
+                << kaldi::PrintableRxfilename(rxfilename);
+    } else {
+      KALDI_WARN << "We fail to read FST header from "
+                 << kaldi::PrintableRxfilename(rxfilename)
+                 << ". A NULL pointer is returned.";
+      return NULL;
+    }
+  }
+  // Check the type of Arc
+  if (hdr.ArcType() != fst::StdArc::Type()) {
+    if (throw_on_err) {
+      KALDI_ERR << "FST with arc type " << hdr.ArcType()
+                << " is not supported.";
+    } else {
+      KALDI_WARN << "Fst with arc type" << hdr.ArcType()
+                 << " is not supported. A NULL pointer is returned.";
+      return NULL;
+    }
+  }
+  // Read the FST
+  FstReadOptions ropts("<unspecified>", &hdr);
+  Fst<StdArc> *fst = Fst<StdArc>::Read(ki.Stream(), ropts);
+  if (!fst) {
+    if (throw_on_err) {
+      KALDI_ERR << "Could not read fst from "
+                << kaldi::PrintableRxfilename(rxfilename);
+    } else {
+      KALDI_WARN << "Could not read fst from "
+                 << kaldi::PrintableRxfilename(rxfilename)
+                 << ". A NULL pointer is returned.";
+      return NULL;
+    }
+  }
+  return fst;
+}
+
+VectorFst<StdArc> *CastOrConvertToVectorFst(Fst<StdArc> *fst) {
+  // This version currently supports ConstFst<StdArc> or VectorFst<StdArc>
+  std::string real_type = fst->Type();
+  KALDI_ASSERT(real_type == "vector" || real_type == "const");
+  if (real_type == "vector") {
+    return dynamic_cast<VectorFst<StdArc> *>(fst);
+  } else {
+    // As the 'fst' can't cast to VectorFst, we create a new
+    // VectorFst<StdArc> initialized by 'fst', and delete 'fst'.
+    VectorFst<StdArc> *new_fst = new VectorFst<StdArc>(*fst);
+    delete fst;
+    return new_fst;
+  }
+}
+
+void ReadFstKaldi(std::string rxfilename, fst::StdVectorFst *ofst) {
+  fst::StdVectorFst *fst = ReadFstKaldi(rxfilename);
+  *ofst = *fst;
+  delete fst;
+}
+
+void WriteFstKaldi(const VectorFst<StdArc> &fst, std::string wxfilename) {
+  if (wxfilename == "") wxfilename = "-";  // interpret "" as stdout,
+  // for compatibility with OpenFst conventions.
+  bool write_binary = true, write_header = false;
+  kaldi::Output ko(wxfilename, write_binary, write_header);
+  FstWriteOptions wopts(kaldi::PrintableWxfilename(wxfilename));
+  fst.Write(ko.Stream(), wopts);
+}
+
+fst::VectorFst<fst::StdArc> *ReadAndPrepareLmFst(std::string rxfilename) {
+  // ReadFstKaldi() will die with exception on failure.
+  fst::VectorFst<fst::StdArc> *ans = fst::ReadFstKaldi(rxfilename);
+  if (ans->Properties(fst::kAcceptor, true) == 0) {
+    // If it's not already an acceptor, project on the output, i.e. copy olabels
+    // to ilabels.  Generally the G.fst's on disk will have the disambiguation
+    // symbol #0 on the input symbols of the backoff arc, and projection will
+    // replace them with epsilons which is what is on the output symbols of
+    // those arcs.
+    fst::Project(ans, fst::PROJECT_OUTPUT);
+  }
+  if (ans->Properties(fst::kILabelSorted, true) == 0) {
+    // Make sure LM is sorted on ilabel.
+    fst::ILabelCompare<fst::StdArc> ilabel_comp;
+    fst::ArcSort(ans, ilabel_comp);
+  }
+  return ans;
+}
+
+}  // end namespace fst
diff --git a/speechx/speechx/kaldi/fstext/kaldi-fst-io.h b/speechx/speechx/kaldi/fstext/kaldi-fst-io.h
new file mode 100644
index 00000000..4938aa48
--- /dev/null
+++ b/speechx/speechx/kaldi/fstext/kaldi-fst-io.h
@@ -0,0 +1,158 @@
+// fstext/kaldi-fst-io.h
+
+// Copyright 2009-2011  Microsoft Corporation
+//           2012-2015  Johns Hopkins University (Author: Daniel Povey)
+//                2013  Guoguo Chen
+
+// See ../../COPYING for clarification regarding multiple authors
+//
+// Licensed under the Apache License, Version 2.0 (the "License");
+// you may not use this file except in compliance with the License.
+// You may obtain a copy of the License at
+//
+//  http://www.apache.org/licenses/LICENSE-2.0
+//
+// THIS CODE IS PROVIDED *AS IS* BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY
+// KIND, EITHER EXPRESS OR IMPLIED, INCLUDING WITHOUT LIMITATION ANY IMPLIED
+// WARRANTIES OR CONDITIONS OF TITLE, FITNESS FOR A PARTICULAR PURPOSE,
+// MERCHANTABLITY OR NON-INFRINGEMENT.
+// See the Apache 2 License for the specific language governing permissions and
+// limitations under the License.
+
+#ifndef KALDI_FSTEXT_KALDI_FST_IO_H_
+#define KALDI_FSTEXT_KALDI_FST_IO_H_
+
+#include <string>
+#include <utility>
+
+#include "fst/fst-decl.h"
+#include "fst/fstlib.h"
+#include "fst/script/print-impl.h"
+
+#include "base/kaldi-common.h"
+
+// Some functions for writing Fsts.
+// I/O for FSTs is a bit of a mess, and not very well integrated with Kaldi's
+// generic I/O mechanisms, because we want files containing just FSTs to
+// be readable by OpenFST's native binaries, which is not compatible
+// with the normal \0B header that identifies Kaldi files as containing
+// binary data.
+// So use the functions here with your eyes open, and with caution!
+namespace fst {
+
+// Read a binary FST using Kaldi I/O mechanisms (pipes, etc.)
+// On error returns NULL. Only supports VectorFst and exists
+// mainly for backward code compabibility.
+VectorFst<StdArc> *ReadFstKaldi(std::string rxfilename);
+
+// Read a binary FST using Kaldi I/O mechanisms (pipes, etc.)
+// If it can't read the FST, if throw_on_err == true it throws using KALDI_ERR;
+// otherwise it prints a warning and returns. Note:this
+// doesn't support the text-mode option that we generally like to support.
+// This version currently supports ConstFst<StdArc> or VectorFst<StdArc>
+// (const-fst can give better performance for decoding). Other
+// types could be also loaded if registered inside OpenFst.
+Fst<StdArc> *ReadFstKaldiGeneric(std::string rxfilename,
+                                 bool throw_on_err = true);
+
+// This function attempts to dynamic_cast the pointer 'fst' (which will likely
+// have been returned by ReadFstGeneric()), to the more derived
+// type VectorFst<StdArc>. If this succeeds, it returns the same pointer;
+// if it fails, it converts the FST type (by creating a new VectorFst<stdArc>
+// initialized by 'fst'), prints a warning, and deletes 'fst'.
+VectorFst<StdArc> *CastOrConvertToVectorFst(Fst<StdArc> *fst);
+
+// Version of ReadFstKaldi() that writes to a pointer.  Assumes
+// the FST is binary with no binary marker.  Crashes on error.
+void ReadFstKaldi(std::string rxfilename, VectorFst<StdArc> *ofst);
+
+// Write an FST using Kaldi I/O mechanisms (pipes, etc.)
+// On error, throws using KALDI_ERR.  For use only in code in fstbin/,
+// as it doesn't support the text-mode option.
+void WriteFstKaldi(const VectorFst<StdArc> &fst, std::string wxfilename);
+
+// This is a more general Kaldi-type-IO mechanism of writing FSTs to
+// streams, supporting binary or text-mode writing.  (note: we just
+// write the integers, symbol tables are not supported).
+// On error, throws using KALDI_ERR.
+template <class Arc>
+void WriteFstKaldi(std::ostream &os, bool binary, const VectorFst<Arc> &fst);
+
+// A generic Kaldi-type-IO mechanism of reading FSTs from streams,
+// supporting binary or text-mode reading/writing.
+template <class Arc>
+void ReadFstKaldi(std::istream &is, bool binary, VectorFst<Arc> *fst);
+
+// Read an FST file for LM (G.fst) and make it an acceptor,
+// and make sure it is sorted on labels
+fst::VectorFst<fst::StdArc> *ReadAndPrepareLmFst(std::string rxfilename);
+
+// This is a Holder class with T = VectorFst<Arc>, that meets the requirements
+// of a Holder class as described in ../util/kaldi-holder.h. This enables us to
+// read/write collections of FSTs indexed by strings, using the Table concept (
+// see ../util/kaldi-table.h).
+// Originally it was only templated on T = VectorFst<StdArc>, but as the keyword
+// spotting stuff introduced more types of FSTs, we made it also templated on
+// the arc.
+template <class Arc>
+class VectorFstTplHolder {
+ public:
+  typedef VectorFst<Arc> T;
+
+  VectorFstTplHolder() : t_(NULL) {}
+
+  static bool Write(std::ostream &os, bool binary, const T &t);
+
+  void Copy(const T &t) {  // copies it into the holder.
+    Clear();
+    t_ = new T(t);
+  }
+
+  // Reads into the holder.
+  bool Read(std::istream &is);
+
+  // It's potentially a binary format, so must read in binary mode (linefeed
+  // translation will corrupt the file.  We don't know till we open the file if
+  // it's really binary, so we need to read in binary mode to be on the safe
+  // side.  Extra linefeeds won't matter, the text-mode reading code ignores
+  // them.
+  static bool IsReadInBinary() { return true; }
+
+  T &Value() {
+    // code error if !t_.
+    if (!t_) KALDI_ERR << "VectorFstTplHolder::Value() called wrongly.";
+    return *t_;
+  }
+
+  void Clear() {
+    if (t_) {
+      delete t_;
+      t_ = NULL;
+    }
+  }
+
+  void Swap(VectorFstTplHolder<Arc> *other) { std::swap(t_, other->t_); }
+
+  bool ExtractRange(const VectorFstTplHolder<Arc> &other,
+                    const std::string &range) {
+    KALDI_ERR << "ExtractRange is not defined for this type of holder.";
+    return false;
+  }
+
+  ~VectorFstTplHolder() { Clear(); }
+  // No destructor.  Assignment and
+  // copy constructor take their default implementations.
+ private:
+  KALDI_DISALLOW_COPY_AND_ASSIGN(VectorFstTplHolder);
+  T *t_;
+};
+
+// Now make the original VectorFstHolder as the typedef of
+// VectorFstHolder<StdArc>.
+typedef VectorFstTplHolder<StdArc> VectorFstHolder;
+
+}  // end namespace fst
+
+#include "fstext/kaldi-fst-io-inl.h"
+
+#endif  // KALDI_FSTEXT_KALDI_FST_IO_H_
diff --git a/speechx/speechx/kaldi/fstext/lattice-utils-inl.h b/speechx/speechx/kaldi/fstext/lattice-utils-inl.h
new file mode 100644
index 00000000..82e5f066
--- /dev/null
+++ b/speechx/speechx/kaldi/fstext/lattice-utils-inl.h
@@ -0,0 +1,267 @@
+// fstext/lattice-utils-inl.h
+
+// Copyright 2009-2012  Microsoft Corporation  Johns Hopkins University (Author:
+// Daniel Povey)
+
+// See ../../COPYING for clarification regarding multiple authors
+//
+// Licensed under the Apache License, Version 2.0 (the "License");
+// you may not use this file except in compliance with the License.
+// You may obtain a copy of the License at
+//
+//  http://www.apache.org/licenses/LICENSE-2.0
+//
+// THIS CODE IS PROVIDED *AS IS* BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY
+// KIND, EITHER EXPRESS OR IMPLIED, INCLUDING WITHOUT LIMITATION ANY IMPLIED
+// WARRANTIES OR CONDITIONS OF TITLE, FITNESS FOR A PARTICULAR PURPOSE,
+// MERCHANTABLITY OR NON-INFRINGEMENT.
+// See the Apache 2 License for the specific language governing permissions and
+// limitations under the License.
+
+#ifndef KALDI_FSTEXT_LATTICE_UTILS_INL_H_
+#define KALDI_FSTEXT_LATTICE_UTILS_INL_H_
+// Do not include this file directly.  It is included by lattice-utils.h
+
+#include <utility>
+#include <vector>
+
+namespace fst {
+
+/* Convert from FST with arc-type Weight, to one with arc-type
+   CompactLatticeWeight.  Uses FactorFst to identify chains
+   of states which can be turned into a single output arc. */
+
+template <class Weight, class Int>
+void ConvertLattice(
+    const ExpandedFst<ArcTpl<Weight> > &ifst,
+    MutableFst<ArcTpl<CompactLatticeWeightTpl<Weight, Int> > > *ofst,
+    bool invert) {
+  typedef ArcTpl<Weight> Arc;
+  typedef typename Arc::StateId StateId;
+  typedef CompactLatticeWeightTpl<Weight, Int> CompactWeight;
+  typedef ArcTpl<CompactWeight> CompactArc;
+
+  VectorFst<ArcTpl<Weight> > ffst;
+  std::vector<std::vector<Int> > labels;
+  if (invert) {  // normal case: want the ilabels as sequences on the arcs of
+    Factor(ifst, &ffst, &labels);  // the output... Factor makes seqs of
+                                   // ilabels.
+  } else {
+    VectorFst<ArcTpl<Weight> > invfst(ifst);
+    Invert(&invfst);
+    Factor(invfst, &ffst, &labels);
+  }
+
+  TopSort(&ffst);  // Put the states in ffst in topological order, which is
+  // easier on the eye when reading the text-form lattices and corresponds to
+  // what we get when we generate the lattices in the decoder.
+
+  ofst->DeleteStates();
+
+  // The states will be numbered exactly the same as the original FST.
+  // Add the states to the new FST.
+  StateId num_states = ffst.NumStates();
+  for (StateId s = 0; s < num_states; s++) {
+    StateId news = ofst->AddState();
+    assert(news == s);
+  }
+  ofst->SetStart(ffst.Start());
+  for (StateId s = 0; s < num_states; s++) {
+    Weight final_weight = ffst.Final(s);
+    if (final_weight != Weight::Zero()) {
+      CompactWeight final_compact_weight(final_weight, std::vector<Int>());
+      ofst->SetFinal(s, final_compact_weight);
+    }
+    for (ArcIterator<ExpandedFst<Arc> > iter(ffst, s); !iter.Done();
+         iter.Next()) {
+      const Arc &arc = iter.Value();
+      KALDI_PARANOID_ASSERT(arc.weight != Weight::Zero());
+      // note: zero-weight arcs not allowed anyway so weight should not be zero,
+      // but no harm in checking.
+      CompactArc compact_arc(arc.olabel, arc.olabel,
+                             CompactWeight(arc.weight, labels[arc.ilabel]),
+                             arc.nextstate);
+      ofst->AddArc(s, compact_arc);
+    }
+  }
+}
+
+template <class Weight, class Int>
+void ConvertLattice(
+    const ExpandedFst<ArcTpl<CompactLatticeWeightTpl<Weight, Int> > > &ifst,
+    MutableFst<ArcTpl<Weight> > *ofst, bool invert) {
+  typedef ArcTpl<Weight> Arc;
+  typedef typename Arc::StateId StateId;
+  typedef typename Arc::Label Label;
+  typedef CompactLatticeWeightTpl<Weight, Int> CompactWeight;
+  typedef ArcTpl<CompactWeight> CompactArc;
+  ofst->DeleteStates();
+  // make the states in the new FST have the same numbers as
+  // the original ones, and add chains of states as necessary
+  // to encode the string-valued weights.
+  StateId num_states = ifst.NumStates();
+  for (StateId s = 0; s < num_states; s++) {
+    StateId news = ofst->AddState();
+    assert(news == s);
+  }
+  ofst->SetStart(ifst.Start());
+  for (StateId s = 0; s < num_states; s++) {
+    CompactWeight final_weight = ifst.Final(s);
+    if (final_weight != CompactWeight::Zero()) {
+      StateId cur_state = s;
+      size_t string_length = final_weight.String().size();
+      for (size_t n = 0; n < string_length; n++) {
+        StateId next_state = ofst->AddState();
+        Label ilabel = 0;
+        Arc arc(ilabel, final_weight.String()[n],
+                (n == 0 ? final_weight.Weight() : Weight::One()), next_state);
+        if (invert) std::swap(arc.ilabel, arc.olabel);
+        ofst->AddArc(cur_state, arc);
+        cur_state = next_state;
+      }
+      ofst->SetFinal(cur_state,
+                     string_length > 0 ? Weight::One() : final_weight.Weight());
+    }
+    for (ArcIterator<ExpandedFst<CompactArc> > iter(ifst, s); !iter.Done();
+         iter.Next()) {
+      const CompactArc &arc = iter.Value();
+      size_t string_length = arc.weight.String().size();
+      StateId cur_state = s;
+      // for all but the last element in the string--
+      // add a temporary state.
+      for (size_t n = 0; n + 1 < string_length; n++) {
+        StateId next_state = ofst->AddState();
+        Label ilabel = (n == 0 ? arc.ilabel : 0),
+              olabel = static_cast<Label>(arc.weight.String()[n]);
+        Weight weight = (n == 0 ? arc.weight.Weight() : Weight::One());
+        Arc new_arc(ilabel, olabel, weight, next_state);
+        if (invert) std::swap(new_arc.ilabel, new_arc.olabel);
+        ofst->AddArc(cur_state, new_arc);
+        cur_state = next_state;
+      }
+      Label ilabel = (string_length <= 1 ? arc.ilabel : 0),
+            olabel = (string_length > 0 ? arc.weight.String()[string_length - 1]
+                                        : 0);
+      Weight weight =
+          (string_length <= 1 ? arc.weight.Weight() : Weight::One());
+      Arc new_arc(ilabel, olabel, weight, arc.nextstate);
+      if (invert) std::swap(new_arc.ilabel, new_arc.olabel);
+      ofst->AddArc(cur_state, new_arc);
+    }
+  }
+}
+
+// This function converts lattices between float and double;
+// it works for both CompactLatticeWeight and LatticeWeight.
+template <class WeightIn, class WeightOut>
+void ConvertLattice(const ExpandedFst<ArcTpl<WeightIn> > &ifst,
+                    MutableFst<ArcTpl<WeightOut> > *ofst) {
+  typedef ArcTpl<WeightIn> ArcIn;
+  typedef ArcTpl<WeightOut> ArcOut;
+  typedef typename ArcIn::StateId StateId;
+  ofst->DeleteStates();
+  // The states will be numbered exactly the same as the original FST.
+  // Add the states to the new FST.
+  StateId num_states = ifst.NumStates();
+  for (StateId s = 0; s < num_states; s++) {
+    StateId news = ofst->AddState();
+    assert(news == s);
+  }
+  ofst->SetStart(ifst.Start());
+  for (StateId s = 0; s < num_states; s++) {
+    WeightIn final_iweight = ifst.Final(s);
+    if (final_iweight != WeightIn::Zero()) {
+      WeightOut final_oweight;
+      ConvertLatticeWeight(final_iweight, &final_oweight);
+      ofst->SetFinal(s, final_oweight);
+    }
+    for (ArcIterator<ExpandedFst<ArcIn> > iter(ifst, s); !iter.Done();
+         iter.Next()) {
+      ArcIn arc = iter.Value();
+      KALDI_PARANOID_ASSERT(arc.weight != WeightIn::Zero());
+      ArcOut oarc;
+      ConvertLatticeWeight(arc.weight, &oarc.weight);
+      oarc.ilabel = arc.ilabel;
+      oarc.olabel = arc.olabel;
+      oarc.nextstate = arc.nextstate;
+      ofst->AddArc(s, oarc);
+    }
+  }
+}
+
+template <class Weight, class ScaleFloat>
+void ScaleLattice(const std::vector<std::vector<ScaleFloat> > &scale,
+                  MutableFst<ArcTpl<Weight> > *fst) {
+  assert(scale.size() == 2 && scale[0].size() == 2 && scale[1].size() == 2);
+  if (scale == DefaultLatticeScale())  // nothing to do.
+    return;
+  typedef ArcTpl<Weight> Arc;
+  typedef MutableFst<Arc> Fst;
+  typedef typename Arc::StateId StateId;
+  StateId num_states = fst->NumStates();
+  for (StateId s = 0; s < num_states; s++) {
+    for (MutableArcIterator<Fst> aiter(fst, s); !aiter.Done(); aiter.Next()) {
+      Arc arc = aiter.Value();
+      arc.weight = Weight(ScaleTupleWeight(arc.weight, scale));
+      aiter.SetValue(arc);
+    }
+    Weight final_weight = fst->Final(s);
+    if (final_weight != Weight::Zero())
+      fst->SetFinal(s, Weight(ScaleTupleWeight(final_weight, scale)));
+  }
+}
+
+template <class Weight, class Int>
+void RemoveAlignmentsFromCompactLattice(
+    MutableFst<ArcTpl<CompactLatticeWeightTpl<Weight, Int> > > *fst) {
+  typedef CompactLatticeWeightTpl<Weight, Int> W;
+  typedef ArcTpl<W> Arc;
+  typedef MutableFst<Arc> Fst;
+  typedef typename Arc::StateId StateId;
+  StateId num_states = fst->NumStates();
+  for (StateId s = 0; s < num_states; s++) {
+    for (MutableArcIterator<Fst> aiter(fst, s); !aiter.Done(); aiter.Next()) {
+      Arc arc = aiter.Value();
+      arc.weight = W(arc.weight.Weight(), std::vector<Int>());
+      aiter.SetValue(arc);
+    }
+    W final_weight = fst->Final(s);
+    if (final_weight != W::Zero())
+      fst->SetFinal(s, W(final_weight.Weight(), std::vector<Int>()));
+  }
+}
+
+template <class Weight, class Int>
+bool CompactLatticeHasAlignment(
+    const ExpandedFst<ArcTpl<CompactLatticeWeightTpl<Weight, Int> > > &fst) {
+  typedef CompactLatticeWeightTpl<Weight, Int> W;
+  typedef ArcTpl<W> Arc;
+  typedef ExpandedFst<Arc> Fst;
+  typedef typename Arc::StateId StateId;
+  StateId num_states = fst.NumStates();
+  for (StateId s = 0; s < num_states; s++) {
+    for (ArcIterator<Fst> aiter(fst, s); !aiter.Done(); aiter.Next()) {
+      const Arc &arc = aiter.Value();
+      if (!arc.weight.String().empty()) return true;
+    }
+    W final_weight = fst.Final(s);
+    if (!final_weight.String().empty()) return true;
+  }
+  return false;
+}
+
+template <class Real>
+void ConvertFstToLattice(const ExpandedFst<ArcTpl<TropicalWeight> > &ifst,
+                         MutableFst<ArcTpl<LatticeWeightTpl<Real> > > *ofst) {
+  int32 num_states_cache = 50000;
+  fst::CacheOptions cache_opts(true, num_states_cache);
+  fst::MapFstOptions mapfst_opts(cache_opts);
+  StdToLatticeMapper<Real> mapper;
+  MapFst<StdArc, ArcTpl<LatticeWeightTpl<Real> >, StdToLatticeMapper<Real> >
+      map_fst(ifst, mapper, mapfst_opts);
+  *ofst = map_fst;
+}
+
+}  // namespace fst
+
+#endif  // KALDI_FSTEXT_LATTICE_UTILS_INL_H_
diff --git a/speechx/speechx/kaldi/fstext/lattice-utils.h b/speechx/speechx/kaldi/fstext/lattice-utils.h
new file mode 100644
index 00000000..72f23a3c
--- /dev/null
+++ b/speechx/speechx/kaldi/fstext/lattice-utils.h
@@ -0,0 +1,259 @@
+// fstext/lattice-utils.h
+
+// Copyright 2009-2011  Microsoft Corporation
+
+// See ../../COPYING for clarification regarding multiple authors
+//
+// Licensed under the Apache License, Version 2.0 (the "License");
+// you may not use this file except in compliance with the License.
+// You may obtain a copy of the License at
+//
+//  http://www.apache.org/licenses/LICENSE-2.0
+//
+// THIS CODE IS PROVIDED *AS IS* BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY
+// KIND, EITHER EXPRESS OR IMPLIED, INCLUDING WITHOUT LIMITATION ANY IMPLIED
+// WARRANTIES OR CONDITIONS OF TITLE, FITNESS FOR A PARTICULAR PURPOSE,
+// MERCHANTABLITY OR NON-INFRINGEMENT.
+// See the Apache 2 License for the specific language governing permissions and
+// limitations under the License.
+
+#ifndef KALDI_FSTEXT_LATTICE_UTILS_H_
+#define KALDI_FSTEXT_LATTICE_UTILS_H_
+
+#include <vector>
+
+#include "fst/fstlib.h"
+#include "fstext/lattice-weight.h"
+
+namespace fst {
+
+// The template ConvertLattice does conversions to and from
+// LatticeWeight FSTs and CompactLatticeWeight FSTs, and
+// between float and double, and to convert from LatticeWeight
+// to TropicalWeight.  It's used in the I/O code for lattices,
+// and for converting lattices to standard FSTs (e.g. for creating
+// decoding graphs from lattices).
+
+/**
+   Convert lattice from a normal FST to a CompactLattice FST.
+   This is a bit like converting to the Gallic semiring, except
+   the semiring behaves in a different way (designed to take
+   the best path).
+   Note: the ilabels end up as the symbols on the arcs of the
+   output acceptor, and the olabels go to the strings.  To make
+   it the other way around (useful for the speech-recognition
+   application), set invert=true [the default].
+*/
+template <class Weight, class Int>
+void ConvertLattice(
+    const ExpandedFst<ArcTpl<Weight> > &ifst,
+    MutableFst<ArcTpl<CompactLatticeWeightTpl<Weight, Int> > > *ofst,
+    bool invert = true);
+
+/**
+   Convert lattice CompactLattice  format to Lattice.  This is a bit
+   like converting from the Gallic semiring.  As for any CompactLattice, "ifst"
+   must be an acceptor (i.e., ilabels and olabels should be identical).  If
+   invert=false, the labels on "ifst" become the ilabels on "ofst" and the
+   strings in the weights of "ifst" becomes the olabels.  If invert=true
+   [default], this is reversed (useful for speech recognition lattices; our
+   standard non-compact format has the words on the output side to match HCLG).
+   */
+template <class Weight, class Int>
+void ConvertLattice(
+    const ExpandedFst<ArcTpl<CompactLatticeWeightTpl<Weight, Int> > > &ifst,
+    MutableFst<ArcTpl<Weight> > *ofst, bool invert = true);
+
+/**
+  Convert between CompactLattices and Lattices of different floating point
+  types... this works between any pair of weight types for which
+  ConvertLatticeWeight is defined (c.f. lattice-weight.h), and also includes
+  conversion from LatticeWeight to TropicalWeight.
+ */
+template <class WeightIn, class WeightOut>
+void ConvertLattice(const ExpandedFst<ArcTpl<WeightIn> > &ifst,
+                    MutableFst<ArcTpl<WeightOut> > *ofst);
+
+// Now define some ConvertLattice functions that require two phases of
+// conversion (don't bother coding these separately as they will be used rarely.
+
+// Lattice with float to CompactLattice with double.
+template <class Int>
+void ConvertLattice(
+    const ExpandedFst<ArcTpl<LatticeWeightTpl<float> > > &ifst,
+    MutableFst<ArcTpl<CompactLatticeWeightTpl<LatticeWeightTpl<double>, Int> > >
+        *ofst) {
+  VectorFst<ArcTpl<CompactLatticeWeightTpl<LatticeWeightTpl<float>, Int> > >
+      fst;
+  ConvertLattice(ifst, &fst);
+  ConvertLattice(fst, ofst);
+}
+
+// Lattice with double to CompactLattice with float.
+template <class Int>
+void ConvertLattice(
+    const ExpandedFst<ArcTpl<LatticeWeightTpl<double> > > &ifst,
+    MutableFst<ArcTpl<CompactLatticeWeightTpl<LatticeWeightTpl<float>, Int> > >
+        *ofst) {
+  VectorFst<ArcTpl<CompactLatticeWeightTpl<LatticeWeightTpl<double>, Int> > >
+      fst;
+  ConvertLattice(ifst, &fst);
+  ConvertLattice(fst, ofst);
+}
+
+/// Converts CompactLattice with double to Lattice with float.
+template <class Int>
+void ConvertLattice(
+    const ExpandedFst<
+        ArcTpl<CompactLatticeWeightTpl<LatticeWeightTpl<double>, Int> > > &ifst,
+    MutableFst<ArcTpl<LatticeWeightTpl<float> > > *ofst) {
+  VectorFst<ArcTpl<CompactLatticeWeightTpl<LatticeWeightTpl<float>, Int> > >
+      fst;
+  ConvertLattice(ifst, &fst);
+  ConvertLattice(fst, ofst);
+}
+
+/// Converts CompactLattice with float to Lattice with double.
+template <class Int>
+void ConvertLattice(
+    const ExpandedFst<
+        ArcTpl<CompactLatticeWeightTpl<LatticeWeightTpl<float>, Int> > > &ifst,
+    MutableFst<ArcTpl<LatticeWeightTpl<double> > > *ofst) {
+  VectorFst<ArcTpl<CompactLatticeWeightTpl<LatticeWeightTpl<double>, Int> > >
+      fst;
+  ConvertLattice(ifst, &fst);
+  ConvertLattice(fst, ofst);
+}
+
+/// Converts TropicalWeight to LatticeWeight (puts all the weight on
+/// the first float in the lattice's pair).
+template <class Real>
+void ConvertFstToLattice(const ExpandedFst<ArcTpl<TropicalWeight> > &ifst,
+                         MutableFst<ArcTpl<LatticeWeightTpl<Real> > > *ofst);
+
+/** Returns a default 2x2 matrix scaling factor for LatticeWeight */
+inline std::vector<std::vector<double> > DefaultLatticeScale() {
+  std::vector<std::vector<double> > ans(2);
+  ans[0].resize(2, 0.0);
+  ans[1].resize(2, 0.0);
+  ans[0][0] = ans[1][1] = 1.0;
+  return ans;
+}
+
+inline std::vector<std::vector<double> > AcousticLatticeScale(double acwt) {
+  std::vector<std::vector<double> > ans(2);
+  ans[0].resize(2, 0.0);
+  ans[1].resize(2, 0.0);
+  ans[0][0] = 1.0;
+  ans[1][1] = acwt;
+  return ans;
+}
+
+inline std::vector<std::vector<double> > GraphLatticeScale(double lmwt) {
+  std::vector<std::vector<double> > ans(2);
+  ans[0].resize(2, 0.0);
+  ans[1].resize(2, 0.0);
+  ans[0][0] = lmwt;
+  ans[1][1] = 1.0;
+  return ans;
+}
+
+inline std::vector<std::vector<double> > LatticeScale(double lmwt,
+                                                      double acwt) {
+  std::vector<std::vector<double> > ans(2);
+  ans[0].resize(2, 0.0);
+  ans[1].resize(2, 0.0);
+  ans[0][0] = lmwt;
+  ans[1][1] = acwt;
+  return ans;
+}
+
+/** Scales the pairs of weights in LatticeWeight or CompactLatticeWeight by
+    viewing the pair (a, b) as a 2-vector and pre-multiplying by the 2x2 matrix
+    in "scale".  E.g. typically scale would equal
+     [ 1   0;
+       0  acwt ]
+    if we want to scale the acoustics by "acwt".
+ */
+template <class Weight, class ScaleFloat>
+void ScaleLattice(const std::vector<std::vector<ScaleFloat> > &scale,
+                  MutableFst<ArcTpl<Weight> > *fst);
+
+/// Removes state-level alignments (the strings that are
+/// part of the weights).
+template <class Weight, class Int>
+void RemoveAlignmentsFromCompactLattice(
+    MutableFst<ArcTpl<CompactLatticeWeightTpl<Weight, Int> > > *fst);
+
+/// Returns true if lattice has alignments, i.e. it has
+/// any nonempty strings inside its weights.
+template <class Weight, class Int>
+bool CompactLatticeHasAlignment(
+    const ExpandedFst<ArcTpl<CompactLatticeWeightTpl<Weight, Int> > > &fst);
+
+/// Class StdToLatticeMapper maps a normal arc (StdArc)
+/// to a LatticeArc by putting the StdArc weight as the first
+/// element of the LatticeWeight.  Useful when doing LM
+/// rescoring.
+template <class Real>
+class StdToLatticeMapper {
+  typedef LatticeWeightTpl<Real> LatticeWeight;
+  typedef ArcTpl<LatticeWeight> LatticeArc;
+
+ public:
+  LatticeArc operator()(const StdArc &arc) {
+    // Note: we have to check whether the arc's weight is zero below,
+    // and if so return (infinity, infinity) and not (infinity, zero),
+    // because (infinity, zero) is not a valid LatticeWeight, which should
+    // either be both finite, or both infinite (i.e. Zero()).
+    return LatticeArc(
+        arc.ilabel, arc.olabel,
+        LatticeWeight(arc.weight.Value(), arc.weight == StdArc::Weight::Zero()
+                                              ? arc.weight.Value()
+                                              : 0.0),
+        arc.nextstate);
+  }
+  MapFinalAction FinalAction() { return MAP_NO_SUPERFINAL; }
+
+  MapSymbolsAction InputSymbolsAction() { return MAP_COPY_SYMBOLS; }
+
+  MapSymbolsAction OutputSymbolsAction() { return MAP_COPY_SYMBOLS; }
+
+  // I believe all properties are preserved.
+  uint64 Properties(uint64 props) { return props; }
+};
+
+/// Class LatticeToStdMapper maps a LatticeArc to a normal arc (StdArc)
+/// by adding the elements of the LatticeArc weight.
+
+template <class Real>
+class LatticeToStdMapper {
+  typedef LatticeWeightTpl<Real> LatticeWeight;
+  typedef ArcTpl<LatticeWeight> LatticeArc;
+
+ public:
+  StdArc operator()(const LatticeArc &arc) {
+    return StdArc(arc.ilabel, arc.olabel,
+                  StdArc::Weight(arc.weight.Value1() + arc.weight.Value2()),
+                  arc.nextstate);
+  }
+  MapFinalAction FinalAction() { return MAP_NO_SUPERFINAL; }
+
+  MapSymbolsAction InputSymbolsAction() { return MAP_COPY_SYMBOLS; }
+
+  MapSymbolsAction OutputSymbolsAction() { return MAP_COPY_SYMBOLS; }
+
+  // I believe all properties are preserved.
+  uint64 Properties(uint64 props) { return props; }
+};
+
+template <class Weight, class Int>
+void PruneCompactLattice(
+    Weight beam,
+    MutableFst<ArcTpl<CompactLatticeWeightTpl<Weight, Int> > > *fst);
+
+}  // end namespace fst
+
+#include "fstext/lattice-utils-inl.h"
+
+#endif  // KALDI_FSTEXT_LATTICE_UTILS_H_
diff --git a/speechx/speechx/kaldi/fstext/lattice-weight.h b/speechx/speechx/kaldi/fstext/lattice-weight.h
new file mode 100644
index 00000000..57d746dd
--- /dev/null
+++ b/speechx/speechx/kaldi/fstext/lattice-weight.h
@@ -0,0 +1,892 @@
+// fstext/lattice-weight.h
+// Copyright 2009-2012  Microsoft Corporation
+//                      Johns Hopkins University (author: Daniel Povey)
+
+// See ../../COPYING for clarification regarding multiple authors
+//
+// Licensed under the Apache License, Version 2.0 (the "License");
+// you may not use this file except in compliance with the License.
+// You may obtain a copy of the License at
+//
+//  http://www.apache.org/licenses/LICENSE-2.0
+//
+// THIS CODE IS PROVIDED *AS IS* BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY
+// KIND, EITHER EXPRESS OR IMPLIED, INCLUDING WITHOUT LIMITATION ANY IMPLIED
+// WARRANTIES OR CONDITIONS OF TITLE, FITNESS FOR A PARTICULAR PURPOSE,
+// MERCHANTABLITY OR NON-INFRINGEMENT.
+// See the Apache 2 License for the specific language governing permissions and
+// limitations under the License.
+
+#ifndef KALDI_FSTEXT_LATTICE_WEIGHT_H_
+#define KALDI_FSTEXT_LATTICE_WEIGHT_H_
+
+#include <algorithm>
+#include <limits>
+#include <string>
+#include <vector>
+
+#include "base/kaldi-common.h"
+#include "fst/fstlib.h"
+
+namespace fst {
+
+// Declare weight type for lattice... will import to namespace kaldi.  has two
+// members, value1_ and value2_, of type BaseFloat (normally equals float).  It
+// is basically the same as the tropical semiring on value1_+value2_, except it
+// keeps track of a and b separately.  More precisely, it is equivalent to the
+// lexicographic semiring on (value1_+value2_), (value1_-value2_)
+
+template <class FloatType>
+class LatticeWeightTpl;
+
+template <class FloatType>
+inline std::ostream &operator<<(std::ostream &strm,
+                                const LatticeWeightTpl<FloatType> &w);
+
+template <class FloatType>
+inline std::istream &operator>>(std::istream &strm,
+                                LatticeWeightTpl<FloatType> &w);
+
+template <class FloatType>
+class LatticeWeightTpl {
+ public:
+  typedef FloatType T;  // normally float.
+  typedef LatticeWeightTpl ReverseWeight;
+
+  inline T Value1() const { return value1_; }
+
+  inline T Value2() const { return value2_; }
+
+  inline void SetValue1(T f) { value1_ = f; }
+
+  inline void SetValue2(T f) { value2_ = f; }
+
+  LatticeWeightTpl() : value1_{}, value2_{} {}
+
+  LatticeWeightTpl(T a, T b) : value1_(a), value2_(b) {}
+
+  LatticeWeightTpl(const LatticeWeightTpl &other)
+      : value1_(other.value1_), value2_(other.value2_) {}
+
+  LatticeWeightTpl &operator=(const LatticeWeightTpl &w) {
+    value1_ = w.value1_;
+    value2_ = w.value2_;
+    return *this;
+  }
+
+  LatticeWeightTpl<FloatType> Reverse() const { return *this; }
+
+  static const LatticeWeightTpl Zero() {
+    return LatticeWeightTpl(std::numeric_limits<T>::infinity(),
+                            std::numeric_limits<T>::infinity());
+  }
+
+  static const LatticeWeightTpl One() { return LatticeWeightTpl(0.0, 0.0); }
+
+  static const std::string &Type() {
+    static const std::string type = (sizeof(T) == 4 ? "lattice4" : "lattice8");
+    return type;
+  }
+
+  static const LatticeWeightTpl NoWeight() {
+    return LatticeWeightTpl(std::numeric_limits<FloatType>::quiet_NaN(),
+                            std::numeric_limits<FloatType>::quiet_NaN());
+  }
+
+  bool Member() const {
+    // value1_ == value1_ tests for NaN.
+    // also test for no -inf, and either both or neither
+    // must be +inf, and
+    if (value1_ != value1_ || value2_ != value2_) return false;  // NaN
+    if (value1_ == -std::numeric_limits<T>::infinity() ||
+        value2_ == -std::numeric_limits<T>::infinity())
+      return false;  // -infty not allowed
+    if (value1_ == std::numeric_limits<T>::infinity() ||
+        value2_ == std::numeric_limits<T>::infinity()) {
+      if (value1_ != std::numeric_limits<T>::infinity() ||
+          value2_ != std::numeric_limits<T>::infinity())
+        return false;  // both must be +infty;
+      // this is necessary so that the semiring has only one zero.
+    }
+    return true;
+  }
+
+  LatticeWeightTpl Quantize(float delta = kDelta) const {
+    if (value1_ + value2_ == -std::numeric_limits<T>::infinity()) {
+      return LatticeWeightTpl(-std::numeric_limits<T>::infinity(),
+                              -std::numeric_limits<T>::infinity());
+    } else if (value1_ + value2_ == std::numeric_limits<T>::infinity()) {
+      return LatticeWeightTpl(std::numeric_limits<T>::infinity(),
+                              std::numeric_limits<T>::infinity());
+    } else if (value1_ + value2_ != value1_ + value2_) {  // NaN
+      return LatticeWeightTpl(value1_ + value2_, value1_ + value2_);
+    } else {
+      return LatticeWeightTpl(floor(value1_ / delta + 0.5F) * delta,
+                              floor(value2_ / delta + 0.5F) * delta);
+    }
+  }
+  static constexpr uint64 Properties() {
+    return kLeftSemiring | kRightSemiring | kCommutative | kPath | kIdempotent;
+  }
+
+  // This is used in OpenFst for binary I/O.  This is OpenFst-style,
+  // not Kaldi-style, I/O.
+  std::istream &Read(std::istream &strm) {
+    // Always read/write as float, even if T is double,
+    // so we can use OpenFst-style read/write and still maintain
+    // compatibility when compiling with different FloatTypes
+    ReadType(strm, &value1_);
+    ReadType(strm, &value2_);
+    return strm;
+  }
+
+  // This is used in OpenFst for binary I/O.  This is OpenFst-style,
+  // not Kaldi-style, I/O.
+  std::ostream &Write(std::ostream &strm) const {
+    WriteType(strm, value1_);
+    WriteType(strm, value2_);
+    return strm;
+  }
+
+  size_t Hash() const {
+    size_t ans;
+    union {
+      T f;
+      size_t s;
+    } u;
+    u.s = 0;
+    u.f = value1_;
+    ans = u.s;
+    u.f = value2_;
+    ans += u.s;
+    return ans;
+  }
+
+ protected:
+  inline static void WriteFloatType(std::ostream &strm, const T &f) {
+    if (f == std::numeric_limits<T>::infinity())
+      strm << "Infinity";
+    else if (f == -std::numeric_limits<T>::infinity())
+      strm << "-Infinity";
+    else if (f != f)
+      strm << "BadNumber";
+    else
+      strm << f;
+  }
+
+  // Internal helper function, used in ReadNoParen.
+  inline static void ReadFloatType(std::istream &strm, T &f) {  // NOLINT
+    std::string s;
+    strm >> s;
+    if (s == "Infinity") {
+      f = std::numeric_limits<T>::infinity();
+    } else if (s == "-Infinity") {
+      f = -std::numeric_limits<T>::infinity();
+    } else if (s == "BadNumber") {
+      f = std::numeric_limits<T>::quiet_NaN();
+    } else {
+      char *p;
+      f = strtod(s.c_str(), &p);
+      if (p < s.c_str() + s.size()) strm.clear(std::ios::badbit);
+    }
+  }
+
+  // Reads LatticeWeight when there are no parentheses around pair terms...
+  // currently the only form supported.
+  inline std::istream &ReadNoParen(std::istream &strm, char separator) {
+    int c;
+    do {
+      c = strm.get();
+    } while (isspace(c));
+
+    std::string s1;
+    while (c != separator) {
+      if (c == EOF) {
+        strm.clear(std::ios::badbit);
+        return strm;
+      }
+      s1 += c;
+      c = strm.get();
+    }
+    std::istringstream strm1(s1);
+    ReadFloatType(strm1, value1_);  // ReadFloatType is class member function
+    // read second element
+    ReadFloatType(strm, value2_);
+    return strm;
+  }
+
+  friend std::istream &operator>>
+      <FloatType>(std::istream &, LatticeWeightTpl<FloatType> &);
+  friend std::ostream &operator<<<FloatType>(
+      std::ostream &, const LatticeWeightTpl<FloatType> &);
+
+ private:
+  T value1_;
+  T value2_;
+};
+
+/* ScaleTupleWeight is a function defined for LatticeWeightTpl and
+   CompactLatticeWeightTpl that mutliplies the pair (value1_, value2_) by a 2x2
+   matrix.  Used, for example, in applying acoustic scaling.
+ */
+template <class FloatType, class ScaleFloatType>
+inline LatticeWeightTpl<FloatType> ScaleTupleWeight(
+    const LatticeWeightTpl<FloatType> &w,
+    const std::vector<std::vector<ScaleFloatType> > &scale) {
+  // Without the next special case we'd get NaNs from infinity * 0
+  if (w.Value1() == std::numeric_limits<FloatType>::infinity())
+    return LatticeWeightTpl<FloatType>::Zero();
+  return LatticeWeightTpl<FloatType>(
+      scale[0][0] * w.Value1() + scale[0][1] * w.Value2(),
+      scale[1][0] * w.Value1() + scale[1][1] * w.Value2());
+}
+
+/* For testing purposes and in case it's ever useful, we define a similar
+   function to apply to LexicographicWeight and the like, templated on
+   TropicalWeight<float> etc.; we use PairWeight which is the base class of
+   LexicographicWeight.
+*/
+template <class FloatType, class ScaleFloatType>
+inline PairWeight<TropicalWeightTpl<FloatType>, TropicalWeightTpl<FloatType> >
+ScaleTupleWeight(const PairWeight<TropicalWeightTpl<FloatType>,
+                                  TropicalWeightTpl<FloatType> > &w,
+                 const std::vector<std::vector<ScaleFloatType> > &scale) {
+  typedef TropicalWeightTpl<FloatType> BaseType;
+  typedef PairWeight<BaseType, BaseType> PairType;
+  const BaseType zero = BaseType::Zero();
+  // Without the next special case we'd get NaNs from infinity * 0
+  if (w.Value1() == zero || w.Value2() == zero) return PairType(zero, zero);
+  FloatType f1 = w.Value1().Value(), f2 = w.Value2().Value();
+  return PairType(BaseType(scale[0][0] * f1 + scale[0][1] * f2),
+                  BaseType(scale[1][0] * f1 + scale[1][1] * f2));
+}
+
+template <class FloatType>
+inline bool operator==(const LatticeWeightTpl<FloatType> &wa,
+                       const LatticeWeightTpl<FloatType> &wb) {
+  // Volatile qualifier thwarts over-aggressive compiler optimizations
+  // that lead to problems esp. with NaturalLess().
+  volatile FloatType va1 = wa.Value1(), va2 = wa.Value2(), vb1 = wb.Value1(),
+                     vb2 = wb.Value2();
+  return (va1 == vb1 && va2 == vb2);
+}
+
+template <class FloatType>
+inline bool operator!=(const LatticeWeightTpl<FloatType> &wa,
+                       const LatticeWeightTpl<FloatType> &wb) {
+  // Volatile qualifier thwarts over-aggressive compiler optimizations
+  // that lead to problems esp. with NaturalLess().
+  volatile FloatType va1 = wa.Value1(), va2 = wa.Value2(), vb1 = wb.Value1(),
+                     vb2 = wb.Value2();
+  return (va1 != vb1 || va2 != vb2);
+}
+
+// We define a Compare function LatticeWeightTpl even though it's
+// not required by the semiring standard-- it's just more efficient
+// to do it this way rather than using the NaturalLess template.
+
+/// Compare returns -1 if w1 < w2, +1 if w1 > w2, and 0 if w1 == w2.
+
+template <class FloatType>
+inline int Compare(const LatticeWeightTpl<FloatType> &w1,
+                   const LatticeWeightTpl<FloatType> &w2) {
+  FloatType f1 = w1.Value1() + w1.Value2(), f2 = w2.Value1() + w2.Value2();
+  if (f1 < f2) {  // having smaller cost means you're larger
+    return 1;
+  } else if (f1 > f2) {  // in the semiring [higher probability]
+    return -1;
+  } else if (w1.Value1() < w2.Value1()) {
+  // mathematically we should be comparing (w1.value1_-w1.value2_ <
+  // w2.value1_-w2.value2_) in the next line, but add w1.value1_+w1.value2_ =
+  // w2.value1_+w2.value2_ to both sides and divide by two, and we get the
+  // simpler equivalent form w1.value1_ < w2.value1_.
+    return 1;
+  } else if (w1.Value1() > w2.Value1()) {
+    return -1;
+  } else {
+    return 0;
+  }
+}
+
+template <class FloatType>
+inline LatticeWeightTpl<FloatType> Plus(const LatticeWeightTpl<FloatType> &w1,
+                                        const LatticeWeightTpl<FloatType> &w2) {
+  return (Compare(w1, w2) >= 0 ? w1 : w2);
+}
+
+// For efficiency, override the NaturalLess template class.
+template <class FloatType>
+class NaturalLess<LatticeWeightTpl<FloatType> > {
+ public:
+  typedef LatticeWeightTpl<FloatType> Weight;
+
+  NaturalLess() {}
+
+  bool operator()(const Weight &w1, const Weight &w2) const {
+    // NaturalLess is a negative order (opposite to normal ordering).
+    // This operator () corresponds to "<" in the negative order, which
+    // corresponds to the ">" in the normal order.
+    return (Compare(w1, w2) == 1);
+  }
+};
+template <>
+class NaturalLess<LatticeWeightTpl<float> > {
+ public:
+  typedef LatticeWeightTpl<float> Weight;
+
+  NaturalLess() {}
+
+  bool operator()(const Weight &w1, const Weight &w2) const {
+    // NaturalLess is a negative order (opposite to normal ordering).
+    // This operator () corresponds to "<" in the negative order, which
+    // corresponds to the ">" in the normal order.
+    return (Compare(w1, w2) == 1);
+  }
+};
+template <>
+class NaturalLess<LatticeWeightTpl<double> > {
+ public:
+  typedef LatticeWeightTpl<double> Weight;
+
+  NaturalLess() {}
+
+  bool operator()(const Weight &w1, const Weight &w2) const {
+    // NaturalLess is a negative order (opposite to normal ordering).
+    // This operator () corresponds to "<" in the negative order, which
+    // corresponds to the ">" in the normal order.
+    return (Compare(w1, w2) == 1);
+  }
+};
+
+template <class FloatType>
+inline LatticeWeightTpl<FloatType> Times(
+    const LatticeWeightTpl<FloatType> &w1,
+    const LatticeWeightTpl<FloatType> &w2) {
+  return LatticeWeightTpl<FloatType>(w1.Value1() + w2.Value1(),
+                                     w1.Value2() + w2.Value2());
+}
+
+// divide w1 by w2 (on left/right/any doesn't matter as
+// commutative).
+template <class FloatType>
+inline LatticeWeightTpl<FloatType> Divide(const LatticeWeightTpl<FloatType> &w1,
+                                          const LatticeWeightTpl<FloatType> &w2,
+                                          DivideType typ = DIVIDE_ANY) {
+  typedef FloatType T;
+  T a = w1.Value1() - w2.Value1(), b = w1.Value2() - w2.Value2();
+  if (a != a || b != b || a == -std::numeric_limits<T>::infinity() ||
+      b == -std::numeric_limits<T>::infinity()) {
+    KALDI_WARN << "LatticeWeightTpl::Divide, NaN or invalid number produced. "
+               << "[dividing by zero?]  Returning zero";
+    return LatticeWeightTpl<T>::Zero();
+  }
+  if (a == std::numeric_limits<T>::infinity() ||
+      b == std::numeric_limits<T>::infinity())
+    return LatticeWeightTpl<T>::Zero();  // not a valid number if only one is
+                                         // infinite.
+  return LatticeWeightTpl<T>(a, b);
+}
+
+template <class FloatType>
+inline bool ApproxEqual(const LatticeWeightTpl<FloatType> &w1,
+                        const LatticeWeightTpl<FloatType> &w2,
+                        float delta = kDelta) {
+  if (w1.Value1() == w2.Value1() && w1.Value2() == w2.Value2())
+    return true;  // handles Zero().
+  return (fabs((w1.Value1() + w1.Value2()) - (w2.Value1() + w2.Value2())) <=
+          delta);
+}
+
+template <class FloatType>
+inline std::ostream &operator<<(std::ostream &strm,
+                                const LatticeWeightTpl<FloatType> &w) {
+  LatticeWeightTpl<FloatType>::WriteFloatType(strm, w.Value1());
+  CHECK(FLAGS_fst_weight_separator.size() == 1);  // NOLINT
+  strm << FLAGS_fst_weight_separator[0];  // comma by default;
+  // may or may not be settable from Kaldi programs.
+  LatticeWeightTpl<FloatType>::WriteFloatType(strm, w.Value2());
+  return strm;
+}
+
+template <class FloatType>
+inline std::istream &operator>>(std::istream &strm,
+                                LatticeWeightTpl<FloatType> &w1) {
+  CHECK(FLAGS_fst_weight_separator.size() == 1);  // NOLINT
+  // separator defaults to ','
+  return w1.ReadNoParen(strm, FLAGS_fst_weight_separator[0]);
+}
+
+// CompactLattice will be an acceptor (accepting the words/output-symbols),
+// with the weights and input-symbol-seqs on the arcs.
+// There must be a total order on W.  We assume for the sake of efficiency
+// that there is a function
+// Compare(W w1, W w2) that returns -1 if w1 < w2, +1 if w1 > w2, and
+// zero if w1 == w2, and Plus for type W returns (Compare(w1,w2) >= 0 ? w1 :
+// w2).
+
+template <class WeightType, class IntType>
+class CompactLatticeWeightTpl {
+ public:
+  typedef WeightType W;
+
+  typedef CompactLatticeWeightTpl<WeightType, IntType> ReverseWeight;
+
+  // Plus is like LexicographicWeight on the pair (weight_, string_), but where
+  // we use standard lexicographic order on string_ [this is not the same as
+  // NaturalLess on the StringWeight equivalent, which does not define a
+  // total order].
+  // Times, Divide obvious... (support both left & right division..)
+  // CommonDivisor would need to be coded separately.
+
+  CompactLatticeWeightTpl() {}
+
+  CompactLatticeWeightTpl(const WeightType &w, const std::vector<IntType> &s)
+      : weight_(w), string_(s) {}
+
+  CompactLatticeWeightTpl &operator=(
+      const CompactLatticeWeightTpl<WeightType, IntType> &w) {
+    weight_ = w.weight_;
+    string_ = w.string_;
+    return *this;
+  }
+
+  const W &Weight() const { return weight_; }
+
+  const std::vector<IntType> &String() const { return string_; }
+
+  void SetWeight(const W &w) { weight_ = w; }
+
+  void SetString(const std::vector<IntType> &s) { string_ = s; }
+
+  static const CompactLatticeWeightTpl<WeightType, IntType> Zero() {
+    return CompactLatticeWeightTpl<WeightType, IntType>(WeightType::Zero(),
+                                                        std::vector<IntType>());
+  }
+
+  static const CompactLatticeWeightTpl<WeightType, IntType> One() {
+    return CompactLatticeWeightTpl<WeightType, IntType>(WeightType::One(),
+                                                        std::vector<IntType>());
+  }
+
+  inline static std::string GetIntSizeString() {
+    char buf[2];
+    buf[0] = '0' + sizeof(IntType);
+    buf[1] = '\0';
+    return buf;
+  }
+  static const std::string &Type() {
+    static const std::string type =
+        "compact" + WeightType::Type() + GetIntSizeString();
+    return type;
+  }
+
+  static const CompactLatticeWeightTpl<WeightType, IntType> NoWeight() {
+    return CompactLatticeWeightTpl<WeightType, IntType>(WeightType::NoWeight(),
+                                                        std::vector<IntType>());
+  }
+
+  CompactLatticeWeightTpl<WeightType, IntType> Reverse() const {
+    size_t s = string_.size();
+    std::vector<IntType> v(s);
+    for (size_t i = 0; i < s; i++) v[i] = string_[s - i - 1];
+    return CompactLatticeWeightTpl<WeightType, IntType>(weight_, v);
+  }
+
+  bool Member() const {
+    // a semiring has only one zero, this is the important property
+    // we're trying to maintain here.  So force string_ to be empty if
+    // w_ == zero.
+    if (!weight_.Member()) return false;
+    if (weight_ == WeightType::Zero())
+      return string_.empty();
+    else
+      return true;
+  }
+
+  CompactLatticeWeightTpl Quantize(float delta = kDelta) const {
+    return CompactLatticeWeightTpl(weight_.Quantize(delta), string_);
+  }
+
+  static constexpr uint64 Properties() {
+    return kLeftSemiring | kRightSemiring | kPath | kIdempotent;
+  }
+
+  // This is used in OpenFst for binary I/O.  This is OpenFst-style,
+  // not Kaldi-style, I/O.
+  std::istream &Read(std::istream &strm) {
+    weight_.Read(strm);
+    if (strm.fail()) {
+      return strm;
+    }
+    int32 sz;
+    ReadType(strm, &sz);
+    if (strm.fail()) {
+      return strm;
+    }
+    if (sz < 0) {
+      KALDI_WARN << "Negative string size!  Read failure";
+      strm.clear(std::ios::badbit);
+      return strm;
+    }
+    string_.resize(sz);
+    for (int32 i = 0; i < sz; i++) {
+      ReadType(strm, &(string_[i]));
+    }
+    return strm;
+  }
+
+  // This is used in OpenFst for binary I/O.  This is OpenFst-style,
+  // not Kaldi-style, I/O.
+  std::ostream &Write(std::ostream &strm) const {
+    weight_.Write(strm);
+    if (strm.fail()) {
+      return strm;
+    }
+    int32 sz = static_cast<int32>(string_.size());
+    WriteType(strm, sz);
+    for (int32 i = 0; i < sz; i++) WriteType(strm, string_[i]);
+    return strm;
+  }
+  size_t Hash() const {
+    size_t ans = weight_.Hash();
+    // any weird numbers here are largish primes
+    size_t sz = string_.size(), mult = 6967;
+    for (size_t i = 0; i < sz; i++) {
+      ans += string_[i] * mult;
+      mult *= 7499;
+    }
+    return ans;
+  }
+
+ private:
+  W weight_;
+  std::vector<IntType> string_;
+};
+
+template <class WeightType, class IntType>
+inline bool operator==(const CompactLatticeWeightTpl<WeightType, IntType> &w1,
+                       const CompactLatticeWeightTpl<WeightType, IntType> &w2) {
+  return (w1.Weight() == w2.Weight() && w1.String() == w2.String());
+}
+
+template <class WeightType, class IntType>
+inline bool operator!=(const CompactLatticeWeightTpl<WeightType, IntType> &w1,
+                       const CompactLatticeWeightTpl<WeightType, IntType> &w2) {
+  return (w1.Weight() != w2.Weight() || w1.String() != w2.String());
+}
+
+template <class WeightType, class IntType>
+inline bool ApproxEqual(const CompactLatticeWeightTpl<WeightType, IntType> &w1,
+                        const CompactLatticeWeightTpl<WeightType, IntType> &w2,
+                        float delta = kDelta) {
+  return (ApproxEqual(w1.Weight(), w2.Weight(), delta) &&
+          w1.String() == w2.String());
+}
+
+// Compare is not part of the standard for weight types, but used internally for
+// efficiency.  The comparison here first compares the weight; if this is the
+// same, it compares the string.  The comparison on strings is: first compare
+// the length, if this is the same, use lexicographical order.  We can't just
+// use the lexicographical order because this would destroy the distributive
+// property of multiplication over addition, taking into account that addition
+// uses Compare.  The string element of "Compare" isn't super-important in
+// practical terms; it's only needed to ensure that Plus always give consistent
+// answers and is symmetric.  It's essentially for tie-breaking, but we need to
+// make sure all the semiring axioms are satisfied otherwise OpenFst might
+// break.
+
+template <class WeightType, class IntType>
+inline int Compare(const CompactLatticeWeightTpl<WeightType, IntType> &w1,
+                   const CompactLatticeWeightTpl<WeightType, IntType> &w2) {
+  int c1 = Compare(w1.Weight(), w2.Weight());
+  if (c1 != 0) return c1;
+  int l1 = w1.String().size(), l2 = w2.String().size();
+  // Use opposite order on the string lengths, so that if the costs are the
+  // same, the shorter string wins.
+  if (l1 > l2)
+    return -1;
+  else if (l1 < l2)
+    return 1;
+  for (int i = 0; i < l1; i++) {
+    if (w1.String()[i] < w2.String()[i])
+      return -1;
+    else if (w1.String()[i] > w2.String()[i])
+      return 1;
+  }
+  return 0;
+}
+
+// For efficiency, override the NaturalLess template class.
+template <class FloatType, class IntType>
+class NaturalLess<
+    CompactLatticeWeightTpl<LatticeWeightTpl<FloatType>, IntType> > {
+ public:
+  typedef CompactLatticeWeightTpl<LatticeWeightTpl<FloatType>, IntType> Weight;
+
+  NaturalLess() {}
+
+  bool operator()(const Weight &w1, const Weight &w2) const {
+    // NaturalLess is a negative order (opposite to normal ordering).
+    // This operator () corresponds to "<" in the negative order, which
+    // corresponds to the ">" in the normal order.
+    return (Compare(w1, w2) == 1);
+  }
+};
+template <>
+class NaturalLess<CompactLatticeWeightTpl<LatticeWeightTpl<float>, int32> > {
+ public:
+  typedef CompactLatticeWeightTpl<LatticeWeightTpl<float>, int32> Weight;
+
+  NaturalLess() {}
+
+  bool operator()(const Weight &w1, const Weight &w2) const {
+    // NaturalLess is a negative order (opposite to normal ordering).
+    // This operator () corresponds to "<" in the negative order, which
+    // corresponds to the ">" in the normal order.
+    return (Compare(w1, w2) == 1);
+  }
+};
+template <>
+class NaturalLess<CompactLatticeWeightTpl<LatticeWeightTpl<double>, int32> > {
+ public:
+  typedef CompactLatticeWeightTpl<LatticeWeightTpl<double>, int32> Weight;
+
+  NaturalLess() {}
+
+  bool operator()(const Weight &w1, const Weight &w2) const {
+    // NaturalLess is a negative order (opposite to normal ordering).
+    // This operator () corresponds to "<" in the negative order, which
+    // corresponds to the ">" in the normal order.
+    return (Compare(w1, w2) == 1);
+  }
+};
+
+// Make sure Compare is defined for TropicalWeight, so everything works
+// if we substitute LatticeWeight for TropicalWeight.
+inline int Compare(const TropicalWeight &w1, const TropicalWeight &w2) {
+  float f1 = w1.Value(), f2 = w2.Value();
+  if (f1 == f2)
+    return 0;
+  else if (f1 > f2)
+    return -1;
+  else
+    return 1;
+}
+
+template <class WeightType, class IntType>
+inline CompactLatticeWeightTpl<WeightType, IntType> Plus(
+    const CompactLatticeWeightTpl<WeightType, IntType> &w1,
+    const CompactLatticeWeightTpl<WeightType, IntType> &w2) {
+  return (Compare(w1, w2) >= 0 ? w1 : w2);
+}
+
+template <class WeightType, class IntType>
+inline CompactLatticeWeightTpl<WeightType, IntType> Times(
+    const CompactLatticeWeightTpl<WeightType, IntType> &w1,
+    const CompactLatticeWeightTpl<WeightType, IntType> &w2) {
+  WeightType w = Times(w1.Weight(), w2.Weight());
+  if (w == WeightType::Zero()) {
+    return CompactLatticeWeightTpl<WeightType, IntType>::Zero();
+    // special case to ensure zero is unique
+  } else {
+    std::vector<IntType> v;
+    v.resize(w1.String().size() + w2.String().size());
+    typename std::vector<IntType>::iterator iter = v.begin();
+    iter = std::copy(w1.String().begin(), w1.String().end(),
+                     iter);  // returns end of first range.
+    std::copy(w2.String().begin(), w2.String().end(), iter);
+    return CompactLatticeWeightTpl<WeightType, IntType>(w, v);
+  }
+}
+
+template <class WeightType, class IntType>
+inline CompactLatticeWeightTpl<WeightType, IntType> Divide(
+    const CompactLatticeWeightTpl<WeightType, IntType> &w1,
+    const CompactLatticeWeightTpl<WeightType, IntType> &w2,
+    DivideType div = DIVIDE_ANY) {
+  if (w1.Weight() == WeightType::Zero()) {
+    if (w2.Weight() != WeightType::Zero()) {
+      return CompactLatticeWeightTpl<WeightType, IntType>::Zero();
+    } else {
+      KALDI_ERR << "Division by zero [0/0]";
+    }
+  } else if (w2.Weight() == WeightType::Zero()) {
+    KALDI_ERR << "Error: division by zero";
+  }
+  WeightType w = Divide(w1.Weight(), w2.Weight());
+
+  const std::vector<IntType> v1 = w1.String(), v2 = w2.String();
+  if (v2.size() > v1.size()) {
+    KALDI_ERR << "Cannot divide, length mismatch";
+  }
+  typename std::vector<IntType>::const_iterator v1b = v1.begin(),
+                                                v1e = v1.end(),
+                                                v2b = v2.begin(),
+                                                v2e = v2.end();
+  if (div == DIVIDE_LEFT) {
+    if (!std::equal(v2b, v2e,
+                    v1b)) {  // v2 must be identical to first part of v1.
+      KALDI_ERR << "Cannot divide, data mismatch";
+    }
+    return CompactLatticeWeightTpl<WeightType, IntType>(
+        w, std::vector<IntType>(v1b + (v2e - v2b),
+                                v1e));  // return last part of v1.
+  } else if (div == DIVIDE_RIGHT) {
+    if (!std::equal(
+            v2b, v2e,
+            v1e - (v2e - v2b))) {  // v2 must be identical to last part of v1.
+      KALDI_ERR << "Cannot divide, data mismatch";
+    }
+    return CompactLatticeWeightTpl<WeightType, IntType>(
+        w, std::vector<IntType>(
+               v1b, v1e - (v2e - v2b)));  // return first part of v1.
+
+  } else {
+    KALDI_ERR << "Cannot divide CompactLatticeWeightTpl with DIVIDE_ANY";
+  }
+  return CompactLatticeWeightTpl<WeightType,
+                                 IntType>::Zero();  // keep compiler happy.
+}
+
+template <class WeightType, class IntType>
+inline std::ostream &operator<<(
+    std::ostream &strm, const CompactLatticeWeightTpl<WeightType, IntType> &w) {
+  strm << w.Weight();
+  CHECK(FLAGS_fst_weight_separator.size() == 1);  // NOLINT
+  strm << FLAGS_fst_weight_separator[0];  // comma by default.
+  for (size_t i = 0; i < w.String().size(); i++) {
+    strm << w.String()[i];
+    if (i + 1 < w.String().size())
+      strm << kStringSeparator;  // '_'; defined in string-weight.h in OpenFst
+                                 // code.
+  }
+  return strm;
+}
+
+template <class WeightType, class IntType>
+inline std::istream &operator>>(
+    std::istream &strm, CompactLatticeWeightTpl<WeightType, IntType> &w) {
+  std::string s;
+  strm >> s;
+  if (strm.fail()) {
+    return strm;
+  }
+  CHECK(FLAGS_fst_weight_separator.size() == 1);  // NOLINT
+  size_t pos = s.find_last_of(FLAGS_fst_weight_separator);  // normally ","
+  if (pos == std::string::npos) {
+    strm.clear(std::ios::badbit);
+    return strm;
+  }
+  // get parts of str before and after the separator (default: ',');
+  std::string s1(s, 0, pos), s2(s, pos + 1);
+  std::istringstream strm1(s1);
+  WeightType weight;
+  strm1 >> weight;
+  w.SetWeight(weight);
+  if (strm1.fail() || !strm1.eof()) {
+    strm.clear(std::ios::badbit);
+    return strm;
+  }
+  // read string part.
+  std::vector<IntType> string;
+  const char *c = s2.c_str();
+  while (*c != '\0') {
+    if (*c == kStringSeparator)  // '_'
+      c++;
+    char *c2;
+    int64_t i = strtol(c, &c2, 10);
+    if (c2 == c || static_cast<int64_t>(static_cast<IntType>(i)) != i) {
+      strm.clear(std::ios::badbit);
+      return strm;
+    }
+    c = c2;
+    string.push_back(static_cast<IntType>(i));
+  }
+  w.SetString(string);
+  return strm;
+}
+
+template <class BaseWeightType, class IntType>
+class CompactLatticeWeightCommonDivisorTpl {
+ public:
+  typedef CompactLatticeWeightTpl<BaseWeightType, IntType> Weight;
+
+  Weight operator()(const Weight &w1, const Weight &w2) const {
+    // First find longest common prefix of the strings.
+    typename std::vector<IntType>::const_iterator s1b = w1.String().begin(),
+                                                  s1e = w1.String().end(),
+                                                  s2b = w2.String().begin(),
+                                                  s2e = w2.String().end();
+    while (s1b < s1e && s2b < s2e && *s1b == *s2b) {
+      s1b++;
+      s2b++;
+    }
+    return Weight(Plus(w1.Weight(), w2.Weight()),
+                  std::vector<IntType>(w1.String().begin(), s1b));
+  }
+};
+
+/** Scales the pair (a, b) of floating-point weights inside a
+    CompactLatticeWeight by premultiplying it (viewed as a vector)
+    by a 2x2 matrix "scale".
+    Assumes there is a ScaleTupleWeight function that applies to "Weight";
+    this currently only works if Weight equals LatticeWeightTpl<FloatType>
+    for some FloatType.
+*/
+template <class Weight, class IntType, class ScaleFloatType>
+inline CompactLatticeWeightTpl<Weight, IntType> ScaleTupleWeight(
+    const CompactLatticeWeightTpl<Weight, IntType> &w,
+    const std::vector<std::vector<ScaleFloatType> > &scale) {
+  return CompactLatticeWeightTpl<Weight, IntType>(
+      Weight(ScaleTupleWeight(w.Weight(), scale)), w.String());
+}
+
+/** Define some ConvertLatticeWeight functions that are used in various lattice
+    conversions... make them all templates, some with no arguments, since some
+    must be templates.*/
+template <class Float1, class Float2>
+inline void ConvertLatticeWeight(const LatticeWeightTpl<Float1> &w_in,
+                                 LatticeWeightTpl<Float2> *w_out) {
+  w_out->SetValue1(w_in.Value1());
+  w_out->SetValue2(w_in.Value2());
+}
+
+template <class Float1, class Float2, class Int>
+inline void ConvertLatticeWeight(
+    const CompactLatticeWeightTpl<LatticeWeightTpl<Float1>, Int> &w_in,
+    CompactLatticeWeightTpl<LatticeWeightTpl<Float2>, Int> *w_out) {
+  LatticeWeightTpl<Float2> weight2(w_in.Weight().Value1(),
+                                   w_in.Weight().Value2());
+  w_out->SetWeight(weight2);
+  w_out->SetString(w_in.String());
+}
+
+// to convert from Lattice to standard FST
+template <class Float1, class Float2>
+inline void ConvertLatticeWeight(const LatticeWeightTpl<Float1> &w_in,
+                                 TropicalWeightTpl<Float2> *w_out) {
+  TropicalWeightTpl<Float2> w1(w_in.Value1());
+  TropicalWeightTpl<Float2> w2(w_in.Value2());
+  *w_out = Times(w1, w2);
+}
+
+template <class Float>
+inline double ConvertToCost(const LatticeWeightTpl<Float> &w) {
+  return static_cast<double>(w.Value1()) + static_cast<double>(w.Value2());
+}
+
+template <class Float, class Int>
+inline double ConvertToCost(
+    const CompactLatticeWeightTpl<LatticeWeightTpl<Float>, Int> &w) {
+  return static_cast<double>(w.Weight().Value1()) +
+         static_cast<double>(w.Weight().Value2());
+}
+
+template <class Float>
+inline double ConvertToCost(const TropicalWeightTpl<Float> &w) {
+  return w.Value();
+}
+
+}  // namespace fst
+
+#endif  // KALDI_FSTEXT_LATTICE_WEIGHT_H_
diff --git a/speechx/speechx/kaldi/fstext/pre-determinize-inl.h b/speechx/speechx/kaldi/fstext/pre-determinize-inl.h
new file mode 100644
index 00000000..632d0415
--- /dev/null
+++ b/speechx/speechx/kaldi/fstext/pre-determinize-inl.h
@@ -0,0 +1,798 @@
+// fstext/pre-determinize-inl.h
+
+// Copyright 2009-2011  Microsoft Corporation
+
+// See ../../COPYING for clarification regarding multiple authors
+//
+// Licensed under the Apache License, Version 2.0 (the "License");
+// you may not use this file except in compliance with the License.
+// You may obtain a copy of the License at
+//
+//  http://www.apache.org/licenses/LICENSE-2.0
+//
+// THIS CODE IS PROVIDED *AS IS* BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY
+// KIND, EITHER EXPRESS OR IMPLIED, INCLUDING WITHOUT LIMITATION ANY IMPLIED
+// WARRANTIES OR CONDITIONS OF TITLE, FITNESS FOR A PARTICULAR PURPOSE,
+// MERCHANTABLITY OR NON-INFRINGEMENT.
+// See the Apache 2 License for the specific language governing permissions and
+// limitations under the License.
+
+#ifndef KALDI_FSTEXT_PRE_DETERMINIZE_INL_H_
+#define KALDI_FSTEXT_PRE_DETERMINIZE_INL_H_
+
+#include <algorithm>
+#include <map>
+#include <set>
+#include <string>
+#include <utility>
+#include <vector>
+
+/* Do not include this file directly.  It is an implementation file included by
+ * PreDeterminize.h */
+
+/*
+  Predeterminization
+
+    This is a function that makes an FST compactly determinizable by inserting
+  symbols on the input side as necessary for disambiguation.  Note that we do
+  not treat epsilon as a real symbol when measuring determinizability in this
+  sense.   The extra symbols are added to the vocabulary, on the input side;
+  these are of the form (prefix)1, (prefix)2, and so on without limit, where
+  (prefix) is some prefix the user provides, e.g. '#' (the function checks that
+  this will not lead to conflicts with symbols already in the FST).  The
+  function tells us how many such symbols it created.
+
+   Note that there is a paper "Generalized optimization algorithm for speech
+  recognition transducers" by Allauzen and Mohri, that deals with a similar
+  issue, but this is a very different algorithm that only aims to ensure
+  determinizability, but not *compact* determinizability.
+
+   Our algorithm is slightly heuristic, and probably not optimal, but does
+  ensure that the output is compactly determinizable, possibly at the expense of
+  inserting unnecessary symbols.  We considered more sophisticated algorithms,
+  but these were extremely complicated and would give the same output for the
+  kinds of inputs that we envisage.
+
+   Suppose the input FST is T.  We want to ensure that in det(T), if we consider
+  the states of det(T) as weighted subsets of states of T, each state of T only
+  appears once in any given subset.  This ensures that det(T) is no larger than
+  T in an appropriate sense.  The way we do this is as follows.  We identify all
+  states in T that have multiple input transitions (counting "being an initial
+  state" as an input transition). Let's call these "problematic" states.  For a
+  problematic state p we stipulate that it can never appear in any state of
+  det(T) unless that state equals (p, \bar{1}) [i.e. p, unweighted].  In order
+  to ensure this, we insert input symbols on the transitions to these
+   problematic states (this may necessitate adding extra states).
+      We also stipulate that the path through det(T) should always be sufficient
+  to tell us the path through T (and we insert extra symbols sufficient to make
+  this so).  This is to simplify the algorithm, so that we don't have to
+  consider the output symbols or weights when predeterminizing.
+
+   The algorithm is as follows.
+
+    (A) Definitions
+
+      (i)  Define a *problematic state* as a state that either has multiple
+  input transitions, or is an initial state and has at least one input
+  transition.
+
+     (ii)  For an arc a, define:
+            i[a] = input symbol on a
+            o[a] = output symbol on a
+            n[a] = dest-state of a
+            p[a] = origin-state of a
+
+           For a state q, define
+            E[q] = set of transitions leaving q.
+           For a set of states Q, define
+            E[Q] = set of transitions leaving some q in Q
+
+    (iii)  For a state s, define Closure(s) as the union of state s, and all
+  states t that are reachable via sequences of arcs a such that i[a]=epsilon and
+  n[a] is not problematic.
+
+           For a set of states S, define Closure(S) as the union of the closures
+  of states s in S.
+
+    (B) Inputs and outputs.
+
+     (i) Inputs and preconditions.  Input is an FST, which should have a symbol
+  table compiled into it, and a prefix (e.g. #) for symbols to be added.  We
+  check that the input FST is trim, and that it does not have any symbols that
+  appear on its arcs, that are equal to the prefix followed by digits.
+
+    (ii) Outputs: The algorithm modifies the FST that is given to it, and
+  returns the number of the highest numbered "extra symbol" inserted.  The extra
+  symbols are numbered #1, #2 and so on without limit (as integers).  They are
+  inserted into the symbol table in a sequential way by calling AvailableKey()
+         for each in turn (this is stipulated in case we need to keep other
+  symbol tables in sync).
+
+     (C) Sub-algorithm: Closure(S).  This requires the array p(s), defined
+  below, which is true if s is problematic.  This also requires, for efficiency,
+  that the arcs be sorted on input label. Input: a set of states S.  [plus, the
+  fst and the array p]. Output: a set of states T. Algorithm: set T <-- S, Q <--
+  S. while Q is nonempty: pop a state s from Q. for each transition a from state
+  s with epsilon on the input label [we can find these efficiently using the
+  sorting on arcs]: If p(n[a]) is false and n[a] is not in T: Insert n[a] into
+  T. Add n[a] to Q. return T.
+
+
+     (D) Main algorithm.
+
+
+       (i) (a) Check preconditions (FST is trim)
+           (b) Make sure there is just one final state (insert epsilon
+  transitions as necessary). (c) Sort arcs on input label (so epsilon arcs are
+  at the start of arc lists).
+
+
+      (ii) Work out the set of problematic states by constructing a boolean
+  array indexed by states, i.e. p(s) which is true if the state is problematic.
+  We can do this by constructing an array t(s) to store the number of
+  transitions into each state [adding one for the initial state], and then
+  setting p(s) = true if t(s) > 1.
+
+           Also create a boolean array d(s), defined for states, and set d(s) =
+  false. This array is purely for sanity-checking that we are processing each
+  state exactly once.
+
+     (iii) Set up an array of integers m(a), indexed by arcs (how exactly we
+  store these is implementation-dependent, but this will probably be a hash from
+  (state, arc-index) to integers.  m(a) will store the extra symbol, if any, to
+  be added to that arc (or -1 if no such symbol; we can also simply have the arc
+  not present in the hash).  The initial value of m(a) is -1 (if array), or
+  undefined (if hash).
+
+      (iv) Initialize a set of sets-of-states S, and a queue of pairs Q, as
+  follows. The pairs in Q are a pair of (set-of-states, integer), where the
+  integer is the number of "special symbols" already used up for that state.
+
+            Note that we use a special indexing for the sets in both S and Q,
+  rather than using std::set.  We use a sorted vector of StateId's.  And in S,
+  we index them by the lowest-numbered state-id.  Because each state is supposed
+  to only ever be a member of one set, if there is an attempt to add another,
+  different set with the same lowest-numbered state-id, we detect an error.
+
+            Let I be the single initial state (OpenFST only supports one).
+            We set:
+              S = { Closure(I) }
+              Push (Closure(I), 0)  onto Q.
+            Then for each state s such that p(s) = true, and s is not an initial
+  state: S <-- S u { Closure(s) } Push (Closure(s), 0)  onto Q.
+
+       (v) While Q is nonempty:
+
+          (a) Pop pair (A, n) from Q (queue discipline is arbitrary).
+
+          (b) For each state s in A, check that d(s) is false, and set d(s) to
+  true. This is for sanity checking only.
+
+          (c)
+             Let S_\eps be the set of epsilon-transitions from members of A to
+  problematic states (i.e. S_\eps = \{ a \in E[A]: i[a]=\epsilon, p(n[a]) = true
+  \}).
+
+             Next, we will define, for each t \neq \epsilon, S_t as the set of
+               transitions from some state s in S with t as the input label,
+  i.e.: S_t = \{ a \in E[A]: i[a] = t \} We further define T_t and U_t as the
+  subsets of S where the destination state is problematic and non-problematic
+  respectively, i.e: T_t = \{ a \in E[A]: i[a] = t, p(n[a]) = true \} U_t = \{ a
+  \in E[A]: i[a] = t, p(n[a]) = false \}
+
+             The easiest way to obtain these sets is probably to have a hash
+  indexed by t that maps to a list of pairs (state, arc-offset) that stores S_t.
+               From this we can work out the sizes of T_t and U_t on the fly.
+
+         (d)
+             for each transition a in S_\eps:
+                m(a) <-- n # Will put symbol n on this transition.
+                n <-- n+1  # Note, same n as in pair (A, n)
+
+         (e)
+             next,
+             for each t\neq epsilon s.t. S_t is nonempty,
+
+                if |S_t| > 1 #if-statement is because if |S_t|=|T_t|=1, no need
+  for prefix. k = 0 for each transition a in T_t: set m(a) to k. set k = k+1
+
+                if |U_t| > 0
+                   Let V_t be the set of destination-states of arcs in U_t.
+                   if Closure(V_t) is not in S:
+                     insert Closure(V_t) into S, and add the pair (Closure(V_t),
+  k) to Q.
+
+       (vi) Check that for each state in the FST, d(s) = true.
+
+      (vii) Let n = max_a m(a).  This is the highest-numbered extra symbol
+  (extra symbols start from zero, in this numbering which doesn't correspond to
+  the symbol-table numbering).  Here we add n+1 extra symbols to the symbol
+  table and store the mappings from 0, 1, ... n to the symbol-id.
+
+     (viii) Set up a hash h from (state, int) to (state-id) such that
+             t = h(s, k)
+            will be the state-id of a newly-created state that has a transition
+  to state s with input-label #k.
+
+      (ix) For each arc a such that m(a) != 0:
+             If i[a] = epsilon (the input label is epsilon):
+                Change i[a] to #m(a). [i.e. prefix then digit m(a)]
+             Otherwise:
+                If t = h(n[a], m(a)) is not defined [where n[a] is the
+  dest-state]: create a new state t with a transition to n[a], with input-label
+  #m(a) and no output-label or weight.  Set h(n[a], m(a)) = t. Change n[a] to
+  h(n[a], m(a)).
+
+
+*/
+namespace fst {
+
+namespace pre_determinize_helpers {
+
+// make it inline to avoid having to put it in a .cc file which most functions
+// here could not go in.
+inline bool HasBannedPrefixPlusDigits(SymbolTable *symTable, std::string prefix,
+                                      std::string *bad_sym) {
+  // returns true if the symbol table contains any string consisting of this
+  // (possibly empty) prefix followed by a nonempty sequence of digits (0 to 9).
+  // requires symTable to be non-NULL.
+  // if bad_sym != NULL, puts the first bad symbol it finds in *bad_sym.
+  assert(symTable != NULL);
+  const char *prefix_ptr = prefix.c_str();
+  size_t prefix_len =
+      strlen(prefix_ptr);  // allowed to be zero but not encouraged.
+  for (SymbolTableIterator siter(*symTable); !siter.Done(); siter.Next()) {
+    const std::string &sym = siter.Symbol();
+    if (!strncmp(prefix_ptr, sym.c_str(), prefix_len)) {  // has prefix.
+      if (isdigit(sym[prefix_len])) {  // we don't allow prefix followed by a
+                                       // digit, as a symbol.
+        // Has at least one digit.
+        size_t pos;
+        for (pos = prefix_len; sym[pos] != '\0'; pos++)
+          if (!isdigit(sym[pos])) break;
+        if (sym[pos] == '\0') {  // All remaining characters were digits.
+          if (bad_sym != NULL) *bad_sym = sym;
+          return true;
+        }
+      }  // else OK because prefix was followed by '\0' or a non-digit.
+    }
+  }
+  return false;  // doesn't have banned symbol.
+}
+
+template <class T>
+void CopySetToVector(const std::set<T> s, std::vector<T> *v) {
+  // adds members of s to v, in sorted order from lowest to highest
+  // (because the set was in sorted order).
+  assert(v != NULL);
+  v->resize(s.size());
+  typename std::set<T>::const_iterator siter = s.begin();
+  typename std::vector<T>::iterator viter = v->begin();
+  for (; siter != s.end(); ++siter, ++viter) {
+    assert(viter != v->end());
+    *viter = *siter;
+  }
+}
+
+// Warning.  This function calls 'new'.
+template <class T>
+std::vector<T> *InsertMember(const std::vector<T> m,
+                             std::vector<std::vector<T> *> *S) {
+  assert(m.size() > 0);
+  T idx = m[0];
+  assert(idx >= (T)0 && idx < (T)S->size());
+  if ((*S)[idx] != NULL) {
+    assert(*((*S)[idx]) == m);
+    // The vectors should be the same.  Otherwise this is a bug in the
+    // algorithm. It could either be a programming error or a deeper conceptual
+    // bug.
+    return NULL;  // nothing was inserted.
+  } else {
+    std::vector<T> *ret = (*S)[idx] = new std::vector<T>(m);  // New copy of m.
+    return ret;                                               // was inserted.
+  }
+}
+
+// See definition of Closure(S) in item A(iii) in the comment above. it's the
+// set of states that are reachable from S via sequences of arcs a such that
+// i[a]=epsilon and n[a] is not problematic.  We assume that the fst is sorted
+// on input label (so epsilon arcs first) The algorithm is described in section
+// (C) above.  We use the same variable for S and T.
+template <class Arc>
+void Closure(MutableFst<Arc> *fst, std::set<typename Arc::StateId> *S,
+             const std::vector<bool> &pVec) {
+  typedef typename Arc::StateId StateId;
+  std::vector<StateId> Q;
+  CopySetToVector(*S, &Q);
+  while (Q.size() != 0) {
+    StateId s = Q.back();
+    Q.pop_back();
+    for (ArcIterator<MutableFst<Arc> > aiter(*fst, s); !aiter.Done();
+         aiter.Next()) {
+      const Arc &arc = aiter.Value();
+      if (arc.ilabel != 0)
+        break;  // Break from the loop: due to sorting there will be no
+      // more transitions with epsilons as input labels.
+      if (!pVec[arc.nextstate]) {  // Next state is not problematic -> we can
+                                   // use this transition.
+        std::pair<typename std::set<StateId>::iterator, bool> p =
+            S->insert(arc.nextstate);
+        if (p.second) {  // True means: was inserted into S (wasn't already
+                         // there).
+          Q.push_back(arc.nextstate);
+        }
+      }
+    }
+  }
+}  // end function Closure.
+
+}  // end namespace pre_determinize_helpers.
+
+template <class Arc, class Int>
+void PreDeterminize(MutableFst<Arc> *fst, typename Arc::Label first_new_sym,
+                    std::vector<Int> *symsOut) {
+  typedef typename Arc::Label Label;
+  typedef typename Arc::StateId StateId;
+  typedef size_t ArcId;  // Our own typedef, not standard OpenFst.  Use size_t
+  // for compatibility with argument of ArcIterator::Seek().
+  typedef typename Arc::Weight Weight;
+  assert(first_new_sym > 0);
+  assert(fst != NULL);
+  if (fst->Start() == kNoStateId) return;  // for empty FST, nothing to do.
+  assert(symsOut != NULL &&
+         symsOut->size() == 0);  // we will output the symbols we add into this.
+
+  {  // (D)(i)(a): check is trim (i.e. connected, in OpenFST parlance).
+    KALDI_VLOG(2) << "PreDeterminize: Checking FST properties";
+    uint64 props = fst->Properties(
+        kAccessible | kCoAccessible,
+        true);  // true-> computes properties if unknown at time when called.
+    if (props !=
+        (kAccessible | kCoAccessible)) {  // All states are not both accessible
+                                          // and co-accessible...
+      KALDI_ERR << "PreDeterminize: FST is not trim";
+    }
+  }
+
+  {  // (D)(i)(b): make single final state.
+    KALDI_VLOG(2) << "PreDeterminize: creating single final state";
+    CreateSuperFinal(fst);
+  }
+
+  {  // (D)(i)(c): sort arcs on input.
+    KALDI_VLOG(2) << "PreDeterminize: sorting arcs on input";
+    ILabelCompare<Arc> icomp;
+    ArcSort(fst, icomp);
+  }
+
+  StateId n_states = 0,
+          max_state =
+              0;  // Compute n_states, max_state = highest-numbered state.
+  {               // compute nStates, maxStates.
+    for (StateIterator<MutableFst<Arc> > iter(*fst); !iter.Done();
+         iter.Next()) {
+      StateId state = iter.Value();
+      assert(state >= 0);
+      n_states++;
+      if (state > max_state) max_state = state;
+    }
+    KALDI_VLOG(2) << "PreDeterminize: n_states = " << (n_states)
+                  << ", max_state =" << (max_state);
+  }
+
+  std::vector<bool> p_vec(max_state + 1, false);  // compute this next.
+  {  // D(ii): computing the array p. ["problematic states, i.e. states with >1
+     // input transition,
+    // counting being the initial state as an input transition"].
+    std::vector<bool> seen_vec(
+        max_state + 1,
+        false);  // rather than counting incoming transitions we just have a
+                 // bool that says we saw at least one.
+
+    seen_vec[fst->Start()] = true;
+    for (StateIterator<MutableFst<Arc> > siter(*fst); !siter.Done();
+         siter.Next()) {
+      for (ArcIterator<MutableFst<Arc> > aiter(*fst, siter.Value());
+           !aiter.Done(); aiter.Next()) {
+        const Arc &arc = aiter.Value();
+        assert(arc.nextstate >= 0 && arc.nextstate < max_state + 1);
+        if (seen_vec[arc.nextstate])
+          p_vec[arc.nextstate] =
+              true;  // now have >1 transition in, so problematic.
+        else
+          seen_vec[arc.nextstate] = true;
+      }
+    }
+  }
+  // D(iii): set up m(a)
+  std::map<std::pair<StateId, ArcId>, size_t> m_map;
+  // This is the array m, indexed by arcs.  It maps to the index of the symbol
+  // we add.
+
+  // WARNING: we should be sure to clean up this memory before exiting.  Do not
+  // return or throw an exception from this function, later than this point,
+  // without cleaning up! Note that the vectors are shared between Q and S (they
+  // "belong to" S.
+  std::vector<std::vector<StateId> *> S(max_state + 1,
+                                        (std::vector<StateId> *)(void *)0);
+  std::vector<std::pair<std::vector<StateId> *, size_t> > Q;
+
+  // D(iv): initialize S and Q.
+  {
+    std::vector<StateId>
+        all_seed_states;  // all "problematic" states, plus initial state (if
+                          // not problematic).
+    if (!p_vec[fst->Start()]) all_seed_states.push_back(fst->Start());
+    for (StateId s = 0; s <= max_state; s++)
+      if (p_vec[s]) all_seed_states.push_back(s);
+
+    for (size_t idx = 0; idx < all_seed_states.size(); idx++) {
+      StateId s = all_seed_states[idx];
+      std::set<StateId> closure_s;
+      closure_s.insert(s);  // insert "seed" state.
+      pre_determinize_helpers::Closure(
+          fst, &closure_s,
+          p_vec);  // follow epsilons to non-problematic states.
+      // Closure in this case whis will usually not add anything, for typical
+      // topologies in speech
+      std::vector<StateId> closure_s_vec;
+      pre_determinize_helpers::CopySetToVector(closure_s, &closure_s_vec);
+      KALDI_ASSERT(closure_s_vec.size() != 0);
+      std::vector<StateId> *ptr =
+          pre_determinize_helpers::InsertMember(closure_s_vec, &S);
+      KALDI_ASSERT(ptr != NULL);  // Or conceptual bug or programming error.
+      Q.push_back(std::pair<std::vector<StateId> *, size_t>(ptr, 0));
+    }
+  }
+
+  std::vector<bool> d_vec(max_state + 1,
+                          false);  // "done vector".  Purely for debugging.
+
+  size_t num_extra_det_states = 0;
+
+  // (D)(v)
+  while (Q.size() != 0) {
+    // (D)(v)(a)
+    std::pair<std::vector<StateId> *, size_t> cur_pair(Q.back());
+    Q.pop_back();
+    const std::vector<StateId> &A(*cur_pair.first);
+    size_t n = cur_pair.second;  // next special symbol to add.
+
+    // (D)(v)(b)
+    for (size_t idx = 0; idx < A.size(); idx++) {
+      assert(d_vec[A[idx]] == false &&
+             "This state has been seen before.  Algorithm error.");
+      d_vec[A[idx]] = true;
+    }
+
+    // From here is (D)(v)(c).  We work out S_\eps and S_t (for t\neq eps)
+    // simultaneously at first.
+    std::map<Label, std::set<std::pair<std::pair<StateId, ArcId>, StateId> > >
+        arc_hash;
+    // arc_hash is a hash with info of all arcs from states in the set A to
+    // non-problematic states.
+    // It is a map from ilabel to pair(pair(start-state, arc-offset),
+    // end-state). Here, arc-offset reflects the order in which we accessed the
+    // arc using the ArcIterator (zero for the first arc).
+
+    {  // This block sets up arc_hash
+      for (size_t idx = 0; idx < A.size(); idx++) {
+        StateId s = A[idx];
+        assert(s >= 0 && s <= max_state);
+        ArcId arc_id = 0;
+        for (ArcIterator<MutableFst<Arc> > aiter(*fst, s); !aiter.Done();
+             aiter.Next(), ++arc_id) {
+          const Arc &arc = aiter.Value();
+
+          std::pair<std::pair<StateId, ArcId>, StateId> this_pair(
+              std::pair<StateId, ArcId>(s, arc_id), arc.nextstate);
+          bool inserted = (arc_hash[arc.ilabel].insert(this_pair)).second;
+          assert(inserted);  // Otherwise we had a duplicate.
+        }
+      }
+    }
+
+    // (D)(v)(d)
+    if (arc_hash.count(0) == 1) {  // We have epsilon transitions out.
+      std::set<std::pair<std::pair<StateId, ArcId>, StateId> > &eps_set =
+          arc_hash[0];
+      typedef typename std::set<
+          std::pair<std::pair<StateId, ArcId>, StateId> >::iterator set_iter_t;
+      for (set_iter_t siter = eps_set.begin(); siter != eps_set.end();
+           ++siter) {
+        const std::pair<std::pair<StateId, ArcId>, StateId> &this_pr = *siter;
+        if (p_vec[this_pr.second]) {  // Eps-transition to problematic state.
+          assert(m_map.count(this_pr.first) == 0);
+          m_map[this_pr.first] = n;
+          n++;
+        }
+      }
+    }
+
+    // (D)(v)(e)
+    {
+      typedef typename std::map<
+          Label,
+          std::set<std::pair<std::pair<StateId, ArcId>, StateId> > >::iterator
+          map_iter_t;
+      typedef typename std::set<
+          std::pair<std::pair<StateId, ArcId>, StateId> >::iterator set_iter_t2;
+      for (map_iter_t miter = arc_hash.begin(); miter != arc_hash.end();
+           ++miter) {
+        Label t = miter->first;
+        std::set<std::pair<std::pair<StateId, ArcId>, StateId> > &S_t =
+            miter->second;
+        if (t != 0) {             // For t != epsilon,
+          std::set<StateId> V_t;  // set of destination non-problem states. Will
+                                  // create this set now.
+
+          // exists_noproblem is true iff |U_t| > 0.
+          size_t k = 0;
+
+          // First loop "for each transition a in T_t" (i.e. transitions to
+          // problematic states) The if-statement if (|S_t|>1) is pushed inside
+          // the loop, as the loop also computes the set V_t.
+          for (set_iter_t2 siter = S_t.begin(); siter != S_t.end(); ++siter) {
+            const std::pair<std::pair<StateId, ArcId>, StateId> &this_pr =
+                *siter;
+            if (p_vec[this_pr.second]) {  // only consider problematic states
+                                          // (just set T_t)
+              if (S_t.size() >
+                  1) {  // This is where we pushed the if-statement in.
+                assert(m_map.count(this_pr.first) == 0);
+                m_map[this_pr.first] = k;
+                k++;
+                num_extra_det_states++;
+              }
+            } else {  // Create the set V_t.
+              V_t.insert(this_pr.second);
+            }
+          }
+          if (V_t.size() != 0) {
+            pre_determinize_helpers::Closure(
+                fst, &V_t,
+                p_vec);  // follow epsilons to non-problematic states.
+            std::vector<StateId> closure_V_t_vec;
+            pre_determinize_helpers::CopySetToVector(V_t, &closure_V_t_vec);
+            std::vector<StateId> *ptr =
+                pre_determinize_helpers::InsertMember(closure_V_t_vec, &S);
+            if (ptr != NULL) {  // was inserted.
+              Q.push_back(std::pair<std::vector<StateId> *, size_t>(ptr, k));
+            }
+          }
+        }
+      }
+    }
+  }  // end while (Q.size() != 0)
+
+  {  // (D)(vi): Check that for each state in the FST, d(s) = true.
+    for (StateIterator<MutableFst<Arc> > siter(*fst); !siter.Done();
+         siter.Next()) {
+      StateId val = siter.Value();
+      assert(d_vec[val] == true);
+    }
+  }
+
+  {  // (D)(vii): compute symbol-table ID's.
+    // sets up symsOut array.
+    int64 n = -1;
+    for (typename std::map<std::pair<StateId, ArcId>, size_t>::iterator m_iter =
+             m_map.begin();
+         m_iter != m_map.end(); ++m_iter) {
+      n = std::max(n,
+                   static_cast<int64>(
+                       m_iter->second));  // m_iter->second is of type size_t.
+    }
+    // At this point n is the highest symbol-id (type size_t) of symbols we must
+    // add.
+    n++;  // This is now the number of symbols we must add.
+    for (size_t i = 0; static_cast<int64>(i) < n; i++)
+      symsOut->push_back(first_new_sym + i);
+  }
+
+  // (D)(viii): set up hash.
+  std::map<std::pair<StateId, size_t>, StateId> h_map;
+
+  {  // D(ix): add extra symbols!  This is where the work gets done.
+    // Core part of this is below, search for (*)
+    size_t n_states_added = 0;
+
+    for (typename std::map<std::pair<StateId, ArcId>, size_t>::iterator m_iter =
+             m_map.begin();
+         m_iter != m_map.end(); ++m_iter) {
+      StateId state = m_iter->first.first;
+      ArcId arcpos = m_iter->first.second;
+      size_t m_a = m_iter->second;
+
+      MutableArcIterator<MutableFst<Arc> > aiter(fst, state);
+      aiter.Seek(arcpos);
+      Arc arc = aiter.Value();
+
+      // (*) core part here.
+      if (arc.ilabel == 0) {
+        arc.ilabel = (*symsOut)[m_a];
+      } else {
+        std::pair<StateId, size_t> pr(arc.nextstate, m_a);
+        if (!h_map.count(pr)) {
+          n_states_added++;
+          StateId newstate = fst->AddState();
+          assert(newstate >= 0);
+          Arc new_arc((*symsOut)[m_a], (Label)0, Weight::One(), arc.nextstate);
+          fst->AddArc(newstate, new_arc);
+          h_map[pr] = newstate;
+        }
+        arc.nextstate = h_map[pr];
+      }
+      aiter.SetValue(arc);
+    }
+
+    KALDI_VLOG(2) << "Added " << (n_states_added)
+                  << " new states and added/changed " << (m_map.size())
+                  << " arcs";
+  }
+  // Now free up memory.
+  for (size_t i = 0; i < S.size(); i++) delete S[i];
+}  // end function PreDeterminize
+
+template <class Label>
+void CreateNewSymbols(SymbolTable *input_sym_table, int nSym,
+                      std::string prefix, std::vector<Label> *symsOut) {
+  // Creates nSym new symbols named (prefix)0, (prefix)1 and so on.
+  // Crashes if it cannot create them because one or more of them were in the
+  // symbol table already.
+  assert(symsOut && symsOut->size() == 0);
+  for (int i = 0; i < nSym; i++) {
+    std::stringstream ss;
+    ss << prefix << i;
+    std::string str = ss.str();
+    if (input_sym_table->Find(str) != -1) {  // should not be present.
+    }
+    assert(symsOut);
+    symsOut->push_back((Label)input_sym_table->AddSymbol(str));
+  }
+}
+
+// see pre-determinize.h for documentation.
+template <class Arc>
+void AddSelfLoops(MutableFst<Arc> *fst,
+                  const std::vector<typename Arc::Label> &isyms,
+                  const std::vector<typename Arc::Label> &osyms) {
+  assert(fst != NULL);
+  assert(isyms.size() == osyms.size());
+  typedef typename Arc::Label Label;
+  typedef typename Arc::StateId StateId;
+  typedef typename Arc::Weight Weight;
+  size_t n = isyms.size();
+  if (n == 0) return;  // Nothing to do.
+
+  // {
+  // the following declarations and statements are for quick detection of these
+  // symbols, which is purely for debugging/checking purposes.
+  Label isyms_min = *std::min_element(isyms.begin(), isyms.end()),
+        isyms_max = *std::max_element(isyms.begin(), isyms.end()),
+        osyms_min = *std::min_element(osyms.begin(), osyms.end()),
+        osyms_max = *std::max_element(osyms.begin(), osyms.end());
+  std::set<Label> isyms_set, osyms_set;
+  for (size_t i = 0; i < isyms.size(); i++) {
+    assert(isyms[i] > 0 &&
+           osyms[i] > 0);  // should not have epsilon or invalid symbols.
+    isyms_set.insert(isyms[i]);
+    osyms_set.insert(osyms[i]);
+  }
+  assert(isyms_set.size() == n && osyms_set.size() == n);
+  // } end block.
+
+  for (StateIterator<MutableFst<Arc> > siter(*fst); !siter.Done();
+       siter.Next()) {
+    StateId state = siter.Value();
+    bool this_state_needs_self_loops = (fst->Final(state) != Weight::Zero());
+    for (ArcIterator<MutableFst<Arc> > aiter(*fst, state); !aiter.Done();
+         aiter.Next()) {
+      const Arc &arc = aiter.Value();
+      // If one of the following asserts fails, it means that the input FST
+      // already had the symbols we are inserting.  This is contrary to the
+      // preconditions of this algorithm.
+      assert(!(arc.ilabel >= isyms_min && arc.ilabel <= isyms_max &&
+               isyms_set.count(arc.ilabel) != 0));
+      assert(!(arc.olabel >= osyms_min && arc.olabel <= osyms_max &&
+               osyms_set.count(arc.olabel) != 0));
+      if (arc.olabel != 0)  // Has non-epsilon output label -> need self loops.
+        this_state_needs_self_loops = true;
+    }
+    if (this_state_needs_self_loops) {
+      for (size_t i = 0; i < n; i++) {
+        Arc arc;
+        arc.ilabel = isyms[i];
+        arc.olabel = osyms[i];
+        arc.weight = Weight::One();
+        arc.nextstate = state;
+        fst->AddArc(state, arc);
+      }
+    }
+  }
+}
+
+template <class Arc>
+int64 DeleteISymbols(MutableFst<Arc> *fst,
+                     std::vector<typename Arc::Label> isyms) {
+  // We could do this using the Mapper concept, but this is much easier to
+  // understand.
+
+  typedef typename Arc::Label Label;
+  typedef typename Arc::StateId StateId;
+
+  int64 num_deleted = 0;
+
+  if (isyms.size() == 0) return 0;
+  Label isyms_min = *std::min_element(isyms.begin(), isyms.end()),
+        isyms_max = *std::max_element(isyms.begin(), isyms.end());
+  bool isyms_consecutive =
+      (isyms_max + 1 - isyms_min == static_cast<Label>(isyms.size()));
+  std::set<Label> isyms_set;
+  if (!isyms_consecutive) {
+    for (size_t i = 0; i < isyms.size(); i++) isyms_set.insert(isyms[i]);
+  }
+
+  for (StateIterator<MutableFst<Arc> > siter(*fst); !siter.Done();
+       siter.Next()) {
+    StateId state = siter.Value();
+    for (MutableArcIterator<MutableFst<Arc> > aiter(fst, state); !aiter.Done();
+         aiter.Next()) {
+      const Arc &arc = aiter.Value();
+      if (arc.ilabel >= isyms_min && arc.ilabel <= isyms_max) {
+        if (isyms_consecutive || isyms_set.count(arc.ilabel) != 0) {
+          num_deleted++;
+          Arc mod_arc(arc);
+          mod_arc.ilabel = 0;  // change label to epsilon.
+          aiter.SetValue(mod_arc);
+        }
+      }
+    }
+  }
+  return num_deleted;
+}
+
+template <class Arc>
+typename Arc::StateId CreateSuperFinal(MutableFst<Arc> *fst) {
+  typedef typename Arc::StateId StateId;
+  typedef typename Arc::Weight Weight;
+  assert(fst != NULL);
+  StateId num_states = fst->NumStates();
+  StateId num_final = 0;
+  std::vector<StateId> final_states;
+  for (StateId s = 0; s < num_states; s++) {
+    if (fst->Final(s) != Weight::Zero()) {
+      num_final++;
+      final_states.push_back(s);
+    }
+  }
+  if (final_states.size() == 1) {
+    if (fst->Final(final_states[0]) == Weight::One()) {
+      ArcIterator<MutableFst<Arc> > iter(*fst, final_states[0]);
+      if (iter.Done()) {
+        // We already have a final state w/ no transitions out and unit weight.
+        // So we're done.
+        return final_states[0];
+      }
+    }
+  }
+
+  StateId final_state = fst->AddState();
+  fst->SetFinal(final_state, Weight::One());
+  for (size_t idx = 0; idx < final_states.size(); idx++) {
+    StateId s = final_states[idx];
+    Weight weight = fst->Final(s);
+    fst->SetFinal(s, Weight::Zero());
+    Arc arc;
+    arc.ilabel = 0;
+    arc.olabel = 0;
+    arc.nextstate = final_state;
+    arc.weight = weight;
+    fst->AddArc(s, arc);
+  }
+  return final_state;
+}
+
+}  // namespace fst
+
+#endif  // KALDI_FSTEXT_PRE_DETERMINIZE_INL_H_
diff --git a/speechx/speechx/kaldi/fstext/pre-determinize.h b/speechx/speechx/kaldi/fstext/pre-determinize.h
new file mode 100644
index 00000000..a49a0e21
--- /dev/null
+++ b/speechx/speechx/kaldi/fstext/pre-determinize.h
@@ -0,0 +1,98 @@
+// fstext/pre-determinize.h
+
+// Copyright 2009-2011  Microsoft Corporation
+
+// See ../../COPYING for clarification regarding multiple authors
+//
+// Licensed under the Apache License, Version 2.0 (the "License");
+// you may not use this file except in compliance with the License.
+// You may obtain a copy of the License at
+//
+//  http://www.apache.org/licenses/LICENSE-2.0
+//
+// THIS CODE IS PROVIDED *AS IS* BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY
+// KIND, EITHER EXPRESS OR IMPLIED, INCLUDING WITHOUT LIMITATION ANY IMPLIED
+// WARRANTIES OR CONDITIONS OF TITLE, FITNESS FOR A PARTICULAR PURPOSE,
+// MERCHANTABLITY OR NON-INFRINGEMENT.
+// See the Apache 2 License for the specific language governing permissions and
+// limitations under the License.
+
+#ifndef KALDI_FSTEXT_PRE_DETERMINIZE_H_
+#define KALDI_FSTEXT_PRE_DETERMINIZE_H_
+
+#include <fst/fst-decl.h>
+#include <fst/fstlib.h>
+
+#include <algorithm>
+#include <map>
+#include <set>
+#include <string>
+#include <vector>
+
+#include "base/kaldi-common.h"
+
+namespace fst {
+
+/* PreDeterminize inserts extra symbols on the input side of an FST as necessary
+   to ensure that, after epsilon removal, it will be compactly determinizable by
+   the determinize* algorithm.  By compactly determinizable we mean that no
+   original FST state is represented in more than one determinized state).
+
+   Caution: this code is now only used in testing.
+
+   The new symbols start from the value "first_new_symbol", which should be
+   higher than the largest-numbered symbol currently in the FST.  The new
+   symbols added are put in the array syms_out, which should be empty at start.
+*/
+
+template <class Arc, class Int>
+void PreDeterminize(MutableFst<Arc> *fst, typename Arc::Label first_new_symbol,
+                    std::vector<Int> *syms_out);
+
+/* CreateNewSymbols is a helper function used inside PreDeterminize, and is also
+   useful when you need to add a number of extra symbols to a different
+   vocabulary from the one modified by PreDeterminize. */
+
+template <class Label>
+void CreateNewSymbols(SymbolTable *inputSymTable, int nSym, std::string prefix,
+                      std::vector<Label> *syms_out);
+
+/** AddSelfLoops is a function you will probably want to use alongside
+   PreDeterminize, to add self-loops to any FSTs that you compose on the left
+   hand side of the one modified by PreDeterminize.
+
+    This function inserts loops with "special symbols" [e.g. \#0, \#1] into an
+   FST. This is done at each final state and each state with non-epsilon output
+   symbols on at least one arc out of it.  This is to ensure that these symbols,
+   when inserted into the input side of an FST we will compose with on the
+   right, can "pass through" this FST.
+
+    At input, isyms and osyms must be vectors of the same size n, corresponding
+    to symbols that currently do not exist in 'fst'.  For each state in n that
+   has non-epsilon symbols on the output side of arcs leaving it, or which is a
+   final state, this function inserts n self-loops with unit weight and one of
+   the n pairs of symbols on its input and output.
+*/
+template <class Arc>
+void AddSelfLoops(MutableFst<Arc> *fst,
+                  const std::vector<typename Arc::Label> &isyms,
+                  const std::vector<typename Arc::Label> &osyms);
+
+/* DeleteSymbols replaces any instances of symbols in the vector symsIn,
+   appearing on the input side, with epsilon. */
+/* It returns the number of instances of symbols deleted. */
+template <class Arc>
+int64 DeleteISymbols(MutableFst<Arc> *fst,
+                     std::vector<typename Arc::Label> symsIn);
+
+/* CreateSuperFinal takes an FST, and creates an equivalent FST with a single
+   final state with no transitions out and unit final weight, by inserting
+   epsilon transitions as necessary. */
+template <class Arc>
+typename Arc::StateId CreateSuperFinal(MutableFst<Arc> *fst);
+
+}  // end namespace fst
+
+#include "fstext/pre-determinize-inl.h"
+
+#endif  // KALDI_FSTEXT_PRE_DETERMINIZE_H_
diff --git a/speechx/speechx/kaldi/fstext/remove-eps-local-inl.h b/speechx/speechx/kaldi/fstext/remove-eps-local-inl.h
new file mode 100644
index 00000000..94176018
--- /dev/null
+++ b/speechx/speechx/kaldi/fstext/remove-eps-local-inl.h
@@ -0,0 +1,318 @@
+// fstext/remove-eps-local-inl.h
+
+// Copyright 2009-2011  Microsoft Corporation
+//                2014  Johns Hopkins University (author: Daniel Povey
+
+// See ../../COPYING for clarification regarding multiple authors
+//
+// Licensed under the Apache License, Version 2.0 (the "License");
+// you may not use this file except in compliance with the License.
+// You may obtain a copy of the License at
+//
+//  http://www.apache.org/licenses/LICENSE-2.0
+//
+// THIS CODE IS PROVIDED *AS IS* BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY
+// KIND, EITHER EXPRESS OR IMPLIED, INCLUDING WITHOUT LIMITATION ANY IMPLIED
+// WARRANTIES OR CONDITIONS OF TITLE, FITNESS FOR A PARTICULAR PURPOSE,
+// MERCHANTABLITY OR NON-INFRINGEMENT.
+// See the Apache 2 License for the specific language governing permissions and
+// limitations under the License.
+
+#ifndef KALDI_FSTEXT_REMOVE_EPS_LOCAL_INL_H_
+#define KALDI_FSTEXT_REMOVE_EPS_LOCAL_INL_H_
+
+#include <vector>
+
+namespace fst {
+
+template <class Weight>
+struct ReweightPlusDefault {
+  inline Weight operator()(const Weight &a, const Weight &b) {
+    return Plus(a, b);
+  }
+};
+
+struct ReweightPlusLogArc {
+  inline TropicalWeight operator()(const TropicalWeight &a,
+                                   const TropicalWeight &b) {
+    LogWeight a_log(a.Value()), b_log(b.Value());
+    return TropicalWeight(Plus(a_log, b_log).Value());
+  }
+};
+
+template <class Arc,
+          class ReweightPlus = ReweightPlusDefault<typename Arc::Weight> >
+class RemoveEpsLocalClass {
+  typedef typename Arc::StateId StateId;
+  typedef typename Arc::Label Label;
+  typedef typename Arc::Weight Weight;
+
+ public:
+  explicit RemoveEpsLocalClass(MutableFst<Arc> *fst) : fst_(fst) {
+    if (fst_->Start() == kNoStateId) return;  // empty.
+    non_coacc_state_ = fst_->AddState();
+    InitNumArcs();
+    StateId num_states = fst_->NumStates();
+    for (StateId s = 0; s < num_states; s++)
+      for (size_t pos = 0; pos < fst_->NumArcs(s); pos++) RemoveEps(s, pos);
+    assert(CheckNumArcs());
+    Connect(fst);  // remove inaccessible states.
+  }
+
+ private:
+  MutableFst<Arc> *fst_;
+  StateId non_coacc_state_;  //  use this to delete arcs: make it nextstate
+  std::vector<StateId> num_arcs_in_;  // The number of arcs into the state, plus
+                                      // one if it's the start state.
+  std::vector<StateId> num_arcs_out_;  // The number of arcs out of the state,
+                                       // plus one if it's a final state.
+  ReweightPlus reweight_plus_;
+
+  bool CanCombineArcs(const Arc &a, const Arc &b, Arc *c) {
+    if (a.ilabel != 0 && b.ilabel != 0) return false;
+    if (a.olabel != 0 && b.olabel != 0) return false;
+    c->weight = Times(a.weight, b.weight);
+    c->ilabel = (a.ilabel != 0 ? a.ilabel : b.ilabel);
+    c->olabel = (a.olabel != 0 ? a.olabel : b.olabel);
+    c->nextstate = b.nextstate;
+    return true;
+  }
+
+  static bool CanCombineFinal(const Arc &a, Weight final_prob,
+                              Weight *final_prob_out) {
+    if (a.ilabel != 0 || a.olabel != 0) {
+      return false;
+    } else {
+      *final_prob_out = Times(a.weight, final_prob);
+      return true;
+    }
+  }
+
+  void InitNumArcs() {  // init num transitions in/out of each state.
+    StateId num_states = fst_->NumStates();
+    num_arcs_in_.resize(num_states);
+    num_arcs_out_.resize(num_states);
+    num_arcs_in_[fst_->Start()]++;  // count start as trans in.
+    for (StateId s = 0; s < num_states; s++) {
+      if (fst_->Final(s) != Weight::Zero())
+        num_arcs_out_[s]++;  // count final as transition.
+      for (ArcIterator<MutableFst<Arc> > aiter(*fst_, s); !aiter.Done();
+           aiter.Next()) {
+        num_arcs_in_[aiter.Value().nextstate]++;
+        num_arcs_out_[s]++;
+      }
+    }
+  }
+
+  bool CheckNumArcs() {  // check num arcs in/out of each state, at end.  Debug.
+    num_arcs_in_[fst_->Start()]--;  // count start as trans in.
+    StateId num_states = fst_->NumStates();
+    for (StateId s = 0; s < num_states; s++) {
+      if (s == non_coacc_state_) continue;
+      if (fst_->Final(s) != Weight::Zero())
+        num_arcs_out_[s]--;  // count final as transition.
+      for (ArcIterator<MutableFst<Arc> > aiter(*fst_, s); !aiter.Done();
+           aiter.Next()) {
+        if (aiter.Value().nextstate == non_coacc_state_) continue;
+        num_arcs_in_[aiter.Value().nextstate]--;
+        num_arcs_out_[s]--;
+      }
+    }
+    for (StateId s = 0; s < num_states; s++) {
+      assert(num_arcs_in_[s] == 0);
+      assert(num_arcs_out_[s] == 0);
+    }
+    return true;  // always does this.  so we can assert it w/o warnings.
+  }
+
+  inline void GetArc(StateId s, size_t pos, Arc *arc) const {
+    ArcIterator<MutableFst<Arc> > aiter(*fst_, s);
+    aiter.Seek(pos);
+    *arc = aiter.Value();
+  }
+
+  inline void SetArc(StateId s, size_t pos, const Arc &arc) {
+    MutableArcIterator<MutableFst<Arc> > aiter(fst_, s);
+    aiter.Seek(pos);
+    aiter.SetValue(arc);
+  }
+
+  void Reweight(StateId s, size_t pos, Weight reweight) {
+    // Reweight is called from RemoveEpsPattern1; it is a step we
+    // do to preserve stochasticity.  This function multiplies the
+    // arc at (s, pos) by reweight and divides all the arcs [+final-prob]
+    // out of the next state by the same.  This is only valid if
+    // the next state has only one arc in and is not the start state.
+    assert(reweight != Weight::Zero());
+    MutableArcIterator<MutableFst<Arc> > aiter(fst_, s);
+    aiter.Seek(pos);
+    Arc arc = aiter.Value();
+    assert(num_arcs_in_[arc.nextstate] == 1);
+    arc.weight = Times(arc.weight, reweight);
+    aiter.SetValue(arc);
+
+    for (MutableArcIterator<MutableFst<Arc> > aiter_next(fst_, arc.nextstate);
+         !aiter_next.Done(); aiter_next.Next()) {
+      Arc nextarc = aiter_next.Value();
+      if (nextarc.nextstate != non_coacc_state_) {
+        nextarc.weight = Divide(nextarc.weight, reweight, DIVIDE_LEFT);
+        aiter_next.SetValue(nextarc);
+      }
+    }
+    Weight final = fst_->Final(arc.nextstate);
+    if (final != Weight::Zero()) {
+      fst_->SetFinal(arc.nextstate, Divide(final, reweight, DIVIDE_LEFT));
+    }
+  }
+
+  // RemoveEpsPattern1 applies where this arc, which is not a
+  // self-loop, enters a state which has only one input transition
+  // [and is not the start state], and has multiple output
+  // transitions [counting being the final-state as a final-transition].
+
+  void RemoveEpsPattern1(StateId s, size_t pos, Arc arc) {
+    const StateId nextstate = arc.nextstate;
+    Weight total_removed = Weight::Zero(),
+           total_kept = Weight::Zero();  // totals out of nextstate.
+    std::vector<Arc> arcs_to_add;        // to add to state s.
+    for (MutableArcIterator<MutableFst<Arc> > aiter_next(fst_, nextstate);
+         !aiter_next.Done(); aiter_next.Next()) {
+      Arc nextarc = aiter_next.Value();
+      if (nextarc.nextstate == non_coacc_state_) continue;  // deleted.
+      Arc combined;
+      if (CanCombineArcs(arc, nextarc, &combined)) {
+        total_removed = reweight_plus_(total_removed, nextarc.weight);
+        num_arcs_out_[nextstate]--;
+        num_arcs_in_[nextarc.nextstate]--;
+        nextarc.nextstate = non_coacc_state_;
+        aiter_next.SetValue(nextarc);
+        arcs_to_add.push_back(combined);
+      } else {
+        total_kept = reweight_plus_(total_kept, nextarc.weight);
+      }
+    }
+
+    {  // now final-state.
+      Weight next_final = fst_->Final(nextstate);
+      if (next_final != Weight::Zero()) {
+        Weight new_final;
+        if (CanCombineFinal(arc, next_final, &new_final)) {
+          total_removed = reweight_plus_(total_removed, next_final);
+          if (fst_->Final(s) == Weight::Zero())
+            num_arcs_out_[s]++;  // final is counted as arc.
+          fst_->SetFinal(s, Plus(fst_->Final(s), new_final));
+          num_arcs_out_[nextstate]--;
+          fst_->SetFinal(nextstate, Weight::Zero());
+        } else {
+          total_kept = reweight_plus_(total_kept, next_final);
+        }
+      }
+    }
+
+    if (total_removed != Weight::Zero()) {  // did something...
+      if (total_kept == Weight::Zero()) {   // removed everything: remove arc.
+        num_arcs_out_[s]--;
+        num_arcs_in_[arc.nextstate]--;
+        arc.nextstate = non_coacc_state_;
+        SetArc(s, pos, arc);
+      } else {
+        // Have to reweight.
+        Weight total = reweight_plus_(total_removed, total_kept);
+        Weight reweight = Divide(total_kept, total, DIVIDE_LEFT);  // <=1
+        Reweight(s, pos, reweight);
+      }
+    }
+    // Now add the arcs we were going to add.
+    for (size_t i = 0; i < arcs_to_add.size(); i++) {
+      num_arcs_out_[s]++;
+      num_arcs_in_[arcs_to_add[i].nextstate]++;
+      fst_->AddArc(s, arcs_to_add[i]);
+    }
+  }
+
+  void RemoveEpsPattern2(StateId s, size_t pos, Arc arc) {
+    // Pattern 2 is where "nextstate" has only one arc out, counting
+    // being-the-final-state as an arc, but possibly multiple arcs in.
+    // Also, nextstate != s.
+
+    const StateId nextstate = arc.nextstate;
+    bool can_delete_next = (num_arcs_in_[nextstate] == 1);  // if
+    // we combine, can delete the corresponding out-arc/final-prob
+    // of nextstate.
+    bool delete_arc = false;  // set to true if this arc to be deleted.
+
+    Weight next_final = fst_->Final(arc.nextstate);
+    if (next_final !=
+        Weight::Zero()) {  // nextstate has no actual arcs out, only final-prob.
+      Weight new_final;
+      if (CanCombineFinal(arc, next_final, &new_final)) {
+        if (fst_->Final(s) == Weight::Zero())
+          num_arcs_out_[s]++;  // final is counted as arc.
+        fst_->SetFinal(s, Plus(fst_->Final(s), new_final));
+        delete_arc = true;  // will delete "arc".
+        if (can_delete_next) {
+          num_arcs_out_[nextstate]--;
+          fst_->SetFinal(nextstate, Weight::Zero());
+        }
+      }
+    } else {  // has an arc but no final prob.
+      MutableArcIterator<MutableFst<Arc> > aiter_next(fst_, nextstate);
+      assert(!aiter_next.Done());
+      while (aiter_next.Value().nextstate == non_coacc_state_) {
+        aiter_next.Next();
+        assert(!aiter_next.Done());
+      }
+      // now aiter_next points to a real arc out of nextstate.
+      Arc nextarc = aiter_next.Value();
+      Arc combined;
+      if (CanCombineArcs(arc, nextarc, &combined)) {
+        delete_arc = true;
+        if (can_delete_next) {  // do it before we invalidate iterators
+          num_arcs_out_[nextstate]--;
+          num_arcs_in_[nextarc.nextstate]--;
+          nextarc.nextstate = non_coacc_state_;
+          aiter_next.SetValue(nextarc);
+        }
+        num_arcs_out_[s]++;
+        num_arcs_in_[combined.nextstate]++;
+        fst_->AddArc(s, combined);
+      }
+    }
+    if (delete_arc) {
+      num_arcs_out_[s]--;
+      num_arcs_in_[nextstate]--;
+      arc.nextstate = non_coacc_state_;
+      SetArc(s, pos, arc);
+    }
+  }
+
+  void RemoveEps(StateId s, size_t pos) {
+    // Tries to do local epsilon-removal for arc sequences starting with this
+    // arc
+    Arc arc;
+    GetArc(s, pos, &arc);
+    StateId nextstate = arc.nextstate;
+    if (nextstate == non_coacc_state_) return;  // deleted arc.
+    if (nextstate == s) return;  // don't handle self-loops: too complex.
+
+    if (num_arcs_in_[nextstate] == 1 && num_arcs_out_[nextstate] > 1) {
+      RemoveEpsPattern1(s, pos, arc);
+    } else if (num_arcs_out_[nextstate] == 1) {
+      RemoveEpsPattern2(s, pos, arc);
+    }
+  }
+};
+
+template <class Arc>
+void RemoveEpsLocal(MutableFst<Arc> *fst) {
+  RemoveEpsLocalClass<Arc> c(fst);  // work gets done in initializer.
+}
+
+void RemoveEpsLocalSpecial(MutableFst<StdArc> *fst) {
+  // work gets done in initializer.
+  RemoveEpsLocalClass<StdArc, ReweightPlusLogArc> c(fst);
+}
+
+}  // end namespace fst.
+
+#endif  // KALDI_FSTEXT_REMOVE_EPS_LOCAL_INL_H_
diff --git a/speechx/speechx/kaldi/fstext/remove-eps-local.h b/speechx/speechx/kaldi/fstext/remove-eps-local.h
new file mode 100644
index 00000000..a2c76d85
--- /dev/null
+++ b/speechx/speechx/kaldi/fstext/remove-eps-local.h
@@ -0,0 +1,57 @@
+// fstext/remove-eps-local.h
+
+// Copyright 2009-2011  Microsoft Corporation
+//                2014  Johns Hopkins University (author: Daniel Povey)
+
+// See ../../COPYING for clarification regarding multiple authors
+//
+// Licensed under the Apache License, Version 2.0 (the "License");
+// you may not use this file except in compliance with the License.
+// You may obtain a copy of the License at
+//
+//  http://www.apache.org/licenses/LICENSE-2.0
+//
+// THIS CODE IS PROVIDED *AS IS* BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY
+// KIND, EITHER EXPRESS OR IMPLIED, INCLUDING WITHOUT LIMITATION ANY IMPLIED
+// WARRANTIES OR CONDITIONS OF TITLE, FITNESS FOR A PARTICULAR PURPOSE,
+// MERCHANTABLITY OR NON-INFRINGEMENT.
+// See the Apache 2 License for the specific language governing permissions and
+// limitations under the License.
+
+#ifndef KALDI_FSTEXT_REMOVE_EPS_LOCAL_H_
+#define KALDI_FSTEXT_REMOVE_EPS_LOCAL_H_
+
+#include <fst/fst-decl.h>
+#include <fst/fstlib.h>
+
+namespace fst {
+
+/// RemoveEpsLocal remove some (but not necessarily all) epsilons in an FST,
+/// using an algorithm that is guaranteed to never increase the number of arcs
+/// in the FST (and will also never increase the number of states).  The
+/// algorithm is not optimal but is reasonably clever.  It does not just remove
+/// epsilon arcs;it also combines pairs of input-epsilon and output-epsilon arcs
+/// into one.
+/// The algorithm preserves equivalence and stochasticity in the given semiring.
+/// If you want to preserve stochasticity in a different semiring (e.g. log),
+/// then use RemoveEpsLocalSpecial, which only works for StdArc but which
+/// preserves stochasticity, where possible (*) in the LogArc sense.  The reason
+/// that we can't just cast to a different semiring is that in that case we
+/// would no longer be able to guarantee equivalence in the original semiring
+/// (this arises from what happens when we combine identical arcs).
+/// (*) by "where possible".. there are situations where we wouldn't be able to
+/// preserve stochasticity in the LogArc sense while maintaining equivalence in
+/// the StdArc sense, so in these situations we maintain equivalence.
+
+template <class Arc>
+void RemoveEpsLocal(MutableFst<Arc> *fst);
+
+/// As RemoveEpsLocal but takes care to preserve stochasticity
+/// when cast to LogArc.
+inline void RemoveEpsLocalSpecial(MutableFst<StdArc> *fst);
+
+}  // namespace fst
+
+#include "fstext/remove-eps-local-inl.h"
+
+#endif  // KALDI_FSTEXT_REMOVE_EPS_LOCAL_H_
diff --git a/speechx/speechx/kaldi/fstext/table-matcher.h b/speechx/speechx/kaldi/fstext/table-matcher.h
new file mode 100644
index 00000000..0b3f695d
--- /dev/null
+++ b/speechx/speechx/kaldi/fstext/table-matcher.h
@@ -0,0 +1,387 @@
+// fstext/table-matcher.h
+
+// Copyright 2009-2011  Microsoft Corporation
+
+// See ../../COPYING for clarification regarding multiple authors
+//
+// Licensed under the Apache License, Version 2.0 (the "License");
+// you may not use this file except in compliance with the License.
+// You may obtain a copy of the License at
+//
+//  http://www.apache.org/licenses/LICENSE-2.0
+//
+// THIS CODE IS PROVIDED *AS IS* BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY
+// KIND, EITHER EXPRESS OR IMPLIED, INCLUDING WITHOUT LIMITATION ANY IMPLIED
+// WARRANTIES OR CONDITIONS OF TITLE, FITNESS FOR A PARTICULAR PURPOSE,
+// MERCHANTABLITY OR NON-INFRINGEMENT.
+// See the Apache 2 License for the specific language governing permissions and
+// limitations under the License.
+
+#ifndef KALDI_FSTEXT_TABLE_MATCHER_H_
+#define KALDI_FSTEXT_TABLE_MATCHER_H_
+
+#include <fst/fst-decl.h>
+#include <fst/fstlib.h>
+
+#include <memory>
+#include <vector>
+
+namespace fst {
+
+/// TableMatcher is a matcher specialized for the case where the output
+/// side of the left FST always has either all-epsilons coming out of
+/// a state, or a majority of the symbol table.  Therefore we can
+/// either store nothing (for the all-epsilon case) or store a lookup
+/// table from Labels to arc offsets.  Since the TableMatcher has to
+/// iterate over all arcs in each left-hand state the first time it sees
+/// it, this matcher type is not efficient if you compose with
+/// something very small on the right-- unless you do it multiple
+/// times and keep the matcher around. To do this requires using the
+/// most advanced form of ComposeFst in Compose.h, that initializes
+/// with ComposeFstImplOptions.
+
+struct TableMatcherOptions {
+  float
+      table_ratio;  // we construct the table if it would be at least this full.
+  int min_table_size;
+  TableMatcherOptions() : table_ratio(0.25), min_table_size(4) {}
+};
+
+// Introducing an "impl" class for TableMatcher because
+// we need to do a shallow copy of the Matcher for when
+// we want to cache tables for multiple compositions.
+template <class F, class BackoffMatcher = SortedMatcher<F> >
+class TableMatcherImpl : public MatcherBase<typename F::Arc> {
+ public:
+  typedef F FST;
+  typedef typename F::Arc Arc;
+  typedef typename Arc::Label Label;
+  typedef typename Arc::StateId StateId;
+  typedef StateId
+      ArcId;  // Use this type to store arc offsets [it's actually size_t
+  // in the Seek function of ArcIterator, but StateId should be big enough].
+  typedef typename Arc::Weight Weight;
+
+ public:
+  TableMatcherImpl(const FST &fst, MatchType match_type,
+                   const TableMatcherOptions &opts = TableMatcherOptions())
+      : match_type_(match_type),
+        fst_(fst.Copy()),
+        loop_(match_type == MATCH_INPUT
+                  ? Arc(kNoLabel, 0, Weight::One(), kNoStateId)
+                  : Arc(0, kNoLabel, Weight::One(), kNoStateId)),
+        aiter_(NULL),
+        s_(kNoStateId),
+        opts_(opts),
+        backoff_matcher_(fst, match_type) {
+    assert(opts_.min_table_size > 0);
+    if (match_type == MATCH_INPUT)
+      assert(fst_->Properties(kILabelSorted, true) == kILabelSorted);
+    else if (match_type == MATCH_OUTPUT)
+      assert(fst_->Properties(kOLabelSorted, true) == kOLabelSorted);
+    else
+      assert(0 && "Invalid FST properties");
+  }
+
+  virtual const FST &GetFst() const { return *fst_; }
+
+  virtual ~TableMatcherImpl() {
+    std::vector<ArcId> *const empty =
+        ((std::vector<ArcId> *)(NULL)) + 1;  // special marker.
+    for (size_t i = 0; i < tables_.size(); i++) {
+      if (tables_[i] != NULL && tables_[i] != empty) delete tables_[i];
+    }
+    delete aiter_;
+    delete fst_;
+  }
+
+  virtual MatchType Type(bool test) const { return match_type_; }
+
+  void SetState(StateId s) {
+    if (aiter_) {
+      delete aiter_;
+      aiter_ = NULL;
+    }
+    if (match_type_ == MATCH_NONE) LOG(FATAL) << "TableMatcher: bad match type";
+    s_ = s;
+    std::vector<ArcId> *const empty =
+        ((std::vector<ArcId> *)(NULL)) + 1;  // special marker.
+    if (static_cast<size_t>(s) >= tables_.size()) {
+      assert(s >= 0);
+      tables_.resize(s + 1, NULL);
+    }
+    std::vector<ArcId> *&this_table_ = tables_[s];  // note: ref to ptr.
+    if (this_table_ == empty) {
+      backoff_matcher_.SetState(s);
+      return;
+    } else if (this_table_ == NULL) {  // NULL means has not been set.
+      ArcId num_arcs = fst_->NumArcs(s);
+      if (num_arcs == 0 || num_arcs < opts_.min_table_size) {
+        this_table_ = empty;
+        backoff_matcher_.SetState(s);
+        return;
+      }
+      ArcIterator<FST> aiter(*fst_, s);
+      aiter.SetFlags(
+          kArcNoCache |
+              (match_type_ == MATCH_OUTPUT ? kArcOLabelValue : kArcILabelValue),
+          kArcNoCache | kArcValueFlags);
+      // the statement above, says: "Don't cache stuff; and I only need the
+      // ilabel/olabel to be computed.
+      aiter.Seek(num_arcs - 1);
+      Label highest_label =
+          (match_type_ == MATCH_OUTPUT ? aiter.Value().olabel
+                                       : aiter.Value().ilabel);
+      if ((highest_label + 1) * opts_.table_ratio > num_arcs) {
+        this_table_ = empty;
+        backoff_matcher_.SetState(s);
+        return;  // table would be too sparse.
+      }
+      // OK, now we are creating the table.
+      this_table_ = new std::vector<ArcId>(highest_label + 1, kNoStateId);
+      ArcId pos = 0;
+      for (aiter.Seek(0); !aiter.Done(); aiter.Next(), pos++) {
+        Label label = (match_type_ == MATCH_OUTPUT ? aiter.Value().olabel
+                                                   : aiter.Value().ilabel);
+        assert(static_cast<size_t>(label) <=
+               static_cast<size_t>(highest_label));  // also checks >= 0.
+        if ((*this_table_)[label] == kNoStateId) (*this_table_)[label] = pos;
+        // set this_table_[label] to first position where arc has this
+        // label.
+      }
+    }
+    // At this point in the code, this_table_ != NULL and != empty.
+    aiter_ = new ArcIterator<FST>(*fst_, s);
+    aiter_->SetFlags(kArcNoCache,
+                     kArcNoCache);  // don't need to cache arcs as may only
+    // need a small subset.
+    loop_.nextstate = s;
+    // aiter_ = NULL;
+    // backoff_matcher_.SetState(s);
+  }
+
+  bool Find(Label match_label) {
+    if (!aiter_) {
+      return backoff_matcher_.Find(match_label);
+    } else {
+      match_label_ = match_label;
+      current_loop_ = (match_label == 0);
+      // kNoLabel means the implicit loop on the other FST --
+      // matches real epsilons but not the self-loop.
+      match_label_ = (match_label_ == kNoLabel ? 0 : match_label_);
+      if (static_cast<size_t>(match_label_) < tables_[s_]->size() &&
+          (*(tables_[s_]))[match_label_] != kNoStateId) {
+        aiter_->Seek((*(tables_[s_]))[match_label_]);  // label exists.
+        return true;
+      }
+      return current_loop_;
+    }
+  }
+  const Arc &Value() const {
+    if (aiter_)
+      return current_loop_ ? loop_ : aiter_->Value();
+    else
+      return backoff_matcher_.Value();
+  }
+
+  void Next() {
+    if (aiter_) {
+      if (current_loop_)
+        current_loop_ = false;
+      else
+        aiter_->Next();
+    } else {
+      backoff_matcher_.Next();
+    }
+  }
+
+  bool Done() const {
+    if (aiter_ != NULL) {
+      if (current_loop_) return false;
+      if (aiter_->Done()) return true;
+      Label label = (match_type_ == MATCH_OUTPUT ? aiter_->Value().olabel
+                                                 : aiter_->Value().ilabel);
+      return (label != match_label_);
+    } else {
+      return backoff_matcher_.Done();
+    }
+  }
+  const Arc &Value() {
+    if (aiter_ != NULL) {
+      return (current_loop_ ? loop_ : aiter_->Value());
+    } else {
+      return backoff_matcher_.Value();
+    }
+  }
+
+  virtual TableMatcherImpl<FST> *Copy(bool safe = false) const {
+    assert(0);  // shouldn't be called.  This is not a "real" matcher,
+    // although we derive from MatcherBase for convenience.
+    return NULL;
+  }
+
+  virtual uint64 Properties(uint64 props) const {
+    return props;
+  }  // simple matcher that does
+     // not change its FST, so properties are properties of FST it is applied to
+
+ private:
+  virtual void SetState_(StateId s) { SetState(s); }
+  virtual bool Find_(Label label) { return Find(label); }
+  virtual bool Done_() const { return Done(); }
+  virtual const Arc &Value_() const { return Value(); }
+  virtual void Next_() { Next(); }
+
+  MatchType match_type_;
+  FST *fst_;
+  bool current_loop_;
+  Label match_label_;
+  Arc loop_;
+  ArcIterator<FST> *aiter_;
+  StateId s_;
+  std::vector<std::vector<ArcId> *> tables_;
+  TableMatcherOptions opts_;
+  BackoffMatcher backoff_matcher_;
+};
+
+template <class F, class BackoffMatcher = SortedMatcher<F> >
+class TableMatcher : public MatcherBase<typename F::Arc> {
+ public:
+  typedef F FST;
+  typedef typename F::Arc Arc;
+  typedef typename Arc::Label Label;
+  typedef typename Arc::StateId StateId;
+  typedef StateId
+      ArcId;  // Use this type to store arc offsets [it's actually size_t
+  // in the Seek function of ArcIterator, but StateId should be big enough].
+  typedef typename Arc::Weight Weight;
+  typedef TableMatcherImpl<F, BackoffMatcher> Impl;
+
+  TableMatcher(const FST &fst, MatchType match_type,
+               const TableMatcherOptions &opts = TableMatcherOptions())
+      : impl_(std::make_shared<Impl>(fst, match_type, opts)) {}
+
+  TableMatcher(const TableMatcher<FST, BackoffMatcher> &matcher,
+               bool safe = false)
+      : impl_(matcher.impl_) {
+    if (safe == true) {
+      LOG(FATAL) << "TableMatcher: Safe copy not supported";
+    }
+  }
+
+  virtual const FST &GetFst() const { return impl_->GetFst(); }
+
+  virtual MatchType Type(bool test) const { return impl_->Type(test); }
+
+  void SetState(StateId s) { return impl_->SetState(s); }
+
+  bool Find(Label match_label) { return impl_->Find(match_label); }
+
+  const Arc &Value() const { return impl_->Value(); }
+
+  void Next() { return impl_->Next(); }
+
+  bool Done() const { return impl_->Done(); }
+
+  const Arc &Value() { return impl_->Value(); }
+
+  virtual TableMatcher<FST, BackoffMatcher> *Copy(bool safe = false) const {
+    return new TableMatcher<FST, BackoffMatcher>(*this, safe);
+  }
+
+  virtual uint64 Properties(uint64 props) const {
+    return impl_->Properties(props);
+  }  // simple matcher that does
+     // not change its FST, so properties are properties of FST it is applied to
+ private:
+  std::shared_ptr<Impl> impl_;
+
+  virtual void SetState_(StateId s) { impl_->SetState(s); }
+  virtual bool Find_(Label label) { return impl_->Find(label); }
+  virtual bool Done_() const { return impl_->Done(); }
+  virtual const Arc &Value_() const { return impl_->Value(); }
+  virtual void Next_() { impl_->Next(); }
+
+  TableMatcher &operator=(const TableMatcher &) = delete;
+};
+
+struct TableComposeOptions : public TableMatcherOptions {
+  bool connect;               // Connect output
+  ComposeFilter filter_type;  // Which pre-defined filter to use
+  MatchType table_match_type;
+
+  explicit TableComposeOptions(const TableMatcherOptions &mo, bool c = true,
+                               ComposeFilter ft = SEQUENCE_FILTER,
+                               MatchType tms = MATCH_OUTPUT)
+      : TableMatcherOptions(mo),
+        connect(c),
+        filter_type(ft),
+        table_match_type(tms) {}
+  TableComposeOptions()
+      : connect(true),
+        filter_type(SEQUENCE_FILTER),
+        table_match_type(MATCH_OUTPUT) {}
+};
+
+template <class Arc>
+void TableCompose(const Fst<Arc> &ifst1, const Fst<Arc> &ifst2,
+                  MutableFst<Arc> *ofst,
+                  const TableComposeOptions &opts = TableComposeOptions()) {
+  typedef Fst<Arc> F;
+  CacheOptions nopts;
+  nopts.gc_limit = 0;  // Cache only the last state for fastest copy.
+  if (opts.table_match_type == MATCH_OUTPUT) {
+    // ComposeFstImplOptions templated on matcher for fst1, matcher for fst2.
+    ComposeFstImplOptions<TableMatcher<F>, SortedMatcher<F> > impl_opts(nopts);
+    impl_opts.matcher1 = new TableMatcher<F>(ifst1, MATCH_OUTPUT, opts);
+    *ofst = ComposeFst<Arc>(ifst1, ifst2, impl_opts);
+  } else {
+    assert(opts.table_match_type == MATCH_INPUT);
+    // ComposeFstImplOptions templated on matcher for fst1, matcher for fst2.
+    ComposeFstImplOptions<SortedMatcher<F>, TableMatcher<F> > impl_opts(nopts);
+    impl_opts.matcher2 = new TableMatcher<F>(ifst2, MATCH_INPUT, opts);
+    *ofst = ComposeFst<Arc>(ifst1, ifst2, impl_opts);
+  }
+  if (opts.connect) Connect(ofst);
+}
+
+/// TableComposeCache lets us do multiple compositions while caching the same
+/// matcher.
+template <class F>
+struct TableComposeCache {
+  TableMatcher<F> *matcher;
+  TableComposeOptions opts;
+  explicit TableComposeCache(
+      const TableComposeOptions &opts = TableComposeOptions())
+      : matcher(NULL), opts(opts) {}
+  ~TableComposeCache() { delete (matcher); }
+};
+
+template <class Arc>
+void TableCompose(const Fst<Arc> &ifst1, const Fst<Arc> &ifst2,
+                  MutableFst<Arc> *ofst, TableComposeCache<Fst<Arc> > *cache) {
+  typedef Fst<Arc> F;
+  assert(cache != NULL);
+  CacheOptions nopts;
+  nopts.gc_limit = 0;  // Cache only the last state for fastest copy.
+  if (cache->opts.table_match_type == MATCH_OUTPUT) {
+    ComposeFstImplOptions<TableMatcher<F>, SortedMatcher<F> > impl_opts(nopts);
+    if (cache->matcher == NULL)
+      cache->matcher = new TableMatcher<F>(ifst1, MATCH_OUTPUT, cache->opts);
+    impl_opts.matcher1 = cache->matcher->Copy();  // not passing "safe": may not
+    // be thread-safe-- anway I don't understand this part.
+    *ofst = ComposeFst<Arc>(ifst1, ifst2, impl_opts);
+  } else {
+    assert(cache->opts.table_match_type == MATCH_INPUT);
+    ComposeFstImplOptions<SortedMatcher<F>, TableMatcher<F> > impl_opts(nopts);
+    if (cache->matcher == NULL)
+      cache->matcher = new TableMatcher<F>(ifst2, MATCH_INPUT, cache->opts);
+    impl_opts.matcher2 = cache->matcher->Copy();
+    *ofst = ComposeFst<Arc>(ifst1, ifst2, impl_opts);
+  }
+  if (cache->opts.connect) Connect(ofst);
+}
+
+}  // namespace fst
+
+#endif  // KALDI_FSTEXT_TABLE_MATCHER_H_
diff --git a/speechx/speechx/kaldi/lat/CMakeLists.txt b/speechx/speechx/kaldi/lat/CMakeLists.txt
new file mode 100644
index 00000000..f22ce8f0
--- /dev/null
+++ b/speechx/speechx/kaldi/lat/CMakeLists.txt
@@ -0,0 +1,6 @@
+
+add_library(kaldi-lat
+determinize-lattice-pruned.cc
+lattice-functions.cc
+)
+target_link_libraries(kaldi-lat PUBLIC kaldi-util)
\ No newline at end of file
diff --git a/speechx/speechx/kaldi/lat/determinize-lattice-pruned-test.cc b/speechx/speechx/kaldi/lat/determinize-lattice-pruned-test.cc
deleted file mode 100644
index f6684f0b..00000000
--- a/speechx/speechx/kaldi/lat/determinize-lattice-pruned-test.cc
+++ /dev/null
@@ -1,147 +0,0 @@
-// lat/determinize-lattice-pruned-test.cc
-
-// Copyright 2009-2012  Microsoft Corporation
-//           2012-2013  Johns Hopkins University (Author: Daniel Povey)
-
-// See ../../COPYING for clarification regarding multiple authors
-//
-// Licensed under the Apache License, Version 2.0 (the "License");
-// you may not use this file except in compliance with the License.
-// You may obtain a copy of the License at
-//
-//  http://www.apache.org/licenses/LICENSE-2.0
-//
-// THIS CODE IS PROVIDED *AS IS* BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY
-// KIND, EITHER EXPRESS OR IMPLIED, INCLUDING WITHOUT LIMITATION ANY IMPLIED
-// WARRANTIES OR CONDITIONS OF TITLE, FITNESS FOR A PARTICULAR PURPOSE,
-// MERCHANTABLITY OR NON-INFRINGEMENT.
-// See the Apache 2 License for the specific language governing permissions and
-// limitations under the License.
-
-#include "lat/determinize-lattice-pruned.h"
-#include "fstext/lattice-utils.h"
-#include "fstext/fst-test-utils.h"
-#include "lat/kaldi-lattice.h"
-#include "lat/lattice-functions.h"
-
-namespace fst {
-// Caution: these tests are not as generic as you might think from all the
-// templates in the code.  They are basically only valid for LatticeArc.
-// This is partly due to the fact that certain templates need to be instantiated
-// in other .cc files in this directory.
-
-// test that determinization proceeds correctly on general
-// FSTs (not guaranteed determinzable, but we use the
-// max-states option to stop it getting out of control).
-template<class Arc> void TestDeterminizeLatticePruned() {
-  typedef kaldi::int32 Int;
-  typedef typename Arc::Weight Weight;
-  typedef ArcTpl<CompactLatticeWeightTpl<Weight, Int> > CompactArc;
-
-  for(int i = 0; i < 100; i++) {
-    RandFstOptions opts;
-    opts.n_states = 4;
-    opts.n_arcs = 10;
-    opts.n_final = 2;
-    opts.allow_empty = false;
-    opts.weight_multiplier = 0.5; // impt for the randomly generated weights
-    opts.acyclic = true;
-    // to be exactly representable in float,
-    // or this test fails because numerical differences can cause symmetry in
-    // weights to be broken, which causes the wrong path to be chosen as far
-    // as the string part is concerned.
-
-    VectorFst<Arc> *fst = RandPairFst<Arc>(opts);
-
-    bool sorted = TopSort(fst);
-    KALDI_ASSERT(sorted);
-
-    ILabelCompare<Arc> ilabel_comp;
-    if (kaldi::Rand() % 2 == 0)
-      ArcSort(fst, ilabel_comp);
-
-    std::cout << "FST before lattice-determinizing is:\n";
-    {
-      FstPrinter<Arc> fstprinter(*fst, NULL, NULL, NULL, false, true, "\t");
-      fstprinter.Print(&std::cout, "standard output");
-    }
-    VectorFst<Arc> det_fst;
-    try {
-      DeterminizeLatticePrunedOptions lat_opts;
-      lat_opts.max_mem = ((kaldi::Rand() % 2 == 0) ? 100 : 1000);
-      lat_opts.max_states = ((kaldi::Rand() % 2 == 0) ? -1 : 20);
-      lat_opts.max_arcs = ((kaldi::Rand() % 2 == 0) ? -1 : 30);
-      bool ans = DeterminizeLatticePruned<Weight>(*fst, 10.0, &det_fst, lat_opts);
-
-      std::cout << "FST after lattice-determinizing is:\n";
-      {
-        FstPrinter<Arc> fstprinter(det_fst, NULL, NULL, NULL, false, true, "\t");
-        fstprinter.Print(&std::cout, "standard output");
-      }
-      KALDI_ASSERT(det_fst.Properties(kIDeterministic, true) & kIDeterministic);
-      // OK, now determinize it a different way and check equivalence.
-      // [note: it's not normal determinization, it's taking the best path
-      // for any input-symbol sequence....
-
-
-      VectorFst<Arc> pruned_fst(*fst);
-      if (pruned_fst.NumStates() != 0)
-        kaldi::PruneLattice(10.0, &pruned_fst);
-
-      VectorFst<CompactArc> compact_pruned_fst, compact_pruned_det_fst;
-      ConvertLattice<Weight, Int>(pruned_fst, &compact_pruned_fst, false);
-      std::cout << "Compact pruned FST is:\n";
-      {
-        FstPrinter<CompactArc> fstprinter(compact_pruned_fst, NULL, NULL, NULL, false, true, "\t");
-        fstprinter.Print(&std::cout, "standard output");
-      }
-      ConvertLattice<Weight, Int>(det_fst, &compact_pruned_det_fst, false);
-
-      std::cout << "Compact version of determinized FST is:\n";
-      {
-        FstPrinter<CompactArc> fstprinter(compact_pruned_det_fst, NULL, NULL, NULL, false, true, "\t");
-        fstprinter.Print(&std::cout, "standard output");
-      }
-
-      if (ans)
-        KALDI_ASSERT(RandEquivalent(compact_pruned_det_fst, compact_pruned_fst, 5/*paths*/, 0.01/*delta*/, kaldi::Rand()/*seed*/, 100/*path length, max*/));
-    } catch (...) {
-      std::cout << "Failed to lattice-determinize this FST (probably not determinizable)\n";
-    }
-    delete fst;
-  }
-}
-
-// test that determinization proceeds without crash on acyclic FSTs
-// (guaranteed determinizable in this sense).
-template<class Arc> void TestDeterminizeLatticePruned2() {
-  typedef typename Arc::Weight Weight;
-  RandFstOptions opts;
-  opts.acyclic = true;
-  for(int i = 0; i < 100; i++) {
-    VectorFst<Arc> *fst = RandPairFst<Arc>(opts);
-    std::cout << "FST before lattice-determinizing is:\n";
-    {
-      FstPrinter<Arc> fstprinter(*fst, NULL, NULL, NULL, false, true, "\t");
-      fstprinter.Print(&std::cout, "standard output");
-    }
-    VectorFst<Arc> ofst;
-    DeterminizeLatticePruned<Weight>(*fst, 10.0, &ofst);
-    std::cout << "FST after lattice-determinizing is:\n";
-    {
-      FstPrinter<Arc> fstprinter(ofst, NULL, NULL, NULL, false, true, "\t");
-      fstprinter.Print(&std::cout, "standard output");
-    }
-    delete fst;
-  }
-}
-
-
-} // end namespace fst
-
-int main() {
-  using namespace fst;
-  TestDeterminizeLatticePruned<kaldi::LatticeArc>();
-  TestDeterminizeLatticePruned2<kaldi::LatticeArc>();
-  std::cout << "Tests succeeded\n";
-}
diff --git a/speechx/speechx/kaldi/lat/determinize-lattice-pruned.cc b/speechx/speechx/kaldi/lat/determinize-lattice-pruned.cc
index dbdd9af4..23186a32 100644
--- a/speechx/speechx/kaldi/lat/determinize-lattice-pruned.cc
+++ b/speechx/speechx/kaldi/lat/determinize-lattice-pruned.cc
@@ -24,8 +24,8 @@
 #include "fstext/determinize-lattice.h" // for LatticeStringRepository
 #include "fstext/fstext-utils.h"
 #include "lat/lattice-functions.h"  // for PruneLattice
-#include "lat/minimize-lattice.h"   // for minimization
-#include "lat/push-lattice.h"       // for minimization
+// #include "lat/minimize-lattice.h"   // for minimization
+// #include "lat/push-lattice.h"       // for minimization
 #include "lat/determinize-lattice-pruned.h"
 
 namespace fst {
@@ -223,6 +223,10 @@ template<class Weight, class IntType> class LatticeDeterminizerPruned {
          iter != initial_hash_.end(); ++iter)
       delete iter->first;
     { InitialSubsetHash tmp; tmp.swap(initial_hash_); }
+    for (size_t i = 0; i < output_states_.size(); i++) {
+      vector<Element> tmp;
+      tmp.swap(output_states_[i]->minimal_subset);
+    }
     { vector<char> tmp;  tmp.swap(isymbol_or_final_); }
     { // Free up the queue.  I'm not sure how to make sure all
       // the memory is really freed (no swap() function)... doesn't really
@@ -1288,222 +1292,222 @@ bool DeterminizeLatticePruned(const ExpandedFst<ArcTpl<Weight> > &ifst,
   return false; // Suppress compiler warning; this code is unreachable.
 }
 
-template<class Weight>
-typename ArcTpl<Weight>::Label DeterminizeLatticeInsertPhones(
-    const kaldi::TransitionInformation &trans_model,
-    MutableFst<ArcTpl<Weight> > *fst) {
-  // Define some types.
-  typedef ArcTpl<Weight> Arc;
-  typedef typename Arc::StateId StateId;
-  typedef typename Arc::Label Label;
-
-  // Work out the first phone symbol. This is more related to the phone
-  // insertion function, so we put it here and make it the returning value of
-  // DeterminizeLatticeInsertPhones().
-  Label first_phone_label = HighestNumberedInputSymbol(*fst) + 1;
-
-  // Insert phones here.
-  for (StateIterator<MutableFst<Arc> > siter(*fst);
-       !siter.Done(); siter.Next()) {
-    StateId state = siter.Value();
-    if (state == fst->Start())
-      continue;
-    for (MutableArcIterator<MutableFst<Arc> > aiter(fst, state);
-         !aiter.Done(); aiter.Next()) {
-      Arc arc = aiter.Value();
-
-      // Note: the words are on the input symbol side and transition-id's are on
-      // the output symbol side.
-      if ((arc.olabel != 0)
-          && (trans_model.TransitionIdIsStartOfPhone(arc.olabel))
-          && (!trans_model.IsSelfLoop(arc.olabel))) {
-        Label phone =
-            static_cast<Label>(trans_model.TransitionIdToPhone(arc.olabel));
-
-        // Skips <eps>.
-        KALDI_ASSERT(phone != 0);
-
-        if (arc.ilabel == 0) {
-          // If there is no word on the arc, insert the phone directly.
-          arc.ilabel = first_phone_label + phone;
-        } else {
-          // Otherwise, add an additional arc.
-          StateId additional_state = fst->AddState();
-          StateId next_state = arc.nextstate;
-          arc.nextstate = additional_state;
-          fst->AddArc(additional_state,
-                      Arc(first_phone_label + phone, 0,
-                          Weight::One(), next_state));
-        }
-      }
-
-      aiter.SetValue(arc);
-    }
-  }
-
-  return first_phone_label;
-}
-
-template<class Weight>
-void DeterminizeLatticeDeletePhones(
-    typename ArcTpl<Weight>::Label first_phone_label,
-    MutableFst<ArcTpl<Weight> > *fst) {
-  // Define some types.
-  typedef ArcTpl<Weight> Arc;
-  typedef typename Arc::StateId StateId;
-  typedef typename Arc::Label Label;
-
-  // Delete phones here.
-  for (StateIterator<MutableFst<Arc> > siter(*fst);
-       !siter.Done(); siter.Next()) {
-    StateId state = siter.Value();
-    for (MutableArcIterator<MutableFst<Arc> > aiter(fst, state);
-         !aiter.Done(); aiter.Next()) {
-      Arc arc = aiter.Value();
-
-      if (arc.ilabel >= first_phone_label)
-        arc.ilabel = 0;
-
-      aiter.SetValue(arc);
-    }
-  }
-}
+// template<class Weight>
+// typename ArcTpl<Weight>::Label DeterminizeLatticeInsertPhones(
+//     const kaldi::TransitionModel &trans_model,
+//     MutableFst<ArcTpl<Weight> > *fst) {
+//   // Define some types.
+//   typedef ArcTpl<Weight> Arc;
+//   typedef typename Arc::StateId StateId;
+//   typedef typename Arc::Label Label;
+//
+//   // Work out the first phone symbol. This is more related to the phone
+//   // insertion function, so we put it here and make it the returning value of
+//   // DeterminizeLatticeInsertPhones().
+//   Label first_phone_label = HighestNumberedInputSymbol(*fst) + 1;
+//
+//   // Insert phones here.
+//   for (StateIterator<MutableFst<Arc> > siter(*fst);
+//        !siter.Done(); siter.Next()) {
+//     StateId state = siter.Value();
+//     if (state == fst->Start())
+//       continue;
+//     for (MutableArcIterator<MutableFst<Arc> > aiter(fst, state);
+//          !aiter.Done(); aiter.Next()) {
+//       Arc arc = aiter.Value();
+//
+//       // Note: the words are on the input symbol side and transition-id's are on
+//       // the output symbol side.
+//       if ((arc.olabel != 0)
+//           && (trans_model.TransitionIdToHmmState(arc.olabel) == 0)
+//           && (!trans_model.IsSelfLoop(arc.olabel))) {
+//         Label phone =
+//             static_cast<Label>(trans_model.TransitionIdToPhone(arc.olabel));
+//
+//         // Skips <eps>.
+//         KALDI_ASSERT(phone != 0);
+//
+//         if (arc.ilabel == 0) {
+//           // If there is no word on the arc, insert the phone directly.
+//           arc.ilabel = first_phone_label + phone;
+//         } else {
+//           // Otherwise, add an additional arc.
+//           StateId additional_state = fst->AddState();
+//           StateId next_state = arc.nextstate;
+//           arc.nextstate = additional_state;
+//           fst->AddArc(additional_state,
+//                       Arc(first_phone_label + phone, 0,
+//                           Weight::One(), next_state));
+//         }
+//       }
+//
+//       aiter.SetValue(arc);
+//     }
+//   }
+//
+//   return first_phone_label;
+// }
+//
+// template<class Weight>
+// void DeterminizeLatticeDeletePhones(
+//     typename ArcTpl<Weight>::Label first_phone_label,
+//     MutableFst<ArcTpl<Weight> > *fst) {
+//   // Define some types.
+//   typedef ArcTpl<Weight> Arc;
+//   typedef typename Arc::StateId StateId;
+//   typedef typename Arc::Label Label;
+//
+//   // Delete phones here.
+//   for (StateIterator<MutableFst<Arc> > siter(*fst);
+//        !siter.Done(); siter.Next()) {
+//     StateId state = siter.Value();
+//     for (MutableArcIterator<MutableFst<Arc> > aiter(fst, state);
+//          !aiter.Done(); aiter.Next()) {
+//       Arc arc = aiter.Value();
+//
+//       if (arc.ilabel >= first_phone_label)
+//         arc.ilabel = 0;
+//
+//       aiter.SetValue(arc);
+//     }
+//   }
+// }
 // instantiate for type LatticeWeight
-template
-void DeterminizeLatticeDeletePhones(
-    ArcTpl<kaldi::LatticeWeight>::Label first_phone_label,
-    MutableFst<ArcTpl<kaldi::LatticeWeight> > *fst);
-
-/** This function does a first pass determinization with phone symbols inserted
-    at phone boundary. It uses a transition model to work out the transition-id
-    to phone map. First, phones will be inserted into the word level lattice.
-    Second, determinization will be applied on top of the phone + word lattice.
-    Finally, the inserted phones will be removed, converting the lattice back to
-    a word level lattice. The output lattice of this pass is not deterministic,
-    since we remove the phone symbols as a last step. It is supposed to be
-    followed by another pass of determinization at the word level. It could also
-    be useful for some other applications such as fMLLR estimation, confidence
-    estimation, discriminative training, etc.
-*/
-template<class Weight, class IntType>
-bool DeterminizeLatticePhonePrunedFirstPass(
-    const kaldi::TransitionInformation &trans_model,
-    double beam,
-    MutableFst<ArcTpl<Weight> > *fst,
-    const DeterminizeLatticePrunedOptions &opts) {
-  // First, insert the phones.
-  typename ArcTpl<Weight>::Label first_phone_label =
-      DeterminizeLatticeInsertPhones(trans_model, fst);
-  TopSort(fst);
-
-  // Second, do determinization with phone inserted.
-  bool ans = DeterminizeLatticePruned<Weight>(*fst, beam, fst, opts);
-
-  // Finally, remove the inserted phones.
-  DeterminizeLatticeDeletePhones(first_phone_label, fst);
-  TopSort(fst);
-
-  return ans;
-}
-
-// "Destructive" version of DeterminizeLatticePhonePruned() where the input
-// lattice might be modified.
-template<class Weight, class IntType>
-bool DeterminizeLatticePhonePruned(
-    const kaldi::TransitionInformation &trans_model,
-    MutableFst<ArcTpl<Weight> > *ifst,
-    double beam,
-    MutableFst<ArcTpl<CompactLatticeWeightTpl<Weight, IntType> > > *ofst,
-    DeterminizeLatticePhonePrunedOptions opts) {
-  // Returning status.
-  bool ans = true;
-
-  // Make sure at least one of opts.phone_determinize and opts.word_determinize
-  // is not false, otherwise calling this function doesn't make any sense.
-  if ((opts.phone_determinize || opts.word_determinize) == false) {
-    KALDI_WARN << "Both --phone-determinize and --word-determinize are set to "
-               << "false, copying lattice without determinization.";
-    // We are expecting the words on the input side.
-    ConvertLattice<Weight, IntType>(*ifst, ofst, false);
-    return ans;
-  }
-
-  // Determinization options.
-  DeterminizeLatticePrunedOptions det_opts;
-  det_opts.delta = opts.delta;
-  det_opts.max_mem = opts.max_mem;
-
-  // If --phone-determinize is true, do the determinization on phone + word
-  // lattices.
-  if (opts.phone_determinize) {
-    KALDI_VLOG(3) << "Doing first pass of determinization on phone + word "
-                  << "lattices.";
-    ans = DeterminizeLatticePhonePrunedFirstPass<Weight, IntType>(
-        trans_model, beam, ifst, det_opts) && ans;
-
-    // If --word-determinize is false, we've finished the job and return here.
-    if (!opts.word_determinize) {
-      // We are expecting the words on the input side.
-      ConvertLattice<Weight, IntType>(*ifst, ofst, false);
-      return ans;
-    }
-  }
-
-  // If --word-determinize is true, do the determinization on word lattices.
-  if (opts.word_determinize) {
-    KALDI_VLOG(3) << "Doing second pass of determinization on word lattices.";
-    ans = DeterminizeLatticePruned<Weight, IntType>(
-        *ifst, beam, ofst, det_opts) && ans;
-  }
-
-  // If --minimize is true, push and minimize after determinization.
-  if (opts.minimize) {
-    KALDI_VLOG(3) << "Pushing and minimizing on word lattices.";
-    ans = PushCompactLatticeStrings<Weight, IntType>(ofst) && ans;
-    ans = PushCompactLatticeWeights<Weight, IntType>(ofst) && ans;
-    ans = MinimizeCompactLattice<Weight, IntType>(ofst) && ans;
-  }
-
-  return ans;
-}
-
-// Normal verson of DeterminizeLatticePhonePruned(), where the input lattice
-// will be kept as unchanged.
-template<class Weight, class IntType>
-bool DeterminizeLatticePhonePruned(
-    const kaldi::TransitionInformation &trans_model,
-    const ExpandedFst<ArcTpl<Weight> > &ifst,
-    double beam,
-    MutableFst<ArcTpl<CompactLatticeWeightTpl<Weight, IntType> > > *ofst,
-    DeterminizeLatticePhonePrunedOptions opts) {
-  VectorFst<ArcTpl<Weight> > temp_fst(ifst);
-  return DeterminizeLatticePhonePruned(trans_model, &temp_fst,
-                                       beam, ofst, opts);
-}
-
-bool DeterminizeLatticePhonePrunedWrapper(
-    const kaldi::TransitionInformation &trans_model,
-    MutableFst<kaldi::LatticeArc> *ifst,
-    double beam,
-    MutableFst<kaldi::CompactLatticeArc> *ofst,
-    DeterminizeLatticePhonePrunedOptions opts) {
-  bool ans = true;
-  Invert(ifst);
-  if (ifst->Properties(fst::kTopSorted, true) == 0) {
-    if (!TopSort(ifst)) {
-      // Cannot topologically sort the lattice -- determinization will fail.
-      KALDI_ERR << "Topological sorting of state-level lattice failed (probably"
-                << " your lexicon has empty words or your LM has epsilon cycles"
-                << ").";
-    }
-  }
-  ILabelCompare<kaldi::LatticeArc> ilabel_comp;
-  ArcSort(ifst, ilabel_comp);
-  ans = DeterminizeLatticePhonePruned<kaldi::LatticeWeight, kaldi::int32>(
-      trans_model, ifst, beam, ofst, opts);
-  Connect(ofst);
-  return ans;
-}
+// template
+// void DeterminizeLatticeDeletePhones(
+//     ArcTpl<kaldi::LatticeWeight>::Label first_phone_label,
+//     MutableFst<ArcTpl<kaldi::LatticeWeight> > *fst);
+//
+// /** This function does a first pass determinization with phone symbols inserted
+//     at phone boundary. It uses a transition model to work out the transition-id
+//     to phone map. First, phones will be inserted into the word level lattice.
+//     Second, determinization will be applied on top of the phone + word lattice.
+//     Finally, the inserted phones will be removed, converting the lattice back to
+//     a word level lattice. The output lattice of this pass is not deterministic,
+//     since we remove the phone symbols as a last step. It is supposed to be
+//     followed by another pass of determinization at the word level. It could also
+//     be useful for some other applications such as fMLLR estimation, confidence
+//     estimation, discriminative training, etc.
+// */
+// template<class Weight, class IntType>
+// bool DeterminizeLatticePhonePrunedFirstPass(
+//     const kaldi::TransitionModel &trans_model,
+//     double beam,
+//     MutableFst<ArcTpl<Weight> > *fst,
+//     const DeterminizeLatticePrunedOptions &opts) {
+//   // First, insert the phones.
+//   typename ArcTpl<Weight>::Label first_phone_label =
+//       DeterminizeLatticeInsertPhones(trans_model, fst);
+//   TopSort(fst);
+//
+//   // Second, do determinization with phone inserted.
+//   bool ans = DeterminizeLatticePruned<Weight>(*fst, beam, fst, opts);
+//
+//   // Finally, remove the inserted phones.
+//   DeterminizeLatticeDeletePhones(first_phone_label, fst);
+//   TopSort(fst);
+//
+//   return ans;
+// }
+//
+// // "Destructive" version of DeterminizeLatticePhonePruned() where the input
+// // lattice might be modified.
+// template<class Weight, class IntType>
+// bool DeterminizeLatticePhonePruned(
+//     const kaldi::TransitionModel &trans_model,
+//     MutableFst<ArcTpl<Weight> > *ifst,
+//     double beam,
+//     MutableFst<ArcTpl<CompactLatticeWeightTpl<Weight, IntType> > > *ofst,
+//     DeterminizeLatticePhonePrunedOptions opts) {
+//   // Returning status.
+//   bool ans = true;
+//
+//   // Make sure at least one of opts.phone_determinize and opts.word_determinize
+//   // is not false, otherwise calling this function doesn't make any sense.
+//   if ((opts.phone_determinize || opts.word_determinize) == false) {
+//     KALDI_WARN << "Both --phone-determinize and --word-determinize are set to "
+//                << "false, copying lattice without determinization.";
+//     // We are expecting the words on the input side.
+//     ConvertLattice<Weight, IntType>(*ifst, ofst, false);
+//     return ans;
+//   }
+//
+//   // Determinization options.
+//   DeterminizeLatticePrunedOptions det_opts;
+//   det_opts.delta = opts.delta;
+//   det_opts.max_mem = opts.max_mem;
+//
+//   // If --phone-determinize is true, do the determinization on phone + word
+//   // lattices.
+//   if (opts.phone_determinize) {
+//     KALDI_VLOG(3) << "Doing first pass of determinization on phone + word "
+//                   << "lattices.";
+//     ans = DeterminizeLatticePhonePrunedFirstPass<Weight, IntType>(
+//         trans_model, beam, ifst, det_opts) && ans;
+//
+//     // If --word-determinize is false, we've finished the job and return here.
+//     if (!opts.word_determinize) {
+//       // We are expecting the words on the input side.
+//       ConvertLattice<Weight, IntType>(*ifst, ofst, false);
+//       return ans;
+//     }
+//   }
+//
+//   // If --word-determinize is true, do the determinization on word lattices.
+//   if (opts.word_determinize) {
+//     KALDI_VLOG(3) << "Doing second pass of determinization on word lattices.";
+//     ans = DeterminizeLatticePruned<Weight, IntType>(
+//         *ifst, beam, ofst, det_opts) && ans;
+//   }
+//
+//   // If --minimize is true, push and minimize after determinization.
+//   if (opts.minimize) {
+//     KALDI_VLOG(3) << "Pushing and minimizing on word lattices.";
+//     ans = PushCompactLatticeStrings<Weight, IntType>(ofst) && ans;
+//     ans = PushCompactLatticeWeights<Weight, IntType>(ofst) && ans;
+//     ans = MinimizeCompactLattice<Weight, IntType>(ofst) && ans;
+//   }
+//
+//   return ans;
+// }
+//
+// // Normal verson of DeterminizeLatticePhonePruned(), where the input lattice
+// // will be kept as unchanged.
+// template<class Weight, class IntType>
+// bool DeterminizeLatticePhonePruned(
+//     const kaldi::TransitionModel &trans_model,
+//     const ExpandedFst<ArcTpl<Weight> > &ifst,
+//     double beam,
+//     MutableFst<ArcTpl<CompactLatticeWeightTpl<Weight, IntType> > > *ofst,
+//     DeterminizeLatticePhonePrunedOptions opts) {
+//   VectorFst<ArcTpl<Weight> > temp_fst(ifst);
+//   return DeterminizeLatticePhonePruned(trans_model, &temp_fst,
+//                                        beam, ofst, opts);
+// }
+//
+// bool DeterminizeLatticePhonePrunedWrapper(
+//     const kaldi::TransitionModel &trans_model,
+//     MutableFst<kaldi::LatticeArc> *ifst,
+//     double beam,
+//     MutableFst<kaldi::CompactLatticeArc> *ofst,
+//     DeterminizeLatticePhonePrunedOptions opts) {
+//   bool ans = true;
+//   Invert(ifst);
+//   if (ifst->Properties(fst::kTopSorted, true) == 0) {
+//     if (!TopSort(ifst)) {
+//       // Cannot topologically sort the lattice -- determinization will fail.
+//       KALDI_ERR << "Topological sorting of state-level lattice failed (probably"
+//                 << " your lexicon has empty words or your LM has epsilon cycles"
+//                 << ").";
+//     }
+//   }
+//   ILabelCompare<kaldi::LatticeArc> ilabel_comp;
+//   ArcSort(ifst, ilabel_comp);
+//   ans = DeterminizeLatticePhonePruned<kaldi::LatticeWeight, kaldi::int32>(
+//       trans_model, ifst, beam, ofst, opts);
+//   Connect(ofst);
+//   return ans;
+// }
 
 // Instantiate the templates for the types we might need.
 // Note: there are actually four templates, each of which
@@ -1522,20 +1526,20 @@ bool DeterminizeLatticePruned<kaldi::LatticeWeight>(
     MutableFst<kaldi::LatticeArc> *ofst,
     DeterminizeLatticePrunedOptions opts);
 
-template
-bool DeterminizeLatticePhonePruned<kaldi::LatticeWeight, kaldi::int32>(
-    const kaldi::TransitionInformation &trans_model,
-    const ExpandedFst<kaldi::LatticeArc> &ifst,
-    double prune,
-    MutableFst<kaldi::CompactLatticeArc> *ofst,
-    DeterminizeLatticePhonePrunedOptions opts);
-
-template
-bool DeterminizeLatticePhonePruned<kaldi::LatticeWeight, kaldi::int32>(
-    const kaldi::TransitionInformation &trans_model,
-    MutableFst<kaldi::LatticeArc> *ifst,
-    double prune,
-    MutableFst<kaldi::CompactLatticeArc> *ofst,
-    DeterminizeLatticePhonePrunedOptions opts);
+// template
+// bool DeterminizeLatticePhonePruned<kaldi::LatticeWeight, kaldi::int32>(
+//     const kaldi::TransitionModel &trans_model,
+//     const ExpandedFst<kaldi::LatticeArc> &ifst,
+//     double prune,
+//     MutableFst<kaldi::CompactLatticeArc> *ofst,
+//     DeterminizeLatticePhonePrunedOptions opts);
+//
+// template
+// bool DeterminizeLatticePhonePruned<kaldi::LatticeWeight, kaldi::int32>(
+//     const kaldi::TransitionModel &trans_model,
+//     MutableFst<kaldi::LatticeArc> *ifst,
+//     double prune,
+//     MutableFst<kaldi::CompactLatticeArc> *ofst,
+//     DeterminizeLatticePhonePrunedOptions opts);
 
 }
diff --git a/speechx/speechx/kaldi/lat/determinize-lattice-pruned.h b/speechx/speechx/kaldi/lat/determinize-lattice-pruned.h
index c268f129..24932113 100644
--- a/speechx/speechx/kaldi/lat/determinize-lattice-pruned.h
+++ b/speechx/speechx/kaldi/lat/determinize-lattice-pruned.h
@@ -28,8 +28,8 @@
 #include <set>
 #include <vector>
 #include "fstext/lattice-weight.h"
-#include "itf/transition-information.h"
-#include "itf/options-itf.h"
+// #include "hmm/transition-model.h"
+#include "util/options-itf.h"
 #include "lat/kaldi-lattice.h"
 
 namespace fst {
@@ -212,82 +212,82 @@ bool DeterminizeLatticePruned(
     MutableFst<ArcTpl<CompactLatticeWeightTpl<Weight, IntType> > > *ofst,
     DeterminizeLatticePrunedOptions opts = DeterminizeLatticePrunedOptions());
 
-/** This function takes in lattices and inserts phones at phone boundaries. It
-    uses the transition model to work out the transition_id to phone map. The
-    returning value is the starting index of the phone label. Typically we pick
-    (maximum_output_label_index + 1) as this value. The inserted phones are then
-    mapped to (returning_value + original_phone_label) in the new lattice. The
-    returning value will be used by DeterminizeLatticeDeletePhones() where it
-    works out the phones according to this value.
-*/
-template<class Weight>
-typename ArcTpl<Weight>::Label DeterminizeLatticeInsertPhones(
-    const kaldi::TransitionInformation &trans_model,
-    MutableFst<ArcTpl<Weight> > *fst);
-
-/** This function takes in lattices and deletes "phones" from them. The "phones"
-    here are actually any label that is larger than first_phone_label because
-    when we insert phones into the lattice, we map the original phone label to
-    (first_phone_label + original_phone_label). It is supposed to be used
-    together with DeterminizeLatticeInsertPhones()
-*/
-template<class Weight>
-void DeterminizeLatticeDeletePhones(
-    typename ArcTpl<Weight>::Label first_phone_label,
-    MutableFst<ArcTpl<Weight> > *fst);
-
-/** This function is a wrapper of DeterminizeLatticePhonePrunedFirstPass() and
-    DeterminizeLatticePruned(). If --phone-determinize is set to true, it first
-    calls DeterminizeLatticePhonePrunedFirstPass() to do the initial pass of
-    determinization on the phone + word lattices. If --word-determinize is set
-    true, it then does a second pass of determinization on the word lattices by
-    calling DeterminizeLatticePruned(). If both are set to false, then it gives
-    a warning and copying the lattices without determinization.
-
-    Note: the point of doing first a phone-level determinization pass and then
-    a word-level determinization pass is that it allows us to determinize
-    deeper lattices without "failing early" and returning a too-small lattice
-    due to the max-mem constraint.  The result should be the same as word-level
-    determinization in general, but for deeper lattices it is a bit faster,
-    despite the fact that we now have two passes of determinization by default.
-*/
-template<class Weight, class IntType>
-bool DeterminizeLatticePhonePruned(
-    const kaldi::TransitionInformation &trans_model,
-    const ExpandedFst<ArcTpl<Weight> > &ifst,
-    double prune,
-    MutableFst<ArcTpl<CompactLatticeWeightTpl<Weight, IntType> > > *ofst,
-    DeterminizeLatticePhonePrunedOptions opts
-      = DeterminizeLatticePhonePrunedOptions());
-
-/** "Destructive" version of DeterminizeLatticePhonePruned() where the input
-    lattice might be changed.
-*/
-template<class Weight, class IntType>
-bool DeterminizeLatticePhonePruned(
-    const kaldi::TransitionInformation &trans_model,
-    MutableFst<ArcTpl<Weight> > *ifst,
-    double prune,
-    MutableFst<ArcTpl<CompactLatticeWeightTpl<Weight, IntType> > > *ofst,
-    DeterminizeLatticePhonePrunedOptions opts
-      = DeterminizeLatticePhonePrunedOptions());
-
-/** This function is a wrapper of DeterminizeLatticePhonePruned() that works for
-    Lattice type FSTs.  It simplifies the calling process by calling
-    TopSort() Invert() and ArcSort() for you.
-    Unlike other determinization routines, the function
-    requires "ifst" to have transition-id's on the input side and words on the
-    output side.
-    This function can be used as the top-level interface to all the determinization
-    code.
-*/
-bool DeterminizeLatticePhonePrunedWrapper(
-    const kaldi::TransitionInformation &trans_model,
-    MutableFst<kaldi::LatticeArc> *ifst,
-    double prune,
-    MutableFst<kaldi::CompactLatticeArc> *ofst,
-    DeterminizeLatticePhonePrunedOptions opts
-      = DeterminizeLatticePhonePrunedOptions());
+// /** This function takes in lattices and inserts phones at phone boundaries. It
+//     uses the transition model to work out the transition_id to phone map. The
+//     returning value is the starting index of the phone label. Typically we pick
+//     (maximum_output_label_index + 1) as this value. The inserted phones are then
+//     mapped to (returning_value + original_phone_label) in the new lattice. The
+//     returning value will be used by DeterminizeLatticeDeletePhones() where it
+//     works out the phones according to this value.
+// */
+// template<class Weight>
+// typename ArcTpl<Weight>::Label DeterminizeLatticeInsertPhones(
+//     const kaldi::TransitionModel &trans_model,
+//     MutableFst<ArcTpl<Weight> > *fst);
+//
+// /** This function takes in lattices and deletes "phones" from them. The "phones"
+//     here are actually any label that is larger than first_phone_label because
+//     when we insert phones into the lattice, we map the original phone label to
+//     (first_phone_label + original_phone_label). It is supposed to be used
+//     together with DeterminizeLatticeInsertPhones()
+// */
+// template<class Weight>
+// void DeterminizeLatticeDeletePhones(
+//     typename ArcTpl<Weight>::Label first_phone_label,
+//     MutableFst<ArcTpl<Weight> > *fst);
+//
+// /** This function is a wrapper of DeterminizeLatticePhonePrunedFirstPass() and
+//     DeterminizeLatticePruned(). If --phone-determinize is set to true, it first
+//     calls DeterminizeLatticePhonePrunedFirstPass() to do the initial pass of
+//     determinization on the phone + word lattices. If --word-determinize is set
+//     true, it then does a second pass of determinization on the word lattices by
+//     calling DeterminizeLatticePruned(). If both are set to false, then it gives
+//     a warning and copying the lattices without determinization.
+//
+//     Note: the point of doing first a phone-level determinization pass and then
+//     a word-level determinization pass is that it allows us to determinize
+//     deeper lattices without "failing early" and returning a too-small lattice
+//     due to the max-mem constraint.  The result should be the same as word-level
+//     determinization in general, but for deeper lattices it is a bit faster,
+//     despite the fact that we now have two passes of determinization by default.
+// */
+// template<class Weight, class IntType>
+// bool DeterminizeLatticePhonePruned(
+//     const kaldi::TransitionModel &trans_model,
+//     const ExpandedFst<ArcTpl<Weight> > &ifst,
+//     double prune,
+//     MutableFst<ArcTpl<CompactLatticeWeightTpl<Weight, IntType> > > *ofst,
+//     DeterminizeLatticePhonePrunedOptions opts
+//       = DeterminizeLatticePhonePrunedOptions());
+//
+// /** "Destructive" version of DeterminizeLatticePhonePruned() where the input
+//     lattice might be changed.
+// */
+// template<class Weight, class IntType>
+// bool DeterminizeLatticePhonePruned(
+//     const kaldi::TransitionModel &trans_model,
+//     MutableFst<ArcTpl<Weight> > *ifst,
+//     double prune,
+//     MutableFst<ArcTpl<CompactLatticeWeightTpl<Weight, IntType> > > *ofst,
+//     DeterminizeLatticePhonePrunedOptions opts
+//       = DeterminizeLatticePhonePrunedOptions());
+//
+// /** This function is a wrapper of DeterminizeLatticePhonePruned() that works for
+//     Lattice type FSTs.  It simplifies the calling process by calling
+//     TopSort() Invert() and ArcSort() for you.
+//     Unlike other determinization routines, the function
+//     requires "ifst" to have transition-id's on the input side and words on the
+//     output side.
+//     This function can be used as the top-level interface to all the determinization
+//     code.
+// */
+// bool DeterminizeLatticePhonePrunedWrapper(
+//     const kaldi::TransitionModel &trans_model,
+//     MutableFst<kaldi::LatticeArc> *ifst,
+//     double prune,
+//     MutableFst<kaldi::CompactLatticeArc> *ofst,
+//     DeterminizeLatticePhonePrunedOptions opts
+//       = DeterminizeLatticePhonePrunedOptions());
 
 /// @} end "addtogroup fst_extensions"
 
diff --git a/speechx/speechx/kaldi/lat/kaldi-lattice.h b/speechx/speechx/kaldi/lat/kaldi-lattice.h
index dc9f3a71..9158bffe 100644
--- a/speechx/speechx/kaldi/lat/kaldi-lattice.h
+++ b/speechx/speechx/kaldi/lat/kaldi-lattice.h
@@ -23,7 +23,7 @@
 
 #include "fstext/fstext-lib.h"
 #include "base/kaldi-common.h"
-#include "util/common-utils.h"
+// #include "util/common-utils.h"
 
 
 namespace kaldi {
@@ -142,13 +142,13 @@ class LatticeHolder {
   T *t_;
 };
 
-typedef TableWriter<LatticeHolder> LatticeWriter;
-typedef SequentialTableReader<LatticeHolder> SequentialLatticeReader;
-typedef RandomAccessTableReader<LatticeHolder> RandomAccessLatticeReader;
-
-typedef TableWriter<CompactLatticeHolder> CompactLatticeWriter;
-typedef SequentialTableReader<CompactLatticeHolder> SequentialCompactLatticeReader;
-typedef RandomAccessTableReader<CompactLatticeHolder> RandomAccessCompactLatticeReader;
+// typedef TableWriter<LatticeHolder> LatticeWriter;
+// typedef SequentialTableReader<LatticeHolder> SequentialLatticeReader;
+// typedef RandomAccessTableReader<LatticeHolder> RandomAccessLatticeReader;
+//
+// typedef TableWriter<CompactLatticeHolder> CompactLatticeWriter;
+// typedef SequentialTableReader<CompactLatticeHolder> SequentialCompactLatticeReader;
+// typedef RandomAccessTableReader<CompactLatticeHolder> RandomAccessCompactLatticeReader;
 
 
 } // namespace kaldi
diff --git a/speechx/speechx/kaldi/lat/lattice-functions.cc b/speechx/speechx/kaldi/lat/lattice-functions.cc
index 354cdf60..e483151c 100644
--- a/speechx/speechx/kaldi/lat/lattice-functions.cc
+++ b/speechx/speechx/kaldi/lat/lattice-functions.cc
@@ -23,211 +23,214 @@
 // limitations under the License.
 
 
-#include "base/kaldi-math.h"
 #include "lat/lattice-functions.h"
+// #include "hmm/transition-model.h"
+// #include "util/stl-utils.h"
+#include "base/kaldi-math.h"
+// #include "hmm/hmm-utils.h"
 
 namespace kaldi {
 using std::map;
 using std::vector;
 
-void GetPerFrameAcousticCosts(const Lattice &nbest,
-                              Vector<BaseFloat> *per_frame_loglikes) {
-  using namespace fst;
-  typedef Lattice::Arc::Weight Weight;
-  vector<BaseFloat> loglikes;
-
-  int32 cur_state = nbest.Start();
-  int32 prev_frame = -1;
-  BaseFloat eps_acwt = 0.0;
-  while(1) {
-    Weight w = nbest.Final(cur_state);
-    if (w != Weight::Zero()) {
-      KALDI_ASSERT(nbest.NumArcs(cur_state) == 0);
-      if (per_frame_loglikes != NULL)  {
-        SubVector<BaseFloat> subvec(&(loglikes[0]), loglikes.size());
-        Vector<BaseFloat> vec(subvec);
-        *per_frame_loglikes = vec;
-      }
-      break;
-    } else {
-      KALDI_ASSERT(nbest.NumArcs(cur_state) == 1);
-      fst::ArcIterator<Lattice> iter(nbest, cur_state);
-      const Lattice::Arc &arc = iter.Value();
-      BaseFloat acwt = arc.weight.Value2();
-      if (arc.ilabel != 0) {
-        if (eps_acwt > 0) {
-          acwt += eps_acwt;
-          eps_acwt = 0.0;
-        }
-        loglikes.push_back(acwt);
-        prev_frame++;
-      } else if (acwt == acwt){
-        if (prev_frame > -1) {
-          loglikes[prev_frame] += acwt;
-        } else {
-          eps_acwt += acwt;
-        }
-      }
-      cur_state = arc.nextstate;
-    }
-  }
-}
-
-int32 LatticeStateTimes(const Lattice &lat, vector<int32> *times) {
-  if (!lat.Properties(fst::kTopSorted, true))
-    KALDI_ERR << "Input lattice must be topologically sorted.";
-  KALDI_ASSERT(lat.Start() == 0);
-  int32 num_states = lat.NumStates();
-  times->clear();
-  times->resize(num_states, -1);
-  (*times)[0] = 0;
-  for (int32 state = 0; state < num_states; state++) {
-    int32 cur_time = (*times)[state];
-    for (fst::ArcIterator<Lattice> aiter(lat, state); !aiter.Done();
-        aiter.Next()) {
-      const LatticeArc &arc = aiter.Value();
-
-      if (arc.ilabel != 0) {  // Non-epsilon input label on arc
-        // next time instance
-        if ((*times)[arc.nextstate] == -1) {
-          (*times)[arc.nextstate] = cur_time + 1;
-        } else {
-          KALDI_ASSERT((*times)[arc.nextstate] == cur_time + 1);
-        }
-      } else {  // epsilon input label on arc
-        // Same time instance
-        if ((*times)[arc.nextstate] == -1)
-          (*times)[arc.nextstate] = cur_time;
-        else
-          KALDI_ASSERT((*times)[arc.nextstate] == cur_time);
-      }
-    }
-  }
-  return (*std::max_element(times->begin(), times->end()));
-}
-
-int32 CompactLatticeStateTimes(const CompactLattice &lat,
-                               vector<int32> *times) {
-  if (!lat.Properties(fst::kTopSorted, true))
-    KALDI_ERR << "Input lattice must be topologically sorted.";
-  KALDI_ASSERT(lat.Start() == 0);
-  int32 num_states = lat.NumStates();
-  times->clear();
-  times->resize(num_states, -1);
-  (*times)[0] = 0;
-  int32 utt_len = -1;
-  for (int32 state = 0; state < num_states; state++) {
-    int32 cur_time = (*times)[state];
-    for (fst::ArcIterator<CompactLattice> aiter(lat, state); !aiter.Done();
-        aiter.Next()) {
-      const CompactLatticeArc &arc = aiter.Value();
-      int32 arc_len = static_cast<int32>(arc.weight.String().size());
-      if ((*times)[arc.nextstate] == -1)
-        (*times)[arc.nextstate] = cur_time + arc_len;
-      else
-        KALDI_ASSERT((*times)[arc.nextstate] == cur_time + arc_len);
-    }
-    if (lat.Final(state) != CompactLatticeWeight::Zero()) {
-      int32 this_utt_len = (*times)[state] + lat.Final(state).String().size();
-      if (utt_len == -1) utt_len = this_utt_len;
-      else {
-        if (this_utt_len != utt_len) {
-          KALDI_WARN << "Utterance does not "
-              "seem to have a consistent length.";
-          utt_len = std::max(utt_len, this_utt_len);
-        }
-      }
-    }
-  }
-  if (utt_len == -1) {
-    KALDI_WARN << "Utterance does not have a final-state.";
-    return 0;
-  }
-  return utt_len;
-}
-
-bool ComputeCompactLatticeAlphas(const CompactLattice &clat,
-                                 vector<double> *alpha) {
-  using namespace fst;
-
-  // typedef the arc, weight types
-  typedef CompactLattice::Arc Arc;
-  typedef Arc::Weight Weight;
-  typedef Arc::StateId StateId;
-
-  //Make sure the lattice is topologically sorted.
-  if (clat.Properties(fst::kTopSorted, true) == 0) {
-    KALDI_WARN << "Input lattice must be topologically sorted.";
-    return false;
-  }
-  if (clat.Start() != 0) {
-    KALDI_WARN << "Input lattice must start from state 0.";
-    return false;
-  }
-
-  int32 num_states = clat.NumStates();
-  (*alpha).resize(0);
-  (*alpha).resize(num_states, kLogZeroDouble);
-
-  // Now propagate alphas forward. Note that we don't acount the weight of the
-  // final state to alpha[final_state] -- we acount it to beta[final_state];
-  (*alpha)[0] = 0.0;
-  for (StateId s = 0; s < num_states; s++) {
-    double this_alpha = (*alpha)[s];
-    for (ArcIterator<CompactLattice> aiter(clat, s);
-         !aiter.Done(); aiter.Next()) {
-      const Arc &arc = aiter.Value();
-      double arc_like = -(arc.weight.Weight().Value1() +
-                          arc.weight.Weight().Value2());
-      (*alpha)[arc.nextstate] = LogAdd((*alpha)[arc.nextstate],
-                                       this_alpha + arc_like);
-    }
-  }
-
-  return true;
-}
-
-bool ComputeCompactLatticeBetas(const CompactLattice &clat,
-                                vector<double> *beta) {
-  using namespace fst;
-
-  // typedef the arc, weight types
-  typedef CompactLattice::Arc Arc;
-  typedef Arc::Weight Weight;
-  typedef Arc::StateId StateId;
-
-  // Make sure the lattice is topologically sorted.
-  if (clat.Properties(fst::kTopSorted, true) == 0) {
-    KALDI_WARN << "Input lattice must be topologically sorted.";
-    return false;
-  }
-  if (clat.Start() != 0) {
-    KALDI_WARN << "Input lattice must start from state 0.";
-    return false;
-  }
-
-  int32 num_states = clat.NumStates();
-  (*beta).resize(0);
-  (*beta).resize(num_states, kLogZeroDouble);
-
-  // Now propagate betas backward. Note that beta[final_state] contains the
-  // weight of the final state in the lattice -- compare that with alpha.
-  for (StateId s = num_states-1; s >= 0; s--) {
-    Weight f = clat.Final(s);
-    double this_beta = -(f.Weight().Value1()+f.Weight().Value2());
-    for (ArcIterator<CompactLattice> aiter(clat, s);
-         !aiter.Done(); aiter.Next()) {
-      const Arc &arc = aiter.Value();
-      double arc_like = -(arc.weight.Weight().Value1() +
-                          arc.weight.Weight().Value2());
-      double arc_beta = (*beta)[arc.nextstate] + arc_like;
-      this_beta = LogAdd(this_beta, arc_beta);
-    }
-    (*beta)[s] = this_beta;
-  }
-
-  return true;
-}
+// void GetPerFrameAcousticCosts(const Lattice &nbest,
+//                               Vector<BaseFloat> *per_frame_loglikes) {
+//   using namespace fst;
+//   typedef Lattice::Arc::Weight Weight;
+//   vector<BaseFloat> loglikes;
+//
+//   int32 cur_state = nbest.Start();
+//   int32 prev_frame = -1;
+//   BaseFloat eps_acwt = 0.0;
+//   while(1) {
+//     Weight w = nbest.Final(cur_state);
+//     if (w != Weight::Zero()) {
+//       KALDI_ASSERT(nbest.NumArcs(cur_state) == 0);
+//       if (per_frame_loglikes != NULL)  {
+//         SubVector<BaseFloat> subvec(&(loglikes[0]), loglikes.size());
+//         Vector<BaseFloat> vec(subvec);
+//         *per_frame_loglikes = vec;
+//       }
+//       break;
+//     } else {
+//       KALDI_ASSERT(nbest.NumArcs(cur_state) == 1);
+//       fst::ArcIterator<Lattice> iter(nbest, cur_state);
+//       const Lattice::Arc &arc = iter.Value();
+//       BaseFloat acwt = arc.weight.Value2();
+//       if (arc.ilabel != 0) {
+//         if (eps_acwt > 0) {
+//           acwt += eps_acwt;
+//           eps_acwt = 0.0;
+//         }
+//         loglikes.push_back(acwt);
+//         prev_frame++;
+//       } else if (acwt == acwt){
+//         if (prev_frame > -1) {
+//           loglikes[prev_frame] += acwt;
+//         } else {
+//           eps_acwt += acwt;
+//         }
+//       }
+//       cur_state = arc.nextstate;
+//     }
+//   }
+// }
+//
+// int32 LatticeStateTimes(const Lattice &lat, vector<int32> *times) {
+//   if (!lat.Properties(fst::kTopSorted, true))
+//     KALDI_ERR << "Input lattice must be topologically sorted.";
+//   KALDI_ASSERT(lat.Start() == 0);
+//   int32 num_states = lat.NumStates();
+//   times->clear();
+//   times->resize(num_states, -1);
+//   (*times)[0] = 0;
+//   for (int32 state = 0; state < num_states; state++) {
+//     int32 cur_time = (*times)[state];
+//     for (fst::ArcIterator<Lattice> aiter(lat, state); !aiter.Done();
+//         aiter.Next()) {
+//       const LatticeArc &arc = aiter.Value();
+//
+//       if (arc.ilabel != 0) {  // Non-epsilon input label on arc
+//         // next time instance
+//         if ((*times)[arc.nextstate] == -1) {
+//           (*times)[arc.nextstate] = cur_time + 1;
+//         } else {
+//           KALDI_ASSERT((*times)[arc.nextstate] == cur_time + 1);
+//         }
+//       } else {  // epsilon input label on arc
+//         // Same time instance
+//         if ((*times)[arc.nextstate] == -1)
+//           (*times)[arc.nextstate] = cur_time;
+//         else
+//           KALDI_ASSERT((*times)[arc.nextstate] == cur_time);
+//       }
+//     }
+//   }
+//   return (*std::max_element(times->begin(), times->end()));
+// }
+//
+// int32 CompactLatticeStateTimes(const CompactLattice &lat,
+//                                vector<int32> *times) {
+//   if (!lat.Properties(fst::kTopSorted, true))
+//     KALDI_ERR << "Input lattice must be topologically sorted.";
+//   KALDI_ASSERT(lat.Start() == 0);
+//   int32 num_states = lat.NumStates();
+//   times->clear();
+//   times->resize(num_states, -1);
+//   (*times)[0] = 0;
+//   int32 utt_len = -1;
+//   for (int32 state = 0; state < num_states; state++) {
+//     int32 cur_time = (*times)[state];
+//     for (fst::ArcIterator<CompactLattice> aiter(lat, state); !aiter.Done();
+//         aiter.Next()) {
+//       const CompactLatticeArc &arc = aiter.Value();
+//       int32 arc_len = static_cast<int32>(arc.weight.String().size());
+//       if ((*times)[arc.nextstate] == -1)
+//         (*times)[arc.nextstate] = cur_time + arc_len;
+//       else
+//         KALDI_ASSERT((*times)[arc.nextstate] == cur_time + arc_len);
+//     }
+//     if (lat.Final(state) != CompactLatticeWeight::Zero()) {
+//       int32 this_utt_len = (*times)[state] + lat.Final(state).String().size();
+//       if (utt_len == -1) utt_len = this_utt_len;
+//       else {
+//         if (this_utt_len != utt_len) {
+//           KALDI_WARN << "Utterance does not "
+//               "seem to have a consistent length.";
+//           utt_len = std::max(utt_len, this_utt_len);
+//         }
+//       }
+//     }
+//   }
+//   if (utt_len == -1) {
+//     KALDI_WARN << "Utterance does not have a final-state.";
+//     return 0;
+//   }
+//   return utt_len;
+// }
+//
+// bool ComputeCompactLatticeAlphas(const CompactLattice &clat,
+//                                  vector<double> *alpha) {
+//   using namespace fst;
+//
+//   // typedef the arc, weight types
+//   typedef CompactLattice::Arc Arc;
+//   typedef Arc::Weight Weight;
+//   typedef Arc::StateId StateId;
+//
+//   //Make sure the lattice is topologically sorted.
+//   if (clat.Properties(fst::kTopSorted, true) == 0) {
+//     KALDI_WARN << "Input lattice must be topologically sorted.";
+//     return false;
+//   }
+//   if (clat.Start() != 0) {
+//     KALDI_WARN << "Input lattice must start from state 0.";
+//     return false;
+//   }
+//
+//   int32 num_states = clat.NumStates();
+//   (*alpha).resize(0);
+//   (*alpha).resize(num_states, kLogZeroDouble);
+//
+//   // Now propagate alphas forward. Note that we don't acount the weight of the
+//   // final state to alpha[final_state] -- we acount it to beta[final_state];
+//   (*alpha)[0] = 0.0;
+//   for (StateId s = 0; s < num_states; s++) {
+//     double this_alpha = (*alpha)[s];
+//     for (ArcIterator<CompactLattice> aiter(clat, s);
+//          !aiter.Done(); aiter.Next()) {
+//       const Arc &arc = aiter.Value();
+//       double arc_like = -(arc.weight.Weight().Value1() +
+//                           arc.weight.Weight().Value2());
+//       (*alpha)[arc.nextstate] = LogAdd((*alpha)[arc.nextstate],
+//                                        this_alpha + arc_like);
+//     }
+//   }
+//
+//   return true;
+// }
+//
+// bool ComputeCompactLatticeBetas(const CompactLattice &clat,
+//                                 vector<double> *beta) {
+//   using namespace fst;
+//
+//   // typedef the arc, weight types
+//   typedef CompactLattice::Arc Arc;
+//   typedef Arc::Weight Weight;
+//   typedef Arc::StateId StateId;
+//
+//   // Make sure the lattice is topologically sorted.
+//   if (clat.Properties(fst::kTopSorted, true) == 0) {
+//     KALDI_WARN << "Input lattice must be topologically sorted.";
+//     return false;
+//   }
+//   if (clat.Start() != 0) {
+//     KALDI_WARN << "Input lattice must start from state 0.";
+//     return false;
+//   }
+//
+//   int32 num_states = clat.NumStates();
+//   (*beta).resize(0);
+//   (*beta).resize(num_states, kLogZeroDouble);
+//
+//   // Now propagate betas backward. Note that beta[final_state] contains the
+//   // weight of the final state in the lattice -- compare that with alpha.
+//   for (StateId s = num_states-1; s >= 0; s--) {
+//     Weight f = clat.Final(s);
+//     double this_beta = -(f.Weight().Value1()+f.Weight().Value2());
+//     for (ArcIterator<CompactLattice> aiter(clat, s);
+//          !aiter.Done(); aiter.Next()) {
+//       const Arc &arc = aiter.Value();
+//       double arc_like = -(arc.weight.Weight().Value1() +
+//                           arc.weight.Weight().Value2());
+//       double arc_beta = (*beta)[arc.nextstate] + arc_like;
+//       this_beta = LogAdd(this_beta, arc_beta);
+//     }
+//     (*beta)[s] = this_beta;
+//   }
+//
+//   return true;
+// }
 
 template<class LatType>  // could be Lattice or CompactLattice
 bool PruneLattice(BaseFloat beam, LatType *lat) {
@@ -315,1566 +318,1675 @@ template bool PruneLattice(BaseFloat beam, Lattice *lat);
 template bool PruneLattice(BaseFloat beam, CompactLattice *lat);
 
 
-BaseFloat LatticeForwardBackward(const Lattice &lat, Posterior *post,
-                                 double *acoustic_like_sum) {
-  // Note, Posterior is defined as follows:  Indexed [frame], then a list
-  // of (transition-id, posterior-probability) pairs.
-  // typedef std::vector<std::vector<std::pair<int32, BaseFloat> > > Posterior;
-  using namespace fst;
-  typedef Lattice::Arc Arc;
-  typedef Arc::Weight Weight;
-  typedef Arc::StateId StateId;
-
-  if (acoustic_like_sum) *acoustic_like_sum = 0.0;
-
-  // Make sure the lattice is topologically sorted.
-  if (lat.Properties(fst::kTopSorted, true) == 0)
-    KALDI_ERR << "Input lattice must be topologically sorted.";
-  KALDI_ASSERT(lat.Start() == 0);
-
-  int32 num_states = lat.NumStates();
-  vector<int32> state_times;
-  int32 max_time = LatticeStateTimes(lat, &state_times);
-  std::vector<double> alpha(num_states, kLogZeroDouble);
-  std::vector<double> &beta(alpha); // we re-use the same memory for
-  // this, but it's semantically distinct so we name it differently.
-  double tot_forward_prob = kLogZeroDouble;
-
-  post->clear();
-  post->resize(max_time);
-
-  alpha[0] = 0.0;
-  // Propagate alphas forward.
-  for (StateId s = 0; s < num_states; s++) {
-    double this_alpha = alpha[s];
-    for (ArcIterator<Lattice> aiter(lat, s); !aiter.Done(); aiter.Next()) {
-      const Arc &arc = aiter.Value();
-      double arc_like = -ConvertToCost(arc.weight);
-      alpha[arc.nextstate] = LogAdd(alpha[arc.nextstate], this_alpha + arc_like);
-    }
-    Weight f = lat.Final(s);
-    if (f != Weight::Zero()) {
-      double final_like = this_alpha - (f.Value1() + f.Value2());
-      tot_forward_prob = LogAdd(tot_forward_prob, final_like);
-      KALDI_ASSERT(state_times[s] == max_time &&
-                   "Lattice is inconsistent (final-prob not at max_time)");
-    }
-  }
-  for (StateId s = num_states-1; s >= 0; s--) {
-    Weight f = lat.Final(s);
-    double this_beta = -(f.Value1() + f.Value2());
-    for (ArcIterator<Lattice> aiter(lat, s); !aiter.Done(); aiter.Next()) {
-      const Arc &arc = aiter.Value();
-      double arc_like = -ConvertToCost(arc.weight),
-          arc_beta = beta[arc.nextstate] + arc_like;
-      this_beta = LogAdd(this_beta, arc_beta);
-      int32 transition_id = arc.ilabel;
-
-      // The following "if" is an optimization to avoid un-needed exp().
-      if (transition_id != 0 || acoustic_like_sum != NULL) {
-        double posterior = Exp(alpha[s] + arc_beta - tot_forward_prob);
-
-        if (transition_id != 0) // Arc has a transition-id on it [not epsilon]
-          (*post)[state_times[s]].push_back(std::make_pair(transition_id,
-                                                           static_cast<kaldi::BaseFloat>(posterior)));
-        if (acoustic_like_sum != NULL)
-          *acoustic_like_sum -= posterior * arc.weight.Value2();
-      }
-    }
-    if (acoustic_like_sum != NULL && f != Weight::Zero()) {
-      double final_logprob = - ConvertToCost(f),
-          posterior = Exp(alpha[s] + final_logprob - tot_forward_prob);
-      *acoustic_like_sum -= posterior * f.Value2();
-    }
-    beta[s] = this_beta;
-  }
-  double tot_backward_prob = beta[0];
-  if (!ApproxEqual(tot_forward_prob, tot_backward_prob, 1e-8)) {
-    KALDI_WARN << "Total forward probability over lattice = " << tot_forward_prob
-              << ", while total backward probability = " << tot_backward_prob;
-  }
-  // Now combine any posteriors with the same transition-id.
-  for (int32 t = 0; t < max_time; t++)
-    MergePairVectorSumming(&((*post)[t]));
-  return tot_backward_prob;
-}
-
-
-void LatticeActivePhones(const Lattice &lat, const TransitionInformation &trans,
-                         const vector<int32> &silence_phones,
-                         vector< std::set<int32> > *active_phones) {
-  KALDI_ASSERT(IsSortedAndUniq(silence_phones));
-  vector<int32> state_times;
-  int32 num_states = lat.NumStates();
-  int32 max_time = LatticeStateTimes(lat, &state_times);
-  active_phones->clear();
-  active_phones->resize(max_time);
-  for (int32 state = 0; state < num_states; state++) {
-    int32 cur_time = state_times[state];
-    for (fst::ArcIterator<Lattice> aiter(lat, state); !aiter.Done();
-        aiter.Next()) {
-      const LatticeArc &arc = aiter.Value();
-      if (arc.ilabel != 0) {  // Non-epsilon arc
-        int32 phone = trans.TransitionIdToPhone(arc.ilabel);
-        if (!std::binary_search(silence_phones.begin(),
-                                silence_phones.end(), phone))
-          (*active_phones)[cur_time].insert(phone);
-      }
-    }  // end looping over arcs
-  }  // end looping over states
-}
-
-void ConvertLatticeToPhones(const TransitionInformation &trans,
-                            Lattice *lat) {
-  typedef LatticeArc Arc;
-  int32 num_states = lat->NumStates();
-  for (int32 state = 0; state < num_states; state++) {
-    for (fst::MutableArcIterator<Lattice> aiter(lat, state); !aiter.Done();
-        aiter.Next()) {
-      Arc arc(aiter.Value());
-      arc.olabel = 0; // remove any word.
-      if ((arc.ilabel != 0) // has a transition-id on input..
-          && (trans.TransitionIdIsStartOfPhone(arc.ilabel))
-          && (!trans.IsSelfLoop(arc.ilabel))) {
-         // && trans.IsFinal(arc.ilabel)) // there is one of these per phone...
-        arc.olabel = trans.TransitionIdToPhone(arc.ilabel);
-      }
-      aiter.SetValue(arc);
-    }  // end looping over arcs
-  }  // end looping over states
-}
-
-
-static inline double LogAddOrMax(bool viterbi, double a, double b) {
-  if (viterbi)
-    return std::max(a, b);
-  else
-    return LogAdd(a, b);
-}
-
-template<typename LatticeType>
-double ComputeLatticeAlphasAndBetas(const LatticeType &lat,
-                                    bool viterbi,
-                                    vector<double> *alpha,
-                                    vector<double> *beta) {
-  typedef typename LatticeType::Arc Arc;
-  typedef typename Arc::Weight Weight;
-  typedef typename Arc::StateId StateId;
-
-  StateId num_states = lat.NumStates();
-  KALDI_ASSERT(lat.Properties(fst::kTopSorted, true) == fst::kTopSorted);
-  KALDI_ASSERT(lat.Start() == 0);
-  alpha->clear();
-  beta->clear();
-  alpha->resize(num_states, kLogZeroDouble);
-  beta->resize(num_states, kLogZeroDouble);
-
-  double tot_forward_prob = kLogZeroDouble;
-  (*alpha)[0] = 0.0;
-  // Propagate alphas forward.
-  for (StateId s = 0; s < num_states; s++) {
-    double this_alpha = (*alpha)[s];
-    for (fst::ArcIterator<LatticeType> aiter(lat, s); !aiter.Done();
-         aiter.Next()) {
-      const Arc &arc = aiter.Value();
-      double arc_like = -ConvertToCost(arc.weight);
-      (*alpha)[arc.nextstate] = LogAddOrMax(viterbi, (*alpha)[arc.nextstate],
-                                                this_alpha + arc_like);
-    }
-    Weight f = lat.Final(s);
-    if (f != Weight::Zero()) {
-      double final_like = this_alpha - ConvertToCost(f);
-      tot_forward_prob = LogAddOrMax(viterbi, tot_forward_prob, final_like);
-    }
-  }
-  for (StateId s = num_states-1; s >= 0; s--) { // it's guaranteed signed.
-    double this_beta = -ConvertToCost(lat.Final(s));
-    for (fst::ArcIterator<LatticeType> aiter(lat, s); !aiter.Done();
-         aiter.Next()) {
-      const Arc &arc = aiter.Value();
-      double arc_like = -ConvertToCost(arc.weight),
-          arc_beta = (*beta)[arc.nextstate] + arc_like;
-      this_beta = LogAddOrMax(viterbi, this_beta, arc_beta);
-    }
-    (*beta)[s] = this_beta;
-  }
-  double tot_backward_prob = (*beta)[lat.Start()];
-  if (!ApproxEqual(tot_forward_prob, tot_backward_prob, 1e-8)) {
-    KALDI_WARN << "Total forward probability over lattice = " << tot_forward_prob
-               << ", while total backward probability = " << tot_backward_prob;
-  }
-  // Split the difference when returning... they should be the same.
-  return 0.5 * (tot_backward_prob + tot_forward_prob);
-}
-
-// instantiate the template for Lattice and CompactLattice
-template
-double ComputeLatticeAlphasAndBetas(const Lattice &lat,
-                                    bool viterbi,
-                                    vector<double> *alpha,
-                                    vector<double> *beta);
-
-template
-double ComputeLatticeAlphasAndBetas(const CompactLattice &lat,
-                                    bool viterbi,
-                                    vector<double> *alpha,
-                                    vector<double> *beta);
-
-
-
-/// This is used in CompactLatticeLimitDepth.
-struct LatticeArcRecord {
-  BaseFloat logprob; // logprob <= 0 is the best Viterbi logprob of this arc,
-                     // minus the overall best-cost of the lattice.
-  CompactLatticeArc::StateId state; // state in the lattice.
-  size_t arc; // arc index within the state.
-  bool operator < (const LatticeArcRecord &other) const {
-    return logprob < other.logprob;
-  }
-};
-
-void CompactLatticeLimitDepth(int32 max_depth_per_frame,
-                              CompactLattice *clat) {
-  typedef CompactLatticeArc Arc;
-  typedef Arc::Weight Weight;
-  typedef Arc::StateId StateId;
-
-  if (clat->Start() == fst::kNoStateId) {
-    KALDI_WARN << "Limiting depth of empty lattice.";
-    return;
-  }
-  if (clat->Properties(fst::kTopSorted, true) == 0) {
-    if (!TopSort(clat))
-      KALDI_ERR << "Topological sorting of lattice failed.";
-  }
-
-  vector<int32> state_times;
-  int32 T = CompactLatticeStateTimes(*clat, &state_times);
-
-  // The alpha and beta quantities here are "viterbi" alphas and beta.
-  std::vector<double> alpha;
-  std::vector<double> beta;
-  bool viterbi = true;
-  double best_prob = ComputeLatticeAlphasAndBetas(*clat, viterbi,
-                                                  &alpha, &beta);
-
-  std::vector<std::vector<LatticeArcRecord> > arc_records(T);
-
-  StateId num_states = clat->NumStates();
-  for (StateId s = 0; s < num_states; s++) {
-    for (fst::ArcIterator<CompactLattice> aiter(*clat, s); !aiter.Done();
-         aiter.Next()) {
-      const Arc &arc = aiter.Value();
-      LatticeArcRecord arc_record;
-      arc_record.state = s;
-      arc_record.arc = aiter.Position();
-      arc_record.logprob =
-          (alpha[s] + beta[arc.nextstate] - ConvertToCost(arc.weight))
-           - best_prob;
-      KALDI_ASSERT(arc_record.logprob < 0.1); // Should be zero or negative.
-      int32 num_frames = arc.weight.String().size(), start_t = state_times[s];
-      for (int32 t = start_t; t < start_t + num_frames; t++) {
-        KALDI_ASSERT(t < T);
-        arc_records[t].push_back(arc_record);
-      }
-    }
-  }
-  StateId dead_state = clat->AddState(); // A non-coaccesible state which we use
-                                         // to remove arcs (make them end
-                                         // there).
-  size_t max_depth = max_depth_per_frame;
-  for (int32 t = 0; t < T; t++) {
-    size_t size = arc_records[t].size();
-    if (size > max_depth) {
-      // we sort from worst to best, so we keep the later-numbered ones,
-      // and delete the lower-numbered ones.
-      size_t cutoff = size - max_depth;
-      std::nth_element(arc_records[t].begin(),
-                       arc_records[t].begin() + cutoff,
-                       arc_records[t].end());
-      for (size_t index = 0; index < cutoff; index++) {
-        LatticeArcRecord record(arc_records[t][index]);
-        fst::MutableArcIterator<CompactLattice> aiter(clat, record.state);
-        aiter.Seek(record.arc);
-        Arc arc = aiter.Value();
-        if (arc.nextstate != dead_state) { // not already killed.
-          arc.nextstate = dead_state;
-          aiter.SetValue(arc);
-        }
-      }
-    }
-  }
-  Connect(clat);
-  TopSortCompactLatticeIfNeeded(clat);
-}
-
-
-void TopSortCompactLatticeIfNeeded(CompactLattice *clat) {
-  if (clat->Properties(fst::kTopSorted, true) == 0) {
-    if (fst::TopSort(clat) == false) {
-      KALDI_ERR << "Topological sorting failed";
-    }
-  }
-}
-
-void TopSortLatticeIfNeeded(Lattice *lat) {
-  if (lat->Properties(fst::kTopSorted, true) == 0) {
-    if (fst::TopSort(lat) == false) {
-      KALDI_ERR << "Topological sorting failed";
-    }
-  }
-}
-
-
-/// Returns the depth of the lattice, defined as the average number of
-/// arcs crossing any given frame.  Returns 1 for empty lattices.
-/// Requires that input is topologically sorted.
-BaseFloat CompactLatticeDepth(const CompactLattice &clat,
-                              int32 *num_frames) {
-  typedef CompactLattice::Arc::StateId StateId;
-  if (clat.Properties(fst::kTopSorted, true) == 0) {
-    KALDI_ERR << "Lattice input to CompactLatticeDepth was not topologically "
-              << "sorted.";
-  }
-  if (clat.Start() == fst::kNoStateId) {
-    *num_frames = 0;
-    return 1.0;
-  }
-  size_t num_arc_frames = 0;
-  int32 t;
-  {
-    vector<int32> state_times;
-    t = CompactLatticeStateTimes(clat, &state_times);
-  }
-  if (num_frames != NULL)
-    *num_frames = t;
-  for (StateId s = 0; s < clat.NumStates(); s++) {
-    for (fst::ArcIterator<CompactLattice> aiter(clat, s); !aiter.Done();
-         aiter.Next()) {
-      const CompactLatticeArc &arc = aiter.Value();
-      num_arc_frames += arc.weight.String().size();
-    }
-    num_arc_frames += clat.Final(s).String().size();
-  }
-  return num_arc_frames / static_cast<BaseFloat>(t);
-}
-
-
-void CompactLatticeDepthPerFrame(const CompactLattice &clat,
-                                 std::vector<int32> *depth_per_frame) {
-  typedef CompactLattice::Arc::StateId StateId;
-  if (clat.Properties(fst::kTopSorted, true) == 0) {
-    KALDI_ERR << "Lattice input to CompactLatticeDepthPerFrame was not "
-              << "topologically sorted.";
-  }
-  if (clat.Start() == fst::kNoStateId) {
-    depth_per_frame->clear();
-    return;
-  }
-  vector<int32> state_times;
-  int32 T = CompactLatticeStateTimes(clat, &state_times);
-
-  depth_per_frame->clear();
-  if (T <= 0) {
-    return;
-  } else {
-    depth_per_frame->resize(T, 0);
-    for (StateId s = 0; s < clat.NumStates(); s++) {
-      int32 start_time = state_times[s];
-      for (fst::ArcIterator<CompactLattice> aiter(clat, s); !aiter.Done();
-           aiter.Next()) {
-        const CompactLatticeArc &arc = aiter.Value();
-        int32 len = arc.weight.String().size();
-        for (int32 t = start_time; t < start_time + len; t++) {
-          KALDI_ASSERT(t < T);
-          (*depth_per_frame)[t]++;
-        }
-      }
-      int32 final_len = clat.Final(s).String().size();
-      for (int32 t = start_time; t < start_time + final_len; t++) {
-        KALDI_ASSERT(t < T);
-        (*depth_per_frame)[t]++;
-      }
-    }
-  }
-}
-
-
-
-void ConvertCompactLatticeToPhones(const TransitionInformation &trans,
-                                   CompactLattice *clat) {
-  typedef CompactLatticeArc Arc;
-  typedef Arc::Weight Weight;
-  int32 num_states = clat->NumStates();
-  for (int32 state = 0; state < num_states; state++) {
-    for (fst::MutableArcIterator<CompactLattice> aiter(clat, state);
-         !aiter.Done();
-         aiter.Next()) {
-      Arc arc(aiter.Value());
-      std::vector<int32> phone_seq;
-      const std::vector<int32> &tid_seq = arc.weight.String();
-      for (std::vector<int32>::const_iterator iter = tid_seq.begin();
-           iter != tid_seq.end(); ++iter) {
-        if (trans.IsFinal(*iter))// note: there is one of these per phone...
-          phone_seq.push_back(trans.TransitionIdToPhone(*iter));
-      }
-      arc.weight.SetString(phone_seq);
-      aiter.SetValue(arc);
-    } // end looping over arcs
-    Weight f = clat->Final(state);
-    if (f != Weight::Zero()) {
-      std::vector<int32> phone_seq;
-      const std::vector<int32> &tid_seq = f.String();
-      for (std::vector<int32>::const_iterator iter = tid_seq.begin();
-           iter != tid_seq.end(); ++iter) {
-        if (trans.IsFinal(*iter))// note: there is one of these per phone...
-          phone_seq.push_back(trans.TransitionIdToPhone(*iter));
-      }
-      f.SetString(phone_seq);
-      clat->SetFinal(state, f);
-    }
-  }  // end looping over states
-}
-
-bool LatticeBoost(const TransitionInformation &trans,
-                  const std::vector<int32> &alignment,
-                  const std::vector<int32> &silence_phones,
-                  BaseFloat b,
-                  BaseFloat max_silence_error,
-                  Lattice *lat) {
-  TopSortLatticeIfNeeded(lat);
-
-  // get all stored properties (test==false means don't test if not known).
-  uint64 props = lat->Properties(fst::kFstProperties,
-                                 false);
-
-  KALDI_ASSERT(IsSortedAndUniq(silence_phones));
-  KALDI_ASSERT(max_silence_error >= 0.0 && max_silence_error <= 1.0);
-  vector<int32> state_times;
-  int32 num_states = lat->NumStates();
-  int32 num_frames = LatticeStateTimes(*lat, &state_times);
-  KALDI_ASSERT(num_frames == static_cast<int32>(alignment.size()));
-  for (int32 state = 0; state < num_states; state++) {
-    int32 cur_time = state_times[state];
-    for (fst::MutableArcIterator<Lattice> aiter(lat, state); !aiter.Done();
-         aiter.Next()) {
-      LatticeArc arc = aiter.Value();
-      if (arc.ilabel != 0) {  // Non-epsilon arc
-        if (arc.ilabel < 0 || arc.ilabel > trans.NumTransitionIds()) {
-          KALDI_WARN << "Lattice has out-of-range transition-ids: "
-                     << "lattice/model mismatch?";
-          return false;
-        }
-        int32 phone = trans.TransitionIdToPhone(arc.ilabel),
-            ref_phone = trans.TransitionIdToPhone(alignment[cur_time]);
-        BaseFloat frame_error;
-        if (phone == ref_phone) {
-          frame_error = 0.0;
-        } else { // an error...
-          if (std::binary_search(silence_phones.begin(), silence_phones.end(), phone))
-            frame_error = max_silence_error;
-          else
-            frame_error = 1.0;
-        }
-        BaseFloat delta_cost = -b * frame_error; // negative cost if
-        // frame is wrong, to boost likelihood of arcs with errors on them.
-        // Add this cost to the graph part.
-        arc.weight.SetValue1(arc.weight.Value1() + delta_cost);
-        aiter.SetValue(arc);
-      }
-    }
-  }
-  // All we changed is the weights, so any properties that were
-  // known before, are still known, except for whether or not the
-  // lattice was weighted.
-  lat->SetProperties(props,
-                     ~(fst::kWeighted|fst::kUnweighted));
-
-  return true;
-}
-
-
-
-BaseFloat LatticeForwardBackwardMpeVariants(
-    const TransitionInformation &trans,
-    const std::vector<int32> &silence_phones,
-    const Lattice &lat,
-    const std::vector<int32> &num_ali,
-    std::string criterion,
-    bool one_silence_class,
-    Posterior *post) {
-  using namespace fst;
-  typedef Lattice::Arc Arc;
-  typedef Arc::Weight Weight;
-  typedef Arc::StateId StateId;
-
-  KALDI_ASSERT(criterion == "mpfe" || criterion == "smbr");
-  bool is_mpfe = (criterion == "mpfe");
-
-  if (lat.Properties(fst::kTopSorted, true) == 0)
-    KALDI_ERR << "Input lattice must be topologically sorted.";
-  KALDI_ASSERT(lat.Start() == 0);
-
-  int32 num_states = lat.NumStates();
-  vector<int32> state_times;
-  int32 max_time = LatticeStateTimes(lat, &state_times);
-  KALDI_ASSERT(max_time == static_cast<int32>(num_ali.size()));
-  std::vector<double> alpha(num_states, kLogZeroDouble),
-      alpha_smbr(num_states, 0), //forward variable for sMBR
-      beta(num_states, kLogZeroDouble),
-      beta_smbr(num_states, 0); //backward variable for sMBR
-
-  double tot_forward_prob = kLogZeroDouble;
-  double tot_forward_score = 0;
-
-  post->clear();
-  post->resize(max_time);
-
-  alpha[0] = 0.0;
-  // First Pass Forward,
-  for (StateId s = 0; s < num_states; s++) {
-    double this_alpha = alpha[s];
-    for (ArcIterator<Lattice> aiter(lat, s); !aiter.Done(); aiter.Next()) {
-      const Arc &arc = aiter.Value();
-      double arc_like = -ConvertToCost(arc.weight);
-      alpha[arc.nextstate] = LogAdd(alpha[arc.nextstate], this_alpha + arc_like);
-    }
-    Weight f = lat.Final(s);
-    if (f != Weight::Zero()) {
-      double final_like = this_alpha - (f.Value1() + f.Value2());
-      tot_forward_prob = LogAdd(tot_forward_prob, final_like);
-      KALDI_ASSERT(state_times[s] == max_time &&
-                   "Lattice is inconsistent (final-prob not at max_time)");
-    }
-  }
-  // First Pass Backward,
-  for (StateId s = num_states-1; s >= 0; s--) {
-    Weight f = lat.Final(s);
-    double this_beta = -(f.Value1() + f.Value2());
-    for (ArcIterator<Lattice> aiter(lat, s); !aiter.Done(); aiter.Next()) {
-      const Arc &arc = aiter.Value();
-      double arc_like = -ConvertToCost(arc.weight),
-          arc_beta = beta[arc.nextstate] + arc_like;
-      this_beta = LogAdd(this_beta, arc_beta);
-    }
-    beta[s] = this_beta;
-  }
-  // First Pass Forward-Backward Check
-  double tot_backward_prob = beta[0];
-  // may loose the condition somehow here 1e-6 (was 1e-8)
-  if (!ApproxEqual(tot_forward_prob, tot_backward_prob, 1e-6)) {
-    KALDI_ERR << "Total forward probability over lattice = " << tot_forward_prob
-              << ", while total backward probability = " << tot_backward_prob;
-  }
-
-  alpha_smbr[0] = 0.0;
-  // Second Pass Forward, calculate forward for MPFE/SMBR
-  for (StateId s = 0; s < num_states; s++) {
-    double this_alpha = alpha[s];
-    for (ArcIterator<Lattice> aiter(lat, s); !aiter.Done(); aiter.Next()) {
-      const Arc &arc = aiter.Value();
-      double arc_like = -ConvertToCost(arc.weight);
-      double frame_acc = 0.0;
-      if (arc.ilabel != 0) {
-        int32 cur_time = state_times[s];
-        int32 phone = trans.TransitionIdToPhone(arc.ilabel),
-            ref_phone = trans.TransitionIdToPhone(num_ali[cur_time]);
-        bool phone_is_sil = std::binary_search(silence_phones.begin(),
-                                               silence_phones.end(),
-                                               phone),
-            ref_phone_is_sil = std::binary_search(silence_phones.begin(),
-                                                  silence_phones.end(),
-                                                  ref_phone),
-            both_sil = phone_is_sil && ref_phone_is_sil;
-        if (!is_mpfe) { // smbr.
-          int32 pdf = trans.TransitionIdToPdf(arc.ilabel),
-              ref_pdf = trans.TransitionIdToPdf(num_ali[cur_time]);
-          if (!one_silence_class)  // old behavior
-            frame_acc = (pdf == ref_pdf && !phone_is_sil) ? 1.0 : 0.0;
-          else
-            frame_acc = (pdf == ref_pdf || both_sil) ? 1.0 : 0.0;
-        } else {
-          if (!one_silence_class)  // old behavior
-            frame_acc = (phone == ref_phone && !phone_is_sil) ? 1.0 : 0.0;
-          else
-            frame_acc = (phone == ref_phone || both_sil) ? 1.0 : 0.0;
-        }
-      }
-      double arc_scale = Exp(alpha[s] + arc_like - alpha[arc.nextstate]);
-      alpha_smbr[arc.nextstate] += arc_scale * (alpha_smbr[s] + frame_acc);
-    }
-    Weight f = lat.Final(s);
-    if (f != Weight::Zero()) {
-      double final_like = this_alpha - (f.Value1() + f.Value2());
-      double arc_scale = Exp(final_like - tot_forward_prob);
-      tot_forward_score += arc_scale * alpha_smbr[s];
-      KALDI_ASSERT(state_times[s] == max_time &&
-                   "Lattice is inconsistent (final-prob not at max_time)");
-    }
-  }
-  // Second Pass Backward, collect Mpe style posteriors
-  for (StateId s = num_states-1; s >= 0; s--) {
-    for (ArcIterator<Lattice> aiter(lat, s); !aiter.Done(); aiter.Next()) {
-      const Arc &arc = aiter.Value();
-      double arc_like = -ConvertToCost(arc.weight),
-          arc_beta = beta[arc.nextstate] + arc_like;
-      double frame_acc = 0.0;
-      int32 transition_id = arc.ilabel;
-      if (arc.ilabel != 0) {
-        int32 cur_time = state_times[s];
-        int32 phone = trans.TransitionIdToPhone(arc.ilabel),
-            ref_phone = trans.TransitionIdToPhone(num_ali[cur_time]);
-        bool phone_is_sil = std::binary_search(silence_phones.begin(),
-                                               silence_phones.end(), phone),
-            ref_phone_is_sil = std::binary_search(silence_phones.begin(),
-                                                  silence_phones.end(),
-                                                  ref_phone),
-            both_sil = phone_is_sil && ref_phone_is_sil;
-        if (!is_mpfe) { // smbr.
-          int32 pdf = trans.TransitionIdToPdf(arc.ilabel),
-              ref_pdf = trans.TransitionIdToPdf(num_ali[cur_time]);
-          if (!one_silence_class)  // old behavior
-            frame_acc = (pdf == ref_pdf && !phone_is_sil) ? 1.0 : 0.0;
-          else
-            frame_acc = (pdf == ref_pdf || both_sil) ? 1.0 : 0.0;
-        } else {
-          if (!one_silence_class)  // old behavior
-            frame_acc = (phone == ref_phone && !phone_is_sil) ? 1.0 : 0.0;
-          else
-            frame_acc = (phone == ref_phone || both_sil) ? 1.0 : 0.0;
-        }
-      }
-      double arc_scale = Exp(beta[arc.nextstate] + arc_like - beta[s]);
-      // check arc_scale NAN,
-      // this is to prevent partial paths in Lattices
-      // i.e., paths don't survive to the final state
-      if (KALDI_ISNAN(arc_scale)) arc_scale = 0;
-      beta_smbr[s] += arc_scale * (beta_smbr[arc.nextstate] + frame_acc);
-
-      if (transition_id != 0) { // Arc has a transition-id on it [not epsilon]
-        double posterior = Exp(alpha[s] + arc_beta - tot_forward_prob);
-        double acc_diff = alpha_smbr[s] + frame_acc + beta_smbr[arc.nextstate]
-                               - tot_forward_score;
-        double posterior_smbr = posterior * acc_diff;
-        (*post)[state_times[s]].push_back(std::make_pair(transition_id,
-                                                         static_cast<BaseFloat>(posterior_smbr)));
-      }
-    }
-  }
-
-  //Second Pass Forward Backward check
-  double tot_backward_score = beta_smbr[0];  // Initial state id == 0
-  // may loose the condition somehow here 1e-5/1e-4
-  if (!ApproxEqual(tot_forward_score, tot_backward_score, 1e-4)) {
-    KALDI_ERR << "Total forward score over lattice = " << tot_forward_score
-              << ", while total backward score = " << tot_backward_score;
-  }
-
-  // Output the computed posteriors
-  for (int32 t = 0; t < max_time; t++)
-    MergePairVectorSumming(&((*post)[t]));
-  return tot_forward_score;
-}
-
-bool CompactLatticeToWordAlignment(const CompactLattice &clat,
-                                   std::vector<int32> *words,
-                                   std::vector<int32> *begin_times,
-                                   std::vector<int32> *lengths) {
-  words->clear();
-  begin_times->clear();
-  lengths->clear();
-  typedef CompactLattice::Arc Arc;
-  typedef Arc::Label Label;
-  typedef CompactLattice::StateId StateId;
-  typedef CompactLattice::Weight Weight;
-  using namespace fst;
-  StateId state = clat.Start();
-  int32 cur_time = 0;
-  if (state == kNoStateId) {
-    KALDI_WARN << "Empty lattice.";
-    return false;
-  }
-  while (1) {
-    Weight final = clat.Final(state);
-    size_t num_arcs = clat.NumArcs(state);
-    if (final != Weight::Zero()) {
-      if (num_arcs != 0) {
-        KALDI_WARN << "Lattice is not linear.";
-        return false;
-      }
-      if (! final.String().empty()) {
-        KALDI_WARN << "Lattice has alignments on final-weight: probably "
-            "was not word-aligned (alignments will be approximate)";
-      }
-      return true;
-    } else {
-      if (num_arcs != 1) {
-        KALDI_WARN << "Lattice is not linear: num-arcs = " << num_arcs;
-        return false;
-      }
-      fst::ArcIterator<CompactLattice> aiter(clat, state);
-      const Arc &arc = aiter.Value();
-      Label word_id = arc.ilabel; // Note: ilabel==olabel, since acceptor.
-      // Also note: word_id may be zero; we output it anyway.
-      int32 length = arc.weight.String().size();
-      words->push_back(word_id);
-      begin_times->push_back(cur_time);
-      lengths->push_back(length);
-      cur_time += length;
-      state = arc.nextstate;
-    }
-  }
-}
-
-
-void CompactLatticeShortestPath(const CompactLattice &clat,
-                                CompactLattice *shortest_path) {
-  using namespace fst;
-  if (clat.Properties(fst::kTopSorted, true) == 0) {
-    CompactLattice clat_copy(clat);
-    if (!TopSort(&clat_copy))
-      KALDI_ERR << "Was not able to topologically sort lattice (cycles found?)";
-    CompactLatticeShortestPath(clat_copy, shortest_path);
-    return;
-  }
-  // Now we can assume it's topologically sorted.
-  shortest_path->DeleteStates();
-  if (clat.Start() == kNoStateId) return;
-  typedef CompactLatticeArc Arc;
-  typedef Arc::StateId StateId;
-  typedef CompactLatticeWeight Weight;
-  vector<std::pair<double, StateId> > best_cost_and_pred(clat.NumStates() + 1);
-  StateId superfinal = clat.NumStates();
-  for (StateId s = 0; s <= clat.NumStates(); s++) {
-    best_cost_and_pred[s].first = std::numeric_limits<double>::infinity();
-    best_cost_and_pred[s].second = fst::kNoStateId;
-  }
-  best_cost_and_pred[clat.Start()].first = 0;
-  for (StateId s = 0; s < clat.NumStates(); s++) {
-    double my_cost = best_cost_and_pred[s].first;
-    for (ArcIterator<CompactLattice> aiter(clat, s);
-         !aiter.Done();
-         aiter.Next()) {
-      const Arc &arc = aiter.Value();
-      double arc_cost = ConvertToCost(arc.weight),
-          next_cost = my_cost + arc_cost;
-      if (next_cost < best_cost_and_pred[arc.nextstate].first) {
-        best_cost_and_pred[arc.nextstate].first = next_cost;
-        best_cost_and_pred[arc.nextstate].second = s;
-      }
-    }
-    double final_cost = ConvertToCost(clat.Final(s)),
-        tot_final = my_cost + final_cost;
-    if (tot_final < best_cost_and_pred[superfinal].first) {
-      best_cost_and_pred[superfinal].first = tot_final;
-      best_cost_and_pred[superfinal].second = s;
-    }
-  }
-  std::vector<StateId> states; // states on best path.
-  StateId cur_state = superfinal, start_state = clat.Start();
-  while (cur_state != start_state) {
-    StateId prev_state = best_cost_and_pred[cur_state].second;
-    if (prev_state == kNoStateId) {
-      KALDI_WARN << "Failure in best-path algorithm for lattice (infinite costs?)";
-      return; // return empty best-path.
-    }
-    states.push_back(prev_state);
-    KALDI_ASSERT(cur_state != prev_state && "Lattice with cycles");
-    cur_state = prev_state;
-  }
-  std::reverse(states.begin(), states.end());
-  for (size_t i = 0; i < states.size(); i++)
-    shortest_path->AddState();
-  for (StateId s = 0; static_cast<size_t>(s) < states.size(); s++) {
-    if (s == 0) shortest_path->SetStart(s);
-    if (static_cast<size_t>(s + 1) < states.size()) { // transition to next state.
-      bool have_arc = false;
-      Arc cur_arc;
-      for (ArcIterator<CompactLattice> aiter(clat, states[s]);
-           !aiter.Done();
-           aiter.Next()) {
-        const Arc &arc = aiter.Value();
-        if (arc.nextstate == states[s+1]) {
-          if (!have_arc ||
-              ConvertToCost(arc.weight) < ConvertToCost(cur_arc.weight)) {
-            cur_arc = arc;
-            have_arc = true;
-          }
-        }
-      }
-      KALDI_ASSERT(have_arc && "Code error.");
-      shortest_path->AddArc(s, Arc(cur_arc.ilabel, cur_arc.olabel,
-                                   cur_arc.weight, s+1));
-    } else { // final-prob.
-      shortest_path->SetFinal(s, clat.Final(states[s]));
-    }
-  }
-}
-
-
-void ExpandCompactLattice(const CompactLattice &clat,
-                          double epsilon,
-                          CompactLattice *expand_clat) {
-  using namespace fst;
-  typedef CompactLattice::Arc Arc;
-  typedef Arc::Weight Weight;
-  typedef Arc::StateId StateId;
-  typedef std::pair<StateId, StateId> StatePair;
-  typedef unordered_map<StatePair, StateId, PairHasher<StateId> > MapType;
-  typedef MapType::iterator IterType;
-
-  if (clat.Start() == kNoStateId) return;
-  // Make sure the input lattice is topologically sorted.
-  if (clat.Properties(kTopSorted, true) == 0) {
-    CompactLattice clat_copy(clat);
-    KALDI_LOG << "Topsort this lattice.";
-    if (!TopSort(&clat_copy))
-      KALDI_ERR << "Was not able to topologically sort lattice (cycles found?)";
-    ExpandCompactLattice(clat_copy, epsilon, expand_clat);
-    return;
-  }
-
-  // Compute backward logprobs betas for the expanded lattice.
-  // Note: the backward logprobs in the original lattice <clat> and the
-  // expanded lattice <expand_clat> are the same.
-  int32 num_states = clat.NumStates();
-  std::vector<double> beta(num_states, kLogZeroDouble);
-  ComputeCompactLatticeBetas(clat, &beta);
-  double tot_backward_logprob = beta[0];
-  std::vector<double> alpha;
-  alpha.push_back(0.0);
-  expand_clat->DeleteStates();
-  MapType state_map; // Map from state pair (orig_state, copy_state) to
-  // copy_state, where orig_state is a state in the original lattice, and
-  // copy_state is its corresponding one in the expanded lattice.
-  unordered_map<StateId, StateId> states; // Map from orig_state to its
-  // copy_state for states with incoming arcs' posteriors <= epsilon.
-  std::queue<StatePair> state_queue;
-
-  // Set start state in the expanded lattice.
-  StateId start_state = expand_clat->AddState();
-  expand_clat->SetStart(start_state);
-  StatePair start_pair(clat.Start(), start_state);
-  state_queue.push(start_pair);
-  std::pair<IterType, bool> result =
-    state_map.insert(std::make_pair(start_pair, start_state));
-  KALDI_ASSERT(result.second == true);
-
-  // Expand <clat> and update forward logprobs alphas in <expand_clat>.
-  while (!state_queue.empty()) {
-    StatePair s = state_queue.front();
-    StateId s1 = s.first,
-            s2 = s.second;
-    state_queue.pop();
-
-    Weight f = clat.Final(s1);
-    if (f != Weight::Zero()) {
-      KALDI_ASSERT(state_map.find(s) != state_map.end());
-      expand_clat->SetFinal(state_map[s], f);
-    }
-
-    for (ArcIterator<CompactLattice> aiter(clat, s1);
-         !aiter.Done(); aiter.Next()) {
-      const Arc &arc = aiter.Value();
-      StateId orig_state = arc.nextstate;
-      double arc_like = -ConvertToCost(arc.weight),
-             this_alpha = alpha[s2] + arc_like,
-             arc_post = Exp(this_alpha + beta[orig_state] -
-                            tot_backward_logprob);
-      // Generate the expanded lattice.
-      StateId copy_state;
-      if (arc_post > epsilon) {
-        copy_state = expand_clat->AddState();
-        StatePair next_pair(orig_state, copy_state);
-        std::pair<IterType, bool> result =
-          state_map.insert(std::make_pair(next_pair, copy_state));
-        KALDI_ASSERT(result.second == true);
-        state_queue.push(next_pair);
-      } else {
-        unordered_map<StateId, StateId>::iterator iter = states.find(orig_state);
-        if (iter == states.end() ) { // The counterpart state of orig_state
-                                   // has not been created in <expand_clat> yet.
-          copy_state = expand_clat->AddState();
-          StatePair next_pair(orig_state, copy_state);
-          std::pair<IterType, bool> result =
-            state_map.insert(std::make_pair(next_pair, copy_state));
-          KALDI_ASSERT(result.second == true);
-          state_queue.push(next_pair);
-          states[orig_state] = copy_state;
-        } else {
-          copy_state = iter->second;
-        }
-      }
-      // Create an arc from state_map[s] to copy_state in the expanded lattice.
-      expand_clat->AddArc(state_map[s], Arc(arc.ilabel, arc.olabel, arc.weight,
-                                            copy_state));
-      // Compute forward logprobs alpha for the expanded lattice.
-      if ((alpha.size() - 1) < copy_state) { // The first time to compute alpha
-                                             // for copy_state in <expand_clat>.
-        alpha.push_back(this_alpha);
-      } else { // Accumulate alpha.
-        alpha[copy_state] = LogAdd(alpha[copy_state], this_alpha);
-      }
-    }
-  } // end while
-}
-
-
-void CompactLatticeBestCostsAndTracebacks(
-    const CompactLattice &clat,
-    CostTraceType *forward_best_cost_and_pred,
-    CostTraceType *backward_best_cost_and_pred) {
-
-  // typedef the arc, weight types
-  typedef CompactLatticeArc Arc;
-  typedef Arc::Weight Weight;
-  typedef Arc::StateId StateId;
-
-  forward_best_cost_and_pred->clear();
-  backward_best_cost_and_pred->clear();
-  forward_best_cost_and_pred->resize(clat.NumStates());
-  backward_best_cost_and_pred->resize(clat.NumStates());
-  // Initialize the cost and predecessor state for each state.
-  for (StateId s = 0; s < clat.NumStates(); s++) {
-    (*forward_best_cost_and_pred)[s].first =
-                                        std::numeric_limits<double>::infinity();
-    (*backward_best_cost_and_pred)[s].first =
-                                        std::numeric_limits<double>::infinity();
-    (*forward_best_cost_and_pred)[s].second = fst::kNoStateId;
-    (*backward_best_cost_and_pred)[s].second = fst::kNoStateId;
-  }
-
-  StateId start_state = clat.Start();
-  (*forward_best_cost_and_pred)[start_state].first = 0;
-  // Transverse the lattice forwardly to compute the best cost from the start
-  // state to each state and the best predecessor state of each state.
-  for (StateId s = 0; s < clat.NumStates(); s++) {
-    double cur_cost = (*forward_best_cost_and_pred)[s].first;
-    for (fst::ArcIterator<CompactLattice> aiter(clat, s);
-         !aiter.Done(); aiter.Next()) {
-      const Arc &arc = aiter.Value();
-      double next_cost = cur_cost + ConvertToCost(arc.weight);
-      if (next_cost < (*forward_best_cost_and_pred)[arc.nextstate].first) {
-        (*forward_best_cost_and_pred)[arc.nextstate].first = next_cost;
-        (*forward_best_cost_and_pred)[arc.nextstate].second = s;
-      }
-    }
-  }
-  // Transverse the lattice backwardly to compute the best cost from a final
-  // state to each state and the best predecessor state of each state.
-  for (StateId s = clat.NumStates() - 1; s >= 0; s--) {
-    double this_cost = ConvertToCost(clat.Final(s));
-    for (fst::ArcIterator<CompactLattice> aiter(clat, s);
-         !aiter.Done(); aiter.Next()) {
-      const Arc &arc = aiter.Value();
-      double next_cost = (*backward_best_cost_and_pred)[arc.nextstate].first +
-        ConvertToCost(arc.weight);
-      if (next_cost < this_cost) {
-        this_cost = next_cost;
-        (*backward_best_cost_and_pred)[s].second = arc.nextstate;
-      }
-    }
-    (*backward_best_cost_and_pred)[s].first = this_cost;
-  }
-}
-
-
-void AddNnlmScoreToCompactLattice(const MapT &nnlm_scores,
-                                  CompactLattice *clat) {
-  if (clat->Start() == fst::kNoStateId) return;
-  // Make sure the input lattice is topologically sorted.
-  if (clat->Properties(fst::kTopSorted, true) == 0) {
-    KALDI_LOG << "Topsort this lattice.";
-    if (!TopSort(clat))
-      KALDI_ERR << "Was not able to topologically sort lattice (cycles found?)";
-    AddNnlmScoreToCompactLattice(nnlm_scores, clat);
-    return;
-  }
-
-  // typedef the arc, weight types
-  typedef CompactLatticeArc Arc;
-  typedef Arc::Weight Weight;
-  typedef Arc::StateId StateId;
-  typedef std::pair<int32, int32> StatePair;
-
-  int32 num_states = clat->NumStates();
-  unordered_map<StatePair, bool, PairHasher<int32> > final_state_check;
-  for (StateId s = 0; s < num_states; s++) {
-    for (fst::MutableArcIterator<CompactLattice> aiter(clat, s);
-         !aiter.Done(); aiter.Next()) {
-      Arc arc(aiter.Value());
-      StatePair arc_index = std::make_pair(static_cast<int32>(s),
-                                           static_cast<int32>(arc.nextstate));
-      MapT::const_iterator it = nnlm_scores.find(arc_index);
-      double nnlm_score;
-      if (it != nnlm_scores.end())
-        nnlm_score = it->second;
-      else
-        KALDI_ERR << "Some arc does not have neural language model score.";
-      if (arc.ilabel != 0) { // if there is a word on this arc
-        LatticeWeight weight = arc.weight.Weight();
-        // Add associated neural LM score to each arc.
-        weight.SetValue1(weight.Value1() + nnlm_score);
-        arc.weight.SetWeight(weight);
-        aiter.SetValue(arc);
-      }
-      Weight clat_final = clat->Final(arc.nextstate);
-      StatePair final_pair = std::make_pair(arc.nextstate, arc.nextstate);
-      // Add neural LM scores to each final state only once.
-      if (clat_final != CompactLatticeWeight::Zero() &&
-          final_state_check.find(final_pair) == final_state_check.end()) {
-        MapT::const_iterator final_it = nnlm_scores.find(final_pair);
-        double final_nnlm_score = 0.0;
-        if (final_it != nnlm_scores.end())
-          final_nnlm_score = final_it->second;
-        // Add neural LM scores to the final weight.
-        Weight final_weight(LatticeWeight(clat_final.Weight().Value1() +
-                                          final_nnlm_score,
-                                          clat_final.Weight().Value2()),
-                                          clat_final.String());
-        clat->SetFinal(arc.nextstate, final_weight);
-        final_state_check[final_pair] = true;
-      }
-    } // end looping over arcs
-  } // end looping over states
-}
-
-void AddWordInsPenToCompactLattice(BaseFloat word_ins_penalty,
-                                   CompactLattice *clat) {
-  typedef CompactLatticeArc Arc;
-  int32 num_states = clat->NumStates();
-
-  //scan the lattice
-  for (int32 state = 0; state < num_states; state++) {
-    for (fst::MutableArcIterator<CompactLattice> aiter(clat, state);
-         !aiter.Done(); aiter.Next()) {
-
-      Arc arc(aiter.Value());
-
-      if (arc.ilabel != 0) { // if there is a word on this arc
-        LatticeWeight weight = arc.weight.Weight();
-        // add word insertion penalty to lattice
-        weight.SetValue1( weight.Value1() + word_ins_penalty);
-        arc.weight.SetWeight(weight);
-        aiter.SetValue(arc);
-      }
-    } // end looping over arcs
-  }  // end looping over states
-}
-
-struct ClatRescoreTuple {
-  ClatRescoreTuple(int32 state, int32 arc, int32 tid):
-      state_id(state), arc_id(arc), tid(tid) { }
-  int32 state_id;
-  int32 arc_id;
-  int32 tid;
-};
-
-/** RescoreCompactLatticeInternal is the internal code for both
-    RescoreCompactLattice and RescoreCompatLatticeSpeedup.  For
-    RescoreCompactLattice, "tmodel" will be NULL and speedup_factor will be 1.0.
- */
-bool RescoreCompactLatticeInternal(
-    const TransitionInformation *tmodel,
-    BaseFloat speedup_factor,
-    DecodableInterface *decodable,
-    CompactLattice *clat) {
-  KALDI_ASSERT(speedup_factor >= 1.0);
-  if (clat->NumStates() == 0) {
-    KALDI_WARN << "Rescoring empty lattice";
-    return false;
-  }
-  if (!clat->Properties(fst::kTopSorted, true)) {
-    if (fst::TopSort(clat) == false) {
-      KALDI_WARN << "Cycles detected in lattice.";
-      return false;
-    }
-  }
-  std::vector<int32> state_times;
-  int32 utt_len = kaldi::CompactLatticeStateTimes(*clat, &state_times);
-
-  std::vector<std::vector<ClatRescoreTuple> > time_to_state(utt_len);
-
-  int32 num_states = clat->NumStates();
-  KALDI_ASSERT(num_states == state_times.size());
-  for (size_t state = 0; state < num_states; state++) {
-    KALDI_ASSERT(state_times[state] >= 0);
-    int32 t = state_times[state];
-    int32 arc_id = 0;
-    for (fst::MutableArcIterator<CompactLattice> aiter(clat, state);
-         !aiter.Done(); aiter.Next(), arc_id++) {
-      CompactLatticeArc arc = aiter.Value();
-      std::vector<int32> arc_string = arc.weight.String();
-
-      for (size_t offset = 0; offset < arc_string.size(); offset++) {
-        if (t < utt_len) { // end state may be past this..
-          int32 tid = arc_string[offset];
-          time_to_state[t+offset].push_back(ClatRescoreTuple(state, arc_id, tid));
-        } else {
-          if (t != utt_len) {
-            KALDI_WARN << "There appears to be lattice/feature mismatch, "
-                       << "aborting.";
-            return false;
-          }
-        }
-      }
-    }
-    if (clat->Final(state) != CompactLatticeWeight::Zero()) {
-      arc_id = -1;
-      std::vector<int32> arc_string = clat->Final(state).String();
-      for (size_t offset = 0; offset < arc_string.size(); offset++) {
-        KALDI_ASSERT(t + offset < utt_len); // already checked in
-        // CompactLatticeStateTimes, so would be code error.
-        time_to_state[t+offset].push_back(
-            ClatRescoreTuple(state, arc_id, arc_string[offset]));
-      }
-    }
-  }
-
-  for (int32 t = 0; t < utt_len; t++) {
-    if ((t < utt_len - 1) && decodable->IsLastFrame(t)) {
-      KALDI_WARN << "Features are too short for lattice: utt-len is "
-                 << utt_len << ", " << t << " is last frame";
-      return false;
-    }
-    // frame_scale is the scale we put on the computed acoustic probs for this
-    // frame.  It will always be 1.0 if tmodel == NULL (i.e. if we are not doing
-    // the "speedup" code).  For frames with multiple pdf-ids it will be one.
-    // For frames with only one pdf-id, it will equal speedup_factor (>=1.0)
-    // with probability 1.0 / speedup_factor, and zero otherwise.  If it is zero,
-    // we can avoid computing the probabilities.
-    BaseFloat frame_scale = 1.0;
-    KALDI_ASSERT(!time_to_state[t].empty());
-    if (tmodel != NULL) {
-      int32 pdf_id = tmodel->TransitionIdToPdf(time_to_state[t][0].tid);
-      bool frame_has_multiple_pdfs = false;
-      for (size_t i = 1; i < time_to_state[t].size(); i++) {
-        if (tmodel->TransitionIdToPdf(time_to_state[t][i].tid) != pdf_id) {
-          frame_has_multiple_pdfs = true;
-          break;
-        }
-      }
-      if (frame_has_multiple_pdfs) {
-        frame_scale = 1.0;
-      } else {
-        if (WithProb(1.0 / speedup_factor)) {
-          frame_scale = speedup_factor;
-        } else {
-          frame_scale = 0.0;
-        }
-      }
-      if (frame_scale == 0.0)
-        continue; // the code below would be pointless.
-    }
-
-    for (size_t i = 0; i < time_to_state[t].size(); i++) {
-      int32 state = time_to_state[t][i].state_id;
-      int32 arc_id = time_to_state[t][i].arc_id;
-      int32 tid = time_to_state[t][i].tid;
-
-      if (arc_id == -1) { // Final state
-        // Access the trans_id
-        CompactLatticeWeight curr_clat_weight = clat->Final(state);
-
-        // Calculate likelihood
-        BaseFloat log_like = decodable->LogLikelihood(t, tid) * frame_scale;
-        // update weight
-        CompactLatticeWeight new_clat_weight = curr_clat_weight;
-        LatticeWeight new_lat_weight = new_clat_weight.Weight();
-        new_lat_weight.SetValue2(-log_like + curr_clat_weight.Weight().Value2());
-        new_clat_weight.SetWeight(new_lat_weight);
-        clat->SetFinal(state, new_clat_weight);
-      } else {
-        fst::MutableArcIterator<CompactLattice> aiter(clat, state);
-
-        aiter.Seek(arc_id);
-        CompactLatticeArc arc = aiter.Value();
-
-        // Calculate likelihood
-        BaseFloat log_like = decodable->LogLikelihood(t, tid) * frame_scale;
-        // update weight
-        LatticeWeight new_weight = arc.weight.Weight();
-        new_weight.SetValue2(-log_like + arc.weight.Weight().Value2());
-        arc.weight.SetWeight(new_weight);
-        aiter.SetValue(arc);
-      }
-    }
-  }
-  return true;
-}
-
-
-bool RescoreCompactLatticeSpeedup(
-    const TransitionInformation &tmodel,
-    BaseFloat speedup_factor,
-    DecodableInterface *decodable,
-    CompactLattice *clat) {
-  return RescoreCompactLatticeInternal(&tmodel, speedup_factor, decodable, clat);
-}
-
-bool RescoreCompactLattice(DecodableInterface *decodable,
-                           CompactLattice *clat) {
-  return RescoreCompactLatticeInternal(NULL, 1.0, decodable, clat);
-}
-
-
-bool RescoreLattice(DecodableInterface *decodable,
-                    Lattice *lat) {
-  if (lat->NumStates() == 0) {
-    KALDI_WARN << "Rescoring empty lattice";
-    return false;
-  }
-  if (!lat->Properties(fst::kTopSorted, true)) {
-    if (fst::TopSort(lat) == false) {
-      KALDI_WARN << "Cycles detected in lattice.";
-      return false;
-    }
-  }
-  std::vector<int32> state_times;
-  int32 utt_len = kaldi::LatticeStateTimes(*lat, &state_times);
-
-  std::vector<std::vector<int32> > time_to_state(utt_len );
-
-  int32 num_states = lat->NumStates();
-  KALDI_ASSERT(num_states == state_times.size());
-  for (size_t state = 0; state < num_states; state++) {
-    int32 t = state_times[state];
-    // Don't check t >= 0 because non-accessible states could have t = -1.
-    KALDI_ASSERT(t <= utt_len);
-    if (t >= 0 && t < utt_len)
-      time_to_state[t].push_back(state);
-  }
-
-  for (int32 t = 0; t < utt_len; t++) {
-    if ((t < utt_len - 1) && decodable->IsLastFrame(t)) {
-      KALDI_WARN << "Features are too short for lattice: utt-len is "
-                 << utt_len << ", " << t << " is last frame";
-      return false;
-    }
-    for (size_t i = 0; i < time_to_state[t].size(); i++) {
-      int32 state = time_to_state[t][i];
-      for (fst::MutableArcIterator<Lattice> aiter(lat, state);
-           !aiter.Done(); aiter.Next()) {
-        LatticeArc arc = aiter.Value();
-        if (arc.ilabel != 0) {
-          int32 trans_id = arc.ilabel; // Note: it doesn't necessarily
-          // have to be a transition-id, just whatever the Decodable
-          // object is expecting, but it's normally a transition-id.
-
-          BaseFloat log_like = decodable->LogLikelihood(t, trans_id);
-          arc.weight.SetValue2(-log_like + arc.weight.Value2());
-          aiter.SetValue(arc);
-        }
-      }
-    }
-  }
-  return true;
-}
-
-
-int32 LongestSentenceLength(const Lattice &lat) {
-  typedef Lattice::Arc Arc;
-  typedef Arc::Label Label;
-  typedef Arc::StateId StateId;
-
-  if (lat.Properties(fst::kTopSorted, true) == 0) {
-    Lattice lat_copy(lat);
-    if (!TopSort(&lat_copy))
-      KALDI_ERR << "Was not able to topologically sort lattice (cycles found?)";
-    return LongestSentenceLength(lat_copy);
-  }
-  std::vector<int32> max_length(lat.NumStates(), 0);
-  int32 lattice_max_length = 0;
-  for (StateId s = 0; s < lat.NumStates(); s++) {
-    int32 this_max_length = max_length[s];
-    for (fst::ArcIterator<Lattice> aiter(lat, s); !aiter.Done(); aiter.Next()) {
-      const Arc &arc = aiter.Value();
-      bool arc_has_word = (arc.olabel != 0);
-      StateId nextstate = arc.nextstate;
-      KALDI_ASSERT(static_cast<size_t>(nextstate) < max_length.size());
-      if (arc_has_word) {
-        // A lattice should ideally not have cycles anyway; a cycle with a word
-        // on is something very bad.
-        KALDI_ASSERT(nextstate > s && "Lattice has cycles with words on.");
-        max_length[nextstate] = std::max(max_length[nextstate],
-                                         this_max_length + 1);
-      } else {
-        max_length[nextstate] = std::max(max_length[nextstate],
-                                         this_max_length);
-      }
-    }
-    if (lat.Final(s) != LatticeWeight::Zero())
-      lattice_max_length = std::max(lattice_max_length, max_length[s]);
-  }
-  return lattice_max_length;
-}
-
-int32 LongestSentenceLength(const CompactLattice &clat) {
-  typedef CompactLattice::Arc Arc;
-  typedef Arc::Label Label;
-  typedef Arc::StateId StateId;
-
-  if (clat.Properties(fst::kTopSorted, true) == 0) {
-    CompactLattice clat_copy(clat);
-    if (!TopSort(&clat_copy))
-      KALDI_ERR << "Was not able to topologically sort lattice (cycles found?)";
-    return LongestSentenceLength(clat_copy);
-  }
-  std::vector<int32> max_length(clat.NumStates(), 0);
-  int32 lattice_max_length = 0;
-  for (StateId s = 0; s < clat.NumStates(); s++) {
-    int32 this_max_length = max_length[s];
-    for (fst::ArcIterator<CompactLattice> aiter(clat, s);
-         !aiter.Done(); aiter.Next()) {
-      const Arc &arc = aiter.Value();
-      bool arc_has_word = (arc.ilabel != 0); // note: olabel == ilabel.
-      // also note: for normal CompactLattice, e.g. as produced by
-      // determinization, all arcs will have nonzero labels, but the user might
-      // decide to remplace some of the labels with zero for some reason, and we
-      // want to support this.
-      StateId nextstate = arc.nextstate;
-      KALDI_ASSERT(static_cast<size_t>(nextstate) < max_length.size());
-      KALDI_ASSERT(nextstate > s && "CompactLattice has cycles");
-      if (arc_has_word)
-        max_length[nextstate] = std::max(max_length[nextstate],
-                                         this_max_length + 1);
-      else
-        max_length[nextstate] = std::max(max_length[nextstate],
-                                         this_max_length);
-    }
-    if (clat.Final(s) != CompactLatticeWeight::Zero())
-      lattice_max_length = std::max(lattice_max_length, max_length[s]);
-  }
-  return lattice_max_length;
-}
-
-void ComposeCompactLatticeDeterministic(
-    const CompactLattice& clat,
-    fst::DeterministicOnDemandFst<fst::StdArc>* det_fst,
-    CompactLattice* composed_clat) {
-  // StdFst::Arc and CompactLatticeArc has the same StateId type.
-  typedef fst::StdArc::StateId StateId;
-  typedef fst::StdArc::Weight Weight1;
-  typedef CompactLatticeArc::Weight Weight2;
-  typedef std::pair<StateId, StateId> StatePair;
-  typedef unordered_map<StatePair, StateId, PairHasher<StateId> > MapType;
-  typedef MapType::iterator IterType;
-
-  // Empties the output FST.
-  KALDI_ASSERT(composed_clat != NULL);
-  composed_clat->DeleteStates();
-
-  MapType state_map;
-  std::queue<StatePair> state_queue;
-
-  // Sets start state in <composed_clat>.
-  StateId start_state = composed_clat->AddState();
-  StatePair start_pair(clat.Start(), det_fst->Start());
-  composed_clat->SetStart(start_state);
-  state_queue.push(start_pair);
-  std::pair<IterType, bool> result =
-      state_map.insert(std::make_pair(start_pair, start_state));
-  KALDI_ASSERT(result.second == true);
-
-  // Starts composition here.
-  while (!state_queue.empty()) {
-    // Gets the first state in the queue.
-    StatePair s = state_queue.front();
-    StateId s1 = s.first;
-    StateId s2 = s.second;
-    state_queue.pop();
-
-
-    Weight2 clat_final = clat.Final(s1);
-    if (clat_final.Weight().Value1() !=
-        std::numeric_limits<BaseFloat>::infinity()) {
-      // Test for whether the final-prob of state s1 was zero.
-      Weight1 det_fst_final = det_fst->Final(s2);
-      if (det_fst_final.Value() !=
-          std::numeric_limits<BaseFloat>::infinity()) {
-        // Test for whether the final-prob of state s2 was zero.  If neither
-        // source-state final prob was zero, then we should create final state
-        // in fst_composed. We compute the product manually since this is more
-        // efficient.
-        Weight2 final_weight(LatticeWeight(clat_final.Weight().Value1() +
-                                           det_fst_final.Value(),
-                                           clat_final.Weight().Value2()),
-                             clat_final.String());
-        // we can assume final_weight is not Zero(), since neither of
-        // the sources was zero.
-        KALDI_ASSERT(state_map.find(s) != state_map.end());
-        composed_clat->SetFinal(state_map[s], final_weight);
-      }
-    }
-
-    // Loops over pair of edges at s1 and s2.
-    for (fst::ArcIterator<CompactLattice> aiter(clat, s1);
-         !aiter.Done(); aiter.Next()) {
-      const CompactLatticeArc& arc1 = aiter.Value();
-      fst::StdArc arc2;
-      StateId next_state1 = arc1.nextstate, next_state2;
-      bool matched = false;
-
-      if (arc1.olabel == 0) {
-        // If the symbol on <arc1> is <epsilon>, we transit to the next state
-        // for <clat>, but keep <det_fst> at the current state.
-        matched = true;
-        next_state2 = s2;
-      } else {
-        // Otherwise try to find the matched arc in <det_fst>.
-        matched = det_fst->GetArc(s2, arc1.olabel, &arc2);
-        if (matched) {
-          next_state2 = arc2.nextstate;
-        }
-      }
-
-      // If matched arc is found in <det_fst>, then we have to add new arcs to
-      // <composed_clat>.
-      if (matched) {
-        StatePair next_state_pair(next_state1, next_state2);
-        IterType siter = state_map.find(next_state_pair);
-        StateId next_state;
-
-        // Adds composed state to <state_map>.
-        if (siter == state_map.end()) {
-          // If the composed state has not been created yet, create it.
-          next_state = composed_clat->AddState();
-          std::pair<const StatePair, StateId> next_state_map(next_state_pair,
-                                                             next_state);
-          std::pair<IterType, bool> result = state_map.insert(next_state_map);
-          KALDI_ASSERT(result.second);
-          state_queue.push(next_state_pair);
-        } else {
-          // If the composed state is already in <state_map>, we can directly
-          // use that.
-          next_state = siter->second;
-        }
-
-        // Adds arc to <composed_clat>.
-        if (arc1.olabel == 0) {
-          composed_clat->AddArc(state_map[s],
-                                CompactLatticeArc(arc1.ilabel, 0,
-                                                  arc1.weight, next_state));
-        } else {
-          Weight2 composed_weight(
-              LatticeWeight(arc1.weight.Weight().Value1() +
-                            arc2.weight.Value(),
-                            arc1.weight.Weight().Value2()),
-              arc1.weight.String());
-          composed_clat->AddArc(state_map[s],
-                                CompactLatticeArc(arc1.ilabel, arc2.olabel,
-                                                  composed_weight, next_state));
-        }
-      }
-    }
-  }
-  fst::Connect(composed_clat);
-}
-
-
-void ComputeAcousticScoresMap(
-    const Lattice &lat,
-    unordered_map<std::pair<int32, int32>, std::pair<BaseFloat, int32>,
-                                        PairHasher<int32> > *acoustic_scores) {
-  // typedef the arc, weight types
-  typedef Lattice::Arc Arc;
-  typedef Arc::Weight LatticeWeight;
-  typedef Arc::StateId StateId;
-
-  acoustic_scores->clear();
-
-  std::vector<int32> state_times;
-  LatticeStateTimes(lat, &state_times);   // Assumes the input is top sorted
-
-  KALDI_ASSERT(lat.Start() == 0);
-
-  for (StateId s = 0; s < lat.NumStates(); s++) {
-    int32 t = state_times[s];
-    for (fst::ArcIterator<Lattice> aiter(lat, s); !aiter.Done();
-          aiter.Next()) {
-      const Arc &arc = aiter.Value();
-      const LatticeWeight &weight = arc.weight;
-
-      int32 tid = arc.ilabel;
-
-      if (tid != 0) {
-        unordered_map<std::pair<int32, int32>, std::pair<BaseFloat, int32>,
-          PairHasher<int32> >::iterator it = acoustic_scores->find(std::make_pair(t, tid));
-        if (it == acoustic_scores->end()) {
-          acoustic_scores->insert(std::make_pair(std::make_pair(t, tid),
-                                          std::make_pair(weight.Value2(), 1)));
-        } else {
-          if (it->second.second == 2
-                && it->second.first / it->second.second != weight.Value2()) {
-            KALDI_VLOG(2) << "Transitions on the same frame have different "
-                          << "acoustic costs for tid " << tid << "; "
-                          << it->second.first / it->second.second
-                          << " vs " << weight.Value2();
-          }
-          it->second.first += weight.Value2();
-          it->second.second++;
-        }
-      } else {
-        // Arcs with epsilon input label (tid) must have 0 acoustic cost
-        KALDI_ASSERT(weight.Value2() == 0);
-      }
-    }
-
-    LatticeWeight f = lat.Final(s);
-    if (f != LatticeWeight::Zero()) {
-      // Final acoustic cost must be 0 as we are reading from
-      // non-determinized, non-compact lattice
-      KALDI_ASSERT(f.Value2() == 0.0);
-    }
-  }
-}
-
-void ReplaceAcousticScoresFromMap(
-    const unordered_map<std::pair<int32, int32>, std::pair<BaseFloat, int32>,
-                                        PairHasher<int32> > &acoustic_scores,
-    Lattice *lat) {
-  // typedef the arc, weight types
-  typedef Lattice::Arc Arc;
-  typedef Arc::Weight LatticeWeight;
-  typedef Arc::StateId StateId;
-
-  TopSortLatticeIfNeeded(lat);
-
-  std::vector<int32> state_times;
-  LatticeStateTimes(*lat, &state_times);
-
-  KALDI_ASSERT(lat->Start() == 0);
-
-  for (StateId s = 0; s < lat->NumStates(); s++) {
-    int32 t = state_times[s];
-    for (fst::MutableArcIterator<Lattice> aiter(lat, s);
-         !aiter.Done(); aiter.Next()) {
-      Arc arc(aiter.Value());
-
-      int32 tid = arc.ilabel;
-      if (tid != 0) {
-        unordered_map<std::pair<int32, int32>, std::pair<BaseFloat, int32>,
-          PairHasher<int32> >::const_iterator it = acoustic_scores.find(std::make_pair(t, tid));
-        if (it == acoustic_scores.end()) {
-          KALDI_ERR << "Could not find tid " << tid << " at time " << t
-                    << " in the acoustic scores map.";
-        } else {
-          arc.weight.SetValue2(it->second.first / it->second.second);
-        }
-      } else {
-        // For epsilon arcs, set acoustic cost to 0.0
-        arc.weight.SetValue2(0.0);
-      }
-      aiter.SetValue(arc);
-    }
-
-    LatticeWeight f = lat->Final(s);
-    if (f != LatticeWeight::Zero()) {
-      // Set final acoustic cost to 0.0
-      f.SetValue2(0.0);
-      lat->SetFinal(s, f);
-    }
-  }
-}
+// BaseFloat LatticeForwardBackward(const Lattice &lat, Posterior *post,
+//                                  double *acoustic_like_sum) {
+//   // Note, Posterior is defined as follows:  Indexed [frame], then a list
+//   // of (transition-id, posterior-probability) pairs.
+//   // typedef std::vector<std::vector<std::pair<int32, BaseFloat> > > Posterior;
+//   using namespace fst;
+//   typedef Lattice::Arc Arc;
+//   typedef Arc::Weight Weight;
+//   typedef Arc::StateId StateId;
+//
+//   if (acoustic_like_sum) *acoustic_like_sum = 0.0;
+//
+//   // Make sure the lattice is topologically sorted.
+//   if (lat.Properties(fst::kTopSorted, true) == 0)
+//     KALDI_ERR << "Input lattice must be topologically sorted.";
+//   KALDI_ASSERT(lat.Start() == 0);
+//
+//   int32 num_states = lat.NumStates();
+//   vector<int32> state_times;
+//   int32 max_time = LatticeStateTimes(lat, &state_times);
+//   std::vector<double> alpha(num_states, kLogZeroDouble);
+//   std::vector<double> &beta(alpha); // we re-use the same memory for
+//   // this, but it's semantically distinct so we name it differently.
+//   double tot_forward_prob = kLogZeroDouble;
+//
+//   post->clear();
+//   post->resize(max_time);
+//
+//   alpha[0] = 0.0;
+//   // Propagate alphas forward.
+//   for (StateId s = 0; s < num_states; s++) {
+//     double this_alpha = alpha[s];
+//     for (ArcIterator<Lattice> aiter(lat, s); !aiter.Done(); aiter.Next()) {
+//       const Arc &arc = aiter.Value();
+//       double arc_like = -ConvertToCost(arc.weight);
+//       alpha[arc.nextstate] = LogAdd(alpha[arc.nextstate], this_alpha + arc_like);
+//     }
+//     Weight f = lat.Final(s);
+//     if (f != Weight::Zero()) {
+//       double final_like = this_alpha - (f.Value1() + f.Value2());
+//       tot_forward_prob = LogAdd(tot_forward_prob, final_like);
+//       KALDI_ASSERT(state_times[s] == max_time &&
+//                    "Lattice is inconsistent (final-prob not at max_time)");
+//     }
+//   }
+//   for (StateId s = num_states-1; s >= 0; s--) {
+//     Weight f = lat.Final(s);
+//     double this_beta = -(f.Value1() + f.Value2());
+//     for (ArcIterator<Lattice> aiter(lat, s); !aiter.Done(); aiter.Next()) {
+//       const Arc &arc = aiter.Value();
+//       double arc_like = -ConvertToCost(arc.weight),
+//           arc_beta = beta[arc.nextstate] + arc_like;
+//       this_beta = LogAdd(this_beta, arc_beta);
+//       int32 transition_id = arc.ilabel;
+//
+//       // The following "if" is an optimization to avoid un-needed exp().
+//       if (transition_id != 0 || acoustic_like_sum != NULL) {
+//         double posterior = Exp(alpha[s] + arc_beta - tot_forward_prob);
+//
+//         if (transition_id != 0) // Arc has a transition-id on it [not epsilon]
+//           (*post)[state_times[s]].push_back(std::make_pair(transition_id,
+//                                                            static_cast<kaldi::BaseFloat>(posterior)));
+//         if (acoustic_like_sum != NULL)
+//           *acoustic_like_sum -= posterior * arc.weight.Value2();
+//       }
+//     }
+//     if (acoustic_like_sum != NULL && f != Weight::Zero()) {
+//       double final_logprob = - ConvertToCost(f),
+//           posterior = Exp(alpha[s] + final_logprob - tot_forward_prob);
+//       *acoustic_like_sum -= posterior * f.Value2();
+//     }
+//     beta[s] = this_beta;
+//   }
+//   double tot_backward_prob = beta[0];
+//   if (!ApproxEqual(tot_forward_prob, tot_backward_prob, 1e-8)) {
+//     KALDI_WARN << "Total forward probability over lattice = " << tot_forward_prob
+//               << ", while total backward probability = " << tot_backward_prob;
+//   }
+//   // Now combine any posteriors with the same transition-id.
+//   for (int32 t = 0; t < max_time; t++)
+//     MergePairVectorSumming(&((*post)[t]));
+//   return tot_backward_prob;
+// }
+//
+//
+// void LatticeActivePhones(const Lattice &lat, const TransitionModel &trans,
+//                          const vector<int32> &silence_phones,
+//                          vector< std::set<int32> > *active_phones) {
+//   KALDI_ASSERT(IsSortedAndUniq(silence_phones));
+//   vector<int32> state_times;
+//   int32 num_states = lat.NumStates();
+//   int32 max_time = LatticeStateTimes(lat, &state_times);
+//   active_phones->clear();
+//   active_phones->resize(max_time);
+//   for (int32 state = 0; state < num_states; state++) {
+//     int32 cur_time = state_times[state];
+//     for (fst::ArcIterator<Lattice> aiter(lat, state); !aiter.Done();
+//         aiter.Next()) {
+//       const LatticeArc &arc = aiter.Value();
+//       if (arc.ilabel != 0) {  // Non-epsilon arc
+//         int32 phone = trans.TransitionIdToPhone(arc.ilabel);
+//         if (!std::binary_search(silence_phones.begin(),
+//                                 silence_phones.end(), phone))
+//           (*active_phones)[cur_time].insert(phone);
+//       }
+//     }  // end looping over arcs
+//   }  // end looping over states
+// }
+//
+// void ConvertLatticeToPhones(const TransitionModel &trans,
+//                             Lattice *lat) {
+//   typedef LatticeArc Arc;
+//   int32 num_states = lat->NumStates();
+//   for (int32 state = 0; state < num_states; state++) {
+//     for (fst::MutableArcIterator<Lattice> aiter(lat, state); !aiter.Done();
+//         aiter.Next()) {
+//       Arc arc(aiter.Value());
+//       arc.olabel = 0; // remove any word.
+//       if ((arc.ilabel != 0) // has a transition-id on input..
+//           && (trans.TransitionIdToHmmState(arc.ilabel) == 0)
+//           && (!trans.IsSelfLoop(arc.ilabel))) {
+//          // && trans.IsFinal(arc.ilabel)) // there is one of these per phone...
+//         arc.olabel = trans.TransitionIdToPhone(arc.ilabel);
+//       }
+//       aiter.SetValue(arc);
+//     }  // end looping over arcs
+//   }  // end looping over states
+// }
+//
+//
+// static inline double LogAddOrMax(bool viterbi, double a, double b) {
+//   if (viterbi)
+//     return std::max(a, b);
+//   else
+//     return LogAdd(a, b);
+// }
+//
+// template<typename LatticeType>
+// double ComputeLatticeAlphasAndBetas(const LatticeType &lat,
+//                                     bool viterbi,
+//                                     vector<double> *alpha,
+//                                     vector<double> *beta) {
+//   typedef typename LatticeType::Arc Arc;
+//   typedef typename Arc::Weight Weight;
+//   typedef typename Arc::StateId StateId;
+//
+//   StateId num_states = lat.NumStates();
+//   KALDI_ASSERT(lat.Properties(fst::kTopSorted, true) == fst::kTopSorted);
+//   KALDI_ASSERT(lat.Start() == 0);
+//   alpha->clear();
+//   beta->clear();
+//   alpha->resize(num_states, kLogZeroDouble);
+//   beta->resize(num_states, kLogZeroDouble);
+//
+//   double tot_forward_prob = kLogZeroDouble;
+//   (*alpha)[0] = 0.0;
+//   // Propagate alphas forward.
+//   for (StateId s = 0; s < num_states; s++) {
+//     double this_alpha = (*alpha)[s];
+//     for (fst::ArcIterator<LatticeType> aiter(lat, s); !aiter.Done();
+//          aiter.Next()) {
+//       const Arc &arc = aiter.Value();
+//       double arc_like = -ConvertToCost(arc.weight);
+//       (*alpha)[arc.nextstate] = LogAddOrMax(viterbi, (*alpha)[arc.nextstate],
+//                                                 this_alpha + arc_like);
+//     }
+//     Weight f = lat.Final(s);
+//     if (f != Weight::Zero()) {
+//       double final_like = this_alpha - ConvertToCost(f);
+//       tot_forward_prob = LogAddOrMax(viterbi, tot_forward_prob, final_like);
+//     }
+//   }
+//   for (StateId s = num_states-1; s >= 0; s--) { // it's guaranteed signed.
+//     double this_beta = -ConvertToCost(lat.Final(s));
+//     for (fst::ArcIterator<LatticeType> aiter(lat, s); !aiter.Done();
+//          aiter.Next()) {
+//       const Arc &arc = aiter.Value();
+//       double arc_like = -ConvertToCost(arc.weight),
+//           arc_beta = (*beta)[arc.nextstate] + arc_like;
+//       this_beta = LogAddOrMax(viterbi, this_beta, arc_beta);
+//     }
+//     (*beta)[s] = this_beta;
+//   }
+//   double tot_backward_prob = (*beta)[lat.Start()];
+//   if (!ApproxEqual(tot_forward_prob, tot_backward_prob, 1e-8)) {
+//     KALDI_WARN << "Total forward probability over lattice = " << tot_forward_prob
+//                << ", while total backward probability = " << tot_backward_prob;
+//   }
+//   // Split the difference when returning... they should be the same.
+//   return 0.5 * (tot_backward_prob + tot_forward_prob);
+// }
+//
+// // instantiate the template for Lattice and CompactLattice
+// template
+// double ComputeLatticeAlphasAndBetas(const Lattice &lat,
+//                                     bool viterbi,
+//                                     vector<double> *alpha,
+//                                     vector<double> *beta);
+//
+// template
+// double ComputeLatticeAlphasAndBetas(const CompactLattice &lat,
+//                                     bool viterbi,
+//                                     vector<double> *alpha,
+//                                     vector<double> *beta);
+//
+//
+//
+// /// This is used in CompactLatticeLimitDepth.
+// struct LatticeArcRecord {
+//   BaseFloat logprob; // logprob <= 0 is the best Viterbi logprob of this arc,
+//                      // minus the overall best-cost of the lattice.
+//   CompactLatticeArc::StateId state; // state in the lattice.
+//   size_t arc; // arc index within the state.
+//   bool operator < (const LatticeArcRecord &other) const {
+//     return logprob < other.logprob;
+//   }
+// };
+//
+// void CompactLatticeLimitDepth(int32 max_depth_per_frame,
+//                               CompactLattice *clat) {
+//   typedef CompactLatticeArc Arc;
+//   typedef Arc::Weight Weight;
+//   typedef Arc::StateId StateId;
+//
+//   if (clat->Start() == fst::kNoStateId) {
+//     KALDI_WARN << "Limiting depth of empty lattice.";
+//     return;
+//   }
+//   if (clat->Properties(fst::kTopSorted, true) == 0) {
+//     if (!TopSort(clat))
+//       KALDI_ERR << "Topological sorting of lattice failed.";
+//   }
+//
+//   vector<int32> state_times;
+//   int32 T = CompactLatticeStateTimes(*clat, &state_times);
+//
+//   // The alpha and beta quantities here are "viterbi" alphas and beta.
+//   std::vector<double> alpha;
+//   std::vector<double> beta;
+//   bool viterbi = true;
+//   double best_prob = ComputeLatticeAlphasAndBetas(*clat, viterbi,
+//                                                   &alpha, &beta);
+//
+//   std::vector<std::vector<LatticeArcRecord> > arc_records(T);
+//
+//   StateId num_states = clat->NumStates();
+//   for (StateId s = 0; s < num_states; s++) {
+//     for (fst::ArcIterator<CompactLattice> aiter(*clat, s); !aiter.Done();
+//          aiter.Next()) {
+//       const Arc &arc = aiter.Value();
+//       LatticeArcRecord arc_record;
+//       arc_record.state = s;
+//       arc_record.arc = aiter.Position();
+//       arc_record.logprob =
+//           (alpha[s] + beta[arc.nextstate] - ConvertToCost(arc.weight))
+//            - best_prob;
+//       KALDI_ASSERT(arc_record.logprob < 0.1); // Should be zero or negative.
+//       int32 num_frames = arc.weight.String().size(), start_t = state_times[s];
+//       for (int32 t = start_t; t < start_t + num_frames; t++) {
+//         KALDI_ASSERT(t < T);
+//         arc_records[t].push_back(arc_record);
+//       }
+//     }
+//   }
+//   StateId dead_state = clat->AddState(); // A non-coaccesible state which we use
+//                                          // to remove arcs (make them end
+//                                          // there).
+//   size_t max_depth = max_depth_per_frame;
+//   for (int32 t = 0; t < T; t++) {
+//     size_t size = arc_records[t].size();
+//     if (size > max_depth) {
+//       // we sort from worst to best, so we keep the later-numbered ones,
+//       // and delete the lower-numbered ones.
+//       size_t cutoff = size - max_depth;
+//       std::nth_element(arc_records[t].begin(),
+//                        arc_records[t].begin() + cutoff,
+//                        arc_records[t].end());
+//       for (size_t index = 0; index < cutoff; index++) {
+//         LatticeArcRecord record(arc_records[t][index]);
+//         fst::MutableArcIterator<CompactLattice> aiter(clat, record.state);
+//         aiter.Seek(record.arc);
+//         Arc arc = aiter.Value();
+//         if (arc.nextstate != dead_state) { // not already killed.
+//           arc.nextstate = dead_state;
+//           aiter.SetValue(arc);
+//         }
+//       }
+//     }
+//   }
+//   Connect(clat);
+//   TopSortCompactLatticeIfNeeded(clat);
+// }
+//
+//
+// void TopSortCompactLatticeIfNeeded(CompactLattice *clat) {
+//   if (clat->Properties(fst::kTopSorted, true) == 0) {
+//     if (fst::TopSort(clat) == false) {
+//       KALDI_ERR << "Topological sorting failed";
+//     }
+//   }
+// }
+//
+// void TopSortLatticeIfNeeded(Lattice *lat) {
+//   if (lat->Properties(fst::kTopSorted, true) == 0) {
+//     if (fst::TopSort(lat) == false) {
+//       KALDI_ERR << "Topological sorting failed";
+//     }
+//   }
+// }
+//
+//
+// /// Returns the depth of the lattice, defined as the average number of
+// /// arcs crossing any given frame.  Returns 1 for empty lattices.
+// /// Requires that input is topologically sorted.
+// BaseFloat CompactLatticeDepth(const CompactLattice &clat,
+//                               int32 *num_frames) {
+//   typedef CompactLattice::Arc::StateId StateId;
+//   if (clat.Properties(fst::kTopSorted, true) == 0) {
+//     KALDI_ERR << "Lattice input to CompactLatticeDepth was not topologically "
+//               << "sorted.";
+//   }
+//   if (clat.Start() == fst::kNoStateId) {
+//     *num_frames = 0;
+//     return 1.0;
+//   }
+//   size_t num_arc_frames = 0;
+//   int32 t;
+//   {
+//     vector<int32> state_times;
+//     t = CompactLatticeStateTimes(clat, &state_times);
+//   }
+//   if (num_frames != NULL)
+//     *num_frames = t;
+//   for (StateId s = 0; s < clat.NumStates(); s++) {
+//     for (fst::ArcIterator<CompactLattice> aiter(clat, s); !aiter.Done();
+//          aiter.Next()) {
+//       const CompactLatticeArc &arc = aiter.Value();
+//       num_arc_frames += arc.weight.String().size();
+//     }
+//     num_arc_frames += clat.Final(s).String().size();
+//   }
+//   return num_arc_frames / static_cast<BaseFloat>(t);
+// }
+//
+//
+// void CompactLatticeDepthPerFrame(const CompactLattice &clat,
+//                                  std::vector<int32> *depth_per_frame) {
+//   typedef CompactLattice::Arc::StateId StateId;
+//   if (clat.Properties(fst::kTopSorted, true) == 0) {
+//     KALDI_ERR << "Lattice input to CompactLatticeDepthPerFrame was not "
+//               << "topologically sorted.";
+//   }
+//   if (clat.Start() == fst::kNoStateId) {
+//     depth_per_frame->clear();
+//     return;
+//   }
+//   vector<int32> state_times;
+//   int32 T = CompactLatticeStateTimes(clat, &state_times);
+//
+//   depth_per_frame->clear();
+//   if (T <= 0) {
+//     return;
+//   } else {
+//     depth_per_frame->resize(T, 0);
+//     for (StateId s = 0; s < clat.NumStates(); s++) {
+//       int32 start_time = state_times[s];
+//       for (fst::ArcIterator<CompactLattice> aiter(clat, s); !aiter.Done();
+//            aiter.Next()) {
+//         const CompactLatticeArc &arc = aiter.Value();
+//         int32 len = arc.weight.String().size();
+//         for (int32 t = start_time; t < start_time + len; t++) {
+//           KALDI_ASSERT(t < T);
+//           (*depth_per_frame)[t]++;
+//         }
+//       }
+//       int32 final_len = clat.Final(s).String().size();
+//       for (int32 t = start_time; t < start_time + final_len; t++) {
+//         KALDI_ASSERT(t < T);
+//         (*depth_per_frame)[t]++;
+//       }
+//     }
+//   }
+// }
+//
+//
+//
+// void ConvertCompactLatticeToPhones(const TransitionModel &trans,
+//                                    CompactLattice *clat) {
+//   typedef CompactLatticeArc Arc;
+//   typedef Arc::Weight Weight;
+//   int32 num_states = clat->NumStates();
+//   for (int32 state = 0; state < num_states; state++) {
+//     for (fst::MutableArcIterator<CompactLattice> aiter(clat, state);
+//          !aiter.Done();
+//          aiter.Next()) {
+//       Arc arc(aiter.Value());
+//       std::vector<int32> phone_seq;
+//       const std::vector<int32> &tid_seq = arc.weight.String();
+//       for (std::vector<int32>::const_iterator iter = tid_seq.begin();
+//            iter != tid_seq.end(); ++iter) {
+//         if (trans.IsFinal(*iter))// note: there is one of these per phone...
+//           phone_seq.push_back(trans.TransitionIdToPhone(*iter));
+//       }
+//       arc.weight.SetString(phone_seq);
+//       aiter.SetValue(arc);
+//     } // end looping over arcs
+//     Weight f = clat->Final(state);
+//     if (f != Weight::Zero()) {
+//       std::vector<int32> phone_seq;
+//       const std::vector<int32> &tid_seq = f.String();
+//       for (std::vector<int32>::const_iterator iter = tid_seq.begin();
+//            iter != tid_seq.end(); ++iter) {
+//         if (trans.IsFinal(*iter))// note: there is one of these per phone...
+//           phone_seq.push_back(trans.TransitionIdToPhone(*iter));
+//       }
+//       f.SetString(phone_seq);
+//       clat->SetFinal(state, f);
+//     }
+//   }  // end looping over states
+// }
+//
+// bool LatticeBoost(const TransitionModel &trans,
+//                   const std::vector<int32> &alignment,
+//                   const std::vector<int32> &silence_phones,
+//                   BaseFloat b,
+//                   BaseFloat max_silence_error,
+//                   Lattice *lat) {
+//   TopSortLatticeIfNeeded(lat);
+//
+//   // get all stored properties (test==false means don't test if not known).
+//   uint64 props = lat->Properties(fst::kFstProperties,
+//                                  false);
+//
+//   KALDI_ASSERT(IsSortedAndUniq(silence_phones));
+//   KALDI_ASSERT(max_silence_error >= 0.0 && max_silence_error <= 1.0);
+//   vector<int32> state_times;
+//   int32 num_states = lat->NumStates();
+//   int32 num_frames = LatticeStateTimes(*lat, &state_times);
+//   KALDI_ASSERT(num_frames == static_cast<int32>(alignment.size()));
+//   for (int32 state = 0; state < num_states; state++) {
+//     int32 cur_time = state_times[state];
+//     for (fst::MutableArcIterator<Lattice> aiter(lat, state); !aiter.Done();
+//          aiter.Next()) {
+//       LatticeArc arc = aiter.Value();
+//       if (arc.ilabel != 0) {  // Non-epsilon arc
+//         if (arc.ilabel < 0 || arc.ilabel > trans.NumTransitionIds()) {
+//           KALDI_WARN << "Lattice has out-of-range transition-ids: "
+//                      << "lattice/model mismatch?";
+//           return false;
+//         }
+//         int32 phone = trans.TransitionIdToPhone(arc.ilabel),
+//             ref_phone = trans.TransitionIdToPhone(alignment[cur_time]);
+//         BaseFloat frame_error;
+//         if (phone == ref_phone) {
+//           frame_error = 0.0;
+//         } else { // an error...
+//           if (std::binary_search(silence_phones.begin(), silence_phones.end(), phone))
+//             frame_error = max_silence_error;
+//           else
+//             frame_error = 1.0;
+//         }
+//         BaseFloat delta_cost = -b * frame_error; // negative cost if
+//         // frame is wrong, to boost likelihood of arcs with errors on them.
+//         // Add this cost to the graph part.
+//         arc.weight.SetValue1(arc.weight.Value1() + delta_cost);
+//         aiter.SetValue(arc);
+//       }
+//     }
+//   }
+//   // All we changed is the weights, so any properties that were
+//   // known before, are still known, except for whether or not the
+//   // lattice was weighted.
+//   lat->SetProperties(props,
+//                      ~(fst::kWeighted|fst::kUnweighted));
+//
+//   return true;
+// }
+//
+//
+//
+// BaseFloat LatticeForwardBackwardMpeVariants(
+//     const TransitionModel &trans,
+//     const std::vector<int32> &silence_phones,
+//     const Lattice &lat,
+//     const std::vector<int32> &num_ali,
+//     std::string criterion,
+//     bool one_silence_class,
+//     Posterior *post) {
+//   using namespace fst;
+//   typedef Lattice::Arc Arc;
+//   typedef Arc::Weight Weight;
+//   typedef Arc::StateId StateId;
+//
+//   KALDI_ASSERT(criterion == "mpfe" || criterion == "smbr");
+//   bool is_mpfe = (criterion == "mpfe");
+//
+//   if (lat.Properties(fst::kTopSorted, true) == 0)
+//     KALDI_ERR << "Input lattice must be topologically sorted.";
+//   KALDI_ASSERT(lat.Start() == 0);
+//
+//   int32 num_states = lat.NumStates();
+//   vector<int32> state_times;
+//   int32 max_time = LatticeStateTimes(lat, &state_times);
+//   KALDI_ASSERT(max_time == static_cast<int32>(num_ali.size()));
+//   std::vector<double> alpha(num_states, kLogZeroDouble),
+//       alpha_smbr(num_states, 0), //forward variable for sMBR
+//       beta(num_states, kLogZeroDouble),
+//       beta_smbr(num_states, 0); //backward variable for sMBR
+//
+//   double tot_forward_prob = kLogZeroDouble;
+//   double tot_forward_score = 0;
+//
+//   post->clear();
+//   post->resize(max_time);
+//
+//   alpha[0] = 0.0;
+//   // First Pass Forward,
+//   for (StateId s = 0; s < num_states; s++) {
+//     double this_alpha = alpha[s];
+//     for (ArcIterator<Lattice> aiter(lat, s); !aiter.Done(); aiter.Next()) {
+//       const Arc &arc = aiter.Value();
+//       double arc_like = -ConvertToCost(arc.weight);
+//       alpha[arc.nextstate] = LogAdd(alpha[arc.nextstate], this_alpha + arc_like);
+//     }
+//     Weight f = lat.Final(s);
+//     if (f != Weight::Zero()) {
+//       double final_like = this_alpha - (f.Value1() + f.Value2());
+//       tot_forward_prob = LogAdd(tot_forward_prob, final_like);
+//       KALDI_ASSERT(state_times[s] == max_time &&
+//                    "Lattice is inconsistent (final-prob not at max_time)");
+//     }
+//   }
+//   // First Pass Backward,
+//   for (StateId s = num_states-1; s >= 0; s--) {
+//     Weight f = lat.Final(s);
+//     double this_beta = -(f.Value1() + f.Value2());
+//     for (ArcIterator<Lattice> aiter(lat, s); !aiter.Done(); aiter.Next()) {
+//       const Arc &arc = aiter.Value();
+//       double arc_like = -ConvertToCost(arc.weight),
+//           arc_beta = beta[arc.nextstate] + arc_like;
+//       this_beta = LogAdd(this_beta, arc_beta);
+//     }
+//     beta[s] = this_beta;
+//   }
+//   // First Pass Forward-Backward Check
+//   double tot_backward_prob = beta[0];
+//   // may loose the condition somehow here 1e-6 (was 1e-8)
+//   if (!ApproxEqual(tot_forward_prob, tot_backward_prob, 1e-6)) {
+//     KALDI_ERR << "Total forward probability over lattice = " << tot_forward_prob
+//               << ", while total backward probability = " << tot_backward_prob;
+//   }
+//
+//   alpha_smbr[0] = 0.0;
+//   // Second Pass Forward, calculate forward for MPFE/SMBR
+//   for (StateId s = 0; s < num_states; s++) {
+//     double this_alpha = alpha[s];
+//     for (ArcIterator<Lattice> aiter(lat, s); !aiter.Done(); aiter.Next()) {
+//       const Arc &arc = aiter.Value();
+//       double arc_like = -ConvertToCost(arc.weight);
+//       double frame_acc = 0.0;
+//       if (arc.ilabel != 0) {
+//         int32 cur_time = state_times[s];
+//         int32 phone = trans.TransitionIdToPhone(arc.ilabel),
+//             ref_phone = trans.TransitionIdToPhone(num_ali[cur_time]);
+//         bool phone_is_sil = std::binary_search(silence_phones.begin(),
+//                                                silence_phones.end(),
+//                                                phone),
+//             ref_phone_is_sil = std::binary_search(silence_phones.begin(),
+//                                                   silence_phones.end(),
+//                                                   ref_phone),
+//             both_sil = phone_is_sil && ref_phone_is_sil;
+//         if (!is_mpfe) { // smbr.
+//           int32 pdf = trans.TransitionIdToPdf(arc.ilabel),
+//               ref_pdf = trans.TransitionIdToPdf(num_ali[cur_time]);
+//           if (!one_silence_class)  // old behavior
+//             frame_acc = (pdf == ref_pdf && !phone_is_sil) ? 1.0 : 0.0;
+//           else
+//             frame_acc = (pdf == ref_pdf || both_sil) ? 1.0 : 0.0;
+//         } else {
+//           if (!one_silence_class)  // old behavior
+//             frame_acc = (phone == ref_phone && !phone_is_sil) ? 1.0 : 0.0;
+//           else
+//             frame_acc = (phone == ref_phone || both_sil) ? 1.0 : 0.0;
+//         }
+//       }
+//       double arc_scale = Exp(alpha[s] + arc_like - alpha[arc.nextstate]);
+//       alpha_smbr[arc.nextstate] += arc_scale * (alpha_smbr[s] + frame_acc);
+//     }
+//     Weight f = lat.Final(s);
+//     if (f != Weight::Zero()) {
+//       double final_like = this_alpha - (f.Value1() + f.Value2());
+//       double arc_scale = Exp(final_like - tot_forward_prob);
+//       tot_forward_score += arc_scale * alpha_smbr[s];
+//       KALDI_ASSERT(state_times[s] == max_time &&
+//                    "Lattice is inconsistent (final-prob not at max_time)");
+//     }
+//   }
+//   // Second Pass Backward, collect Mpe style posteriors
+//   for (StateId s = num_states-1; s >= 0; s--) {
+//     for (ArcIterator<Lattice> aiter(lat, s); !aiter.Done(); aiter.Next()) {
+//       const Arc &arc = aiter.Value();
+//       double arc_like = -ConvertToCost(arc.weight),
+//           arc_beta = beta[arc.nextstate] + arc_like;
+//       double frame_acc = 0.0;
+//       int32 transition_id = arc.ilabel;
+//       if (arc.ilabel != 0) {
+//         int32 cur_time = state_times[s];
+//         int32 phone = trans.TransitionIdToPhone(arc.ilabel),
+//             ref_phone = trans.TransitionIdToPhone(num_ali[cur_time]);
+//         bool phone_is_sil = std::binary_search(silence_phones.begin(),
+//                                                silence_phones.end(), phone),
+//             ref_phone_is_sil = std::binary_search(silence_phones.begin(),
+//                                                   silence_phones.end(),
+//                                                   ref_phone),
+//             both_sil = phone_is_sil && ref_phone_is_sil;
+//         if (!is_mpfe) { // smbr.
+//           int32 pdf = trans.TransitionIdToPdf(arc.ilabel),
+//               ref_pdf = trans.TransitionIdToPdf(num_ali[cur_time]);
+//           if (!one_silence_class)  // old behavior
+//             frame_acc = (pdf == ref_pdf && !phone_is_sil) ? 1.0 : 0.0;
+//           else
+//             frame_acc = (pdf == ref_pdf || both_sil) ? 1.0 : 0.0;
+//         } else {
+//           if (!one_silence_class)  // old behavior
+//             frame_acc = (phone == ref_phone && !phone_is_sil) ? 1.0 : 0.0;
+//           else
+//             frame_acc = (phone == ref_phone || both_sil) ? 1.0 : 0.0;
+//         }
+//       }
+//       double arc_scale = Exp(beta[arc.nextstate] + arc_like - beta[s]);
+//       // check arc_scale NAN,
+//       // this is to prevent partial paths in Lattices
+//       // i.e., paths don't survive to the final state
+//       if (KALDI_ISNAN(arc_scale)) arc_scale = 0;
+//       beta_smbr[s] += arc_scale * (beta_smbr[arc.nextstate] + frame_acc);
+//
+//       if (transition_id != 0) { // Arc has a transition-id on it [not epsilon]
+//         double posterior = Exp(alpha[s] + arc_beta - tot_forward_prob);
+//         double acc_diff = alpha_smbr[s] + frame_acc + beta_smbr[arc.nextstate]
+//                                - tot_forward_score;
+//         double posterior_smbr = posterior * acc_diff;
+//         (*post)[state_times[s]].push_back(std::make_pair(transition_id,
+//                                                          static_cast<BaseFloat>(posterior_smbr)));
+//       }
+//     }
+//   }
+//
+//   //Second Pass Forward Backward check
+//   double tot_backward_score = beta_smbr[0];  // Initial state id == 0
+//   // may loose the condition somehow here 1e-5/1e-4
+//   if (!ApproxEqual(tot_forward_score, tot_backward_score, 1e-4)) {
+//     KALDI_ERR << "Total forward score over lattice = " << tot_forward_score
+//               << ", while total backward score = " << tot_backward_score;
+//   }
+//
+//   // Output the computed posteriors
+//   for (int32 t = 0; t < max_time; t++)
+//     MergePairVectorSumming(&((*post)[t]));
+//   return tot_forward_score;
+// }
+//
+// bool CompactLatticeToWordAlignment(const CompactLattice &clat,
+//                                    std::vector<int32> *words,
+//                                    std::vector<int32> *begin_times,
+//                                    std::vector<int32> *lengths) {
+//   words->clear();
+//   begin_times->clear();
+//   lengths->clear();
+//   typedef CompactLattice::Arc Arc;
+//   typedef Arc::Label Label;
+//   typedef CompactLattice::StateId StateId;
+//   typedef CompactLattice::Weight Weight;
+//   using namespace fst;
+//   StateId state = clat.Start();
+//   int32 cur_time = 0;
+//   if (state == kNoStateId) {
+//     KALDI_WARN << "Empty lattice.";
+//     return false;
+//   }
+//   while (1) {
+//     Weight final = clat.Final(state);
+//     size_t num_arcs = clat.NumArcs(state);
+//     if (final != Weight::Zero()) {
+//       if (num_arcs != 0) {
+//         KALDI_WARN << "Lattice is not linear.";
+//         return false;
+//       }
+//       if (! final.String().empty()) {
+//         KALDI_WARN << "Lattice has alignments on final-weight: probably "
+//             "was not word-aligned (alignments will be approximate)";
+//       }
+//       return true;
+//     } else {
+//       if (num_arcs != 1) {
+//         KALDI_WARN << "Lattice is not linear: num-arcs = " << num_arcs;
+//         return false;
+//       }
+//       fst::ArcIterator<CompactLattice> aiter(clat, state);
+//       const Arc &arc = aiter.Value();
+//       Label word_id = arc.ilabel; // Note: ilabel==olabel, since acceptor.
+//       // Also note: word_id may be zero; we output it anyway.
+//       int32 length = arc.weight.String().size();
+//       words->push_back(word_id);
+//       begin_times->push_back(cur_time);
+//       lengths->push_back(length);
+//       cur_time += length;
+//       state = arc.nextstate;
+//     }
+//   }
+// }
+//
+//
+// bool CompactLatticeToWordProns(
+//     const TransitionModel &tmodel,
+//     const CompactLattice &clat,
+//     std::vector<int32> *words,
+//     std::vector<int32> *begin_times,
+//     std::vector<int32> *lengths,
+//     std::vector<std::vector<int32> > *prons,
+//     std::vector<std::vector<int32> > *phone_lengths) {
+//   words->clear();
+//   begin_times->clear();
+//   lengths->clear();
+//   prons->clear();
+//   phone_lengths->clear();
+//   typedef CompactLattice::Arc Arc;
+//   typedef Arc::Label Label;
+//   typedef CompactLattice::StateId StateId;
+//   typedef CompactLattice::Weight Weight;
+//   using namespace fst;
+//   StateId state = clat.Start();
+//   int32 cur_time = 0;
+//   if (state == kNoStateId) {
+//     KALDI_WARN << "Empty lattice.";
+//     return false;
+//   }
+//   while (1) {
+//     Weight final = clat.Final(state);
+//     size_t num_arcs = clat.NumArcs(state);
+//     if (final != Weight::Zero()) {
+//       if (num_arcs != 0) {
+//         KALDI_WARN << "Lattice is not linear.";
+//         return false;
+//       }
+//       if (! final.String().empty()) {
+//         KALDI_WARN << "Lattice has alignments on final-weight: probably "
+//             "was not word-aligned (alignments will be approximate)";
+//       }
+//       return true;
+//     } else {
+//       if (num_arcs != 1) {
+//         KALDI_WARN << "Lattice is not linear: num-arcs = " << num_arcs;
+//         return false;
+//       }
+//       fst::ArcIterator<CompactLattice> aiter(clat, state);
+//       const Arc &arc = aiter.Value();
+//       Label word_id = arc.ilabel; // Note: ilabel==olabel, since acceptor.
+//       // Also note: word_id may be zero; we output it anyway.
+//       int32 length = arc.weight.String().size();
+//       words->push_back(word_id);
+//       begin_times->push_back(cur_time);
+//       lengths->push_back(length);
+//       const std::vector<int32> &arc_alignment = arc.weight.String();
+//       std::vector<std::vector<int32> > split_alignment;
+//       SplitToPhones(tmodel, arc_alignment, &split_alignment);
+//       std::vector<int32> phones(split_alignment.size());
+//       std::vector<int32> plengths(split_alignment.size());
+//       for (size_t i = 0; i < split_alignment.size(); i++) {
+//         KALDI_ASSERT(!split_alignment[i].empty());
+//         phones[i] = tmodel.TransitionIdToPhone(split_alignment[i][0]);
+//         plengths[i] = split_alignment[i].size();
+//       }
+//       prons->push_back(phones);
+//       phone_lengths->push_back(plengths);
+//
+//       cur_time += length;
+//       state = arc.nextstate;
+//     }
+//   }
+// }
+//
+//
+//
+// void CompactLatticeShortestPath(const CompactLattice &clat,
+//                                 CompactLattice *shortest_path) {
+//   using namespace fst;
+//   if (clat.Properties(fst::kTopSorted, true) == 0) {
+//     CompactLattice clat_copy(clat);
+//     if (!TopSort(&clat_copy))
+//       KALDI_ERR << "Was not able to topologically sort lattice (cycles found?)";
+//     CompactLatticeShortestPath(clat_copy, shortest_path);
+//     return;
+//   }
+//   // Now we can assume it's topologically sorted.
+//   shortest_path->DeleteStates();
+//   if (clat.Start() == kNoStateId) return;
+//   typedef CompactLatticeArc Arc;
+//   typedef Arc::StateId StateId;
+//   typedef CompactLatticeWeight Weight;
+//   vector<std::pair<double, StateId> > best_cost_and_pred(clat.NumStates() + 1);
+//   StateId superfinal = clat.NumStates();
+//   for (StateId s = 0; s <= clat.NumStates(); s++) {
+//     best_cost_and_pred[s].first = std::numeric_limits<double>::infinity();
+//     best_cost_and_pred[s].second = fst::kNoStateId;
+//   }
+//   best_cost_and_pred[clat.Start()].first = 0;
+//   for (StateId s = 0; s < clat.NumStates(); s++) {
+//     double my_cost = best_cost_and_pred[s].first;
+//     for (ArcIterator<CompactLattice> aiter(clat, s);
+//          !aiter.Done();
+//          aiter.Next()) {
+//       const Arc &arc = aiter.Value();
+//       double arc_cost = ConvertToCost(arc.weight),
+//           next_cost = my_cost + arc_cost;
+//       if (next_cost < best_cost_and_pred[arc.nextstate].first) {
+//         best_cost_and_pred[arc.nextstate].first = next_cost;
+//         best_cost_and_pred[arc.nextstate].second = s;
+//       }
+//     }
+//     double final_cost = ConvertToCost(clat.Final(s)),
+//         tot_final = my_cost + final_cost;
+//     if (tot_final < best_cost_and_pred[superfinal].first) {
+//       best_cost_and_pred[superfinal].first = tot_final;
+//       best_cost_and_pred[superfinal].second = s;
+//     }
+//   }
+//   std::vector<StateId> states; // states on best path.
+//   StateId cur_state = superfinal, start_state = clat.Start();
+//   while (cur_state != start_state) {
+//     StateId prev_state = best_cost_and_pred[cur_state].second;
+//     if (prev_state == kNoStateId) {
+//       KALDI_WARN << "Failure in best-path algorithm for lattice (infinite costs?)";
+//       return; // return empty best-path.
+//     }
+//     states.push_back(prev_state);
+//     KALDI_ASSERT(cur_state != prev_state && "Lattice with cycles");
+//     cur_state = prev_state;
+//   }
+//   std::reverse(states.begin(), states.end());
+//   for (size_t i = 0; i < states.size(); i++)
+//     shortest_path->AddState();
+//   for (StateId s = 0; static_cast<size_t>(s) < states.size(); s++) {
+//     if (s == 0) shortest_path->SetStart(s);
+//     if (static_cast<size_t>(s + 1) < states.size()) { // transition to next state.
+//       bool have_arc = false;
+//       Arc cur_arc;
+//       for (ArcIterator<CompactLattice> aiter(clat, states[s]);
+//            !aiter.Done();
+//            aiter.Next()) {
+//         const Arc &arc = aiter.Value();
+//         if (arc.nextstate == states[s+1]) {
+//           if (!have_arc ||
+//               ConvertToCost(arc.weight) < ConvertToCost(cur_arc.weight)) {
+//             cur_arc = arc;
+//             have_arc = true;
+//           }
+//         }
+//       }
+//       KALDI_ASSERT(have_arc && "Code error.");
+//       shortest_path->AddArc(s, Arc(cur_arc.ilabel, cur_arc.olabel,
+//                                    cur_arc.weight, s+1));
+//     } else { // final-prob.
+//       shortest_path->SetFinal(s, clat.Final(states[s]));
+//     }
+//   }
+// }
+//
+//
+// void ExpandCompactLattice(const CompactLattice &clat,
+//                           double epsilon,
+//                           CompactLattice *expand_clat) {
+//   using namespace fst;
+//   typedef CompactLattice::Arc Arc;
+//   typedef Arc::Weight Weight;
+//   typedef Arc::StateId StateId;
+//   typedef std::pair<StateId, StateId> StatePair;
+//   typedef unordered_map<StatePair, StateId, PairHasher<StateId> > MapType;
+//   typedef MapType::iterator IterType;
+//
+//   if (clat.Start() == kNoStateId) return;
+//   // Make sure the input lattice is topologically sorted.
+//   if (clat.Properties(kTopSorted, true) == 0) {
+//     CompactLattice clat_copy(clat);
+//     KALDI_LOG << "Topsort this lattice.";
+//     if (!TopSort(&clat_copy))
+//       KALDI_ERR << "Was not able to topologically sort lattice (cycles found?)";
+//     ExpandCompactLattice(clat_copy, epsilon, expand_clat);
+//     return;
+//   }
+//
+//   // Compute backward logprobs betas for the expanded lattice.
+//   // Note: the backward logprobs in the original lattice <clat> and the
+//   // expanded lattice <expand_clat> are the same.
+//   int32 num_states = clat.NumStates();
+//   std::vector<double> beta(num_states, kLogZeroDouble);
+//   ComputeCompactLatticeBetas(clat, &beta);
+//   double tot_backward_logprob = beta[0];
+//   std::vector<double> alpha;
+//   alpha.push_back(0.0);
+//   expand_clat->DeleteStates();
+//   MapType state_map; // Map from state pair (orig_state, copy_state) to
+//   // copy_state, where orig_state is a state in the original lattice, and
+//   // copy_state is its corresponding one in the expanded lattice.
+//   unordered_map<StateId, StateId> states; // Map from orig_state to its
+//   // copy_state for states with incoming arcs' posteriors <= epsilon.
+//   std::queue<StatePair> state_queue;
+//
+//   // Set start state in the expanded lattice.
+//   StateId start_state = expand_clat->AddState();
+//   expand_clat->SetStart(start_state);
+//   StatePair start_pair(clat.Start(), start_state);
+//   state_queue.push(start_pair);
+//   std::pair<IterType, bool> result =
+//     state_map.insert(std::make_pair(start_pair, start_state));
+//   KALDI_ASSERT(result.second == true);
+//
+//   // Expand <clat> and update forward logprobs alphas in <expand_clat>.
+//   while (!state_queue.empty()) {
+//     StatePair s = state_queue.front();
+//     StateId s1 = s.first,
+//             s2 = s.second;
+//     state_queue.pop();
+//
+//     Weight f = clat.Final(s1);
+//     if (f != Weight::Zero()) {
+//       KALDI_ASSERT(state_map.find(s) != state_map.end());
+//       expand_clat->SetFinal(state_map[s], f);
+//     }
+//
+//     for (ArcIterator<CompactLattice> aiter(clat, s1);
+//          !aiter.Done(); aiter.Next()) {
+//       const Arc &arc = aiter.Value();
+//       StateId orig_state = arc.nextstate;
+//       double arc_like = -ConvertToCost(arc.weight),
+//              this_alpha = alpha[s2] + arc_like,
+//              arc_post = Exp(this_alpha + beta[orig_state] -
+//                             tot_backward_logprob);
+//       // Generate the expanded lattice.
+//       StateId copy_state;
+//       if (arc_post > epsilon) {
+//         copy_state = expand_clat->AddState();
+//         StatePair next_pair(orig_state, copy_state);
+//         std::pair<IterType, bool> result =
+//           state_map.insert(std::make_pair(next_pair, copy_state));
+//         KALDI_ASSERT(result.second == true);
+//         state_queue.push(next_pair);
+//       } else {
+//         unordered_map<StateId, StateId>::iterator iter = states.find(orig_state);
+//         if (iter == states.end() ) { // The counterpart state of orig_state
+//                                    // has not been created in <expand_clat> yet.
+//           copy_state = expand_clat->AddState();
+//           StatePair next_pair(orig_state, copy_state);
+//           std::pair<IterType, bool> result =
+//             state_map.insert(std::make_pair(next_pair, copy_state));
+//           KALDI_ASSERT(result.second == true);
+//           state_queue.push(next_pair);
+//           states[orig_state] = copy_state;
+//         } else {
+//           copy_state = iter->second;
+//         }
+//       }
+//       // Create an arc from state_map[s] to copy_state in the expanded lattice.
+//       expand_clat->AddArc(state_map[s], Arc(arc.ilabel, arc.olabel, arc.weight,
+//                                             copy_state));
+//       // Compute forward logprobs alpha for the expanded lattice.
+//       if ((alpha.size() - 1) < copy_state) { // The first time to compute alpha
+//                                              // for copy_state in <expand_clat>.
+//         alpha.push_back(this_alpha);
+//       } else { // Accumulate alpha.
+//         alpha[copy_state] = LogAdd(alpha[copy_state], this_alpha);
+//       }
+//     }
+//   } // end while
+// }
+//
+//
+// void CompactLatticeBestCostsAndTracebacks(
+//     const CompactLattice &clat,
+//     CostTraceType *forward_best_cost_and_pred,
+//     CostTraceType *backward_best_cost_and_pred) {
+//
+//   // typedef the arc, weight types
+//   typedef CompactLatticeArc Arc;
+//   typedef Arc::Weight Weight;
+//   typedef Arc::StateId StateId;
+//
+//   forward_best_cost_and_pred->clear();
+//   backward_best_cost_and_pred->clear();
+//   forward_best_cost_and_pred->resize(clat.NumStates());
+//   backward_best_cost_and_pred->resize(clat.NumStates());
+//   // Initialize the cost and predecessor state for each state.
+//   for (StateId s = 0; s < clat.NumStates(); s++) {
+//     (*forward_best_cost_and_pred)[s].first =
+//                                         std::numeric_limits<double>::infinity();
+//     (*backward_best_cost_and_pred)[s].first =
+//                                         std::numeric_limits<double>::infinity();
+//     (*forward_best_cost_and_pred)[s].second = fst::kNoStateId;
+//     (*backward_best_cost_and_pred)[s].second = fst::kNoStateId;
+//   }
+//
+//   StateId start_state = clat.Start();
+//   (*forward_best_cost_and_pred)[start_state].first = 0;
+//   // Transverse the lattice forwardly to compute the best cost from the start
+//   // state to each state and the best predecessor state of each state.
+//   for (StateId s = 0; s < clat.NumStates(); s++) {
+//     double cur_cost = (*forward_best_cost_and_pred)[s].first;
+//     for (fst::ArcIterator<CompactLattice> aiter(clat, s);
+//          !aiter.Done(); aiter.Next()) {
+//       const Arc &arc = aiter.Value();
+//       double next_cost = cur_cost + ConvertToCost(arc.weight);
+//       if (next_cost < (*forward_best_cost_and_pred)[arc.nextstate].first) {
+//         (*forward_best_cost_and_pred)[arc.nextstate].first = next_cost;
+//         (*forward_best_cost_and_pred)[arc.nextstate].second = s;
+//       }
+//     }
+//   }
+//   // Transverse the lattice backwardly to compute the best cost from a final
+//   // state to each state and the best predecessor state of each state.
+//   for (StateId s = clat.NumStates() - 1; s >= 0; s--) {
+//     double this_cost = ConvertToCost(clat.Final(s));
+//     for (fst::ArcIterator<CompactLattice> aiter(clat, s);
+//          !aiter.Done(); aiter.Next()) {
+//       const Arc &arc = aiter.Value();
+//       double next_cost = (*backward_best_cost_and_pred)[arc.nextstate].first +
+//         ConvertToCost(arc.weight);
+//       if (next_cost < this_cost) {
+//         this_cost = next_cost;
+//         (*backward_best_cost_and_pred)[s].second = arc.nextstate;
+//       }
+//     }
+//     (*backward_best_cost_and_pred)[s].first = this_cost;
+//   }
+// }
+//
+//
+// void AddNnlmScoreToCompactLattice(const MapT &nnlm_scores,
+//                                   CompactLattice *clat) {
+//   if (clat->Start() == fst::kNoStateId) return;
+//   // Make sure the input lattice is topologically sorted.
+//   if (clat->Properties(fst::kTopSorted, true) == 0) {
+//     KALDI_LOG << "Topsort this lattice.";
+//     if (!TopSort(clat))
+//       KALDI_ERR << "Was not able to topologically sort lattice (cycles found?)";
+//     AddNnlmScoreToCompactLattice(nnlm_scores, clat);
+//     return;
+//   }
+//
+//   // typedef the arc, weight types
+//   typedef CompactLatticeArc Arc;
+//   typedef Arc::Weight Weight;
+//   typedef Arc::StateId StateId;
+//   typedef std::pair<int32, int32> StatePair;
+//
+//   int32 num_states = clat->NumStates();
+//   unordered_map<StatePair, bool, PairHasher<int32> > final_state_check;
+//   for (StateId s = 0; s < num_states; s++) {
+//     for (fst::MutableArcIterator<CompactLattice> aiter(clat, s);
+//          !aiter.Done(); aiter.Next()) {
+//       Arc arc(aiter.Value());
+//       StatePair arc_index = std::make_pair(static_cast<int32>(s),
+//                                            static_cast<int32>(arc.nextstate));
+//       MapT::const_iterator it = nnlm_scores.find(arc_index);
+//       double nnlm_score;
+//       if (it != nnlm_scores.end())
+//         nnlm_score = it->second;
+//       else
+//         KALDI_ERR << "Some arc does not have neural language model score.";
+//       if (arc.ilabel != 0) { // if there is a word on this arc
+//         LatticeWeight weight = arc.weight.Weight();
+//         // Add associated neural LM score to each arc.
+//         weight.SetValue1(weight.Value1() + nnlm_score);
+//         arc.weight.SetWeight(weight);
+//         aiter.SetValue(arc);
+//       }
+//       Weight clat_final = clat->Final(arc.nextstate);
+//       StatePair final_pair = std::make_pair(arc.nextstate, arc.nextstate);
+//       // Add neural LM scores to each final state only once.
+//       if (clat_final != CompactLatticeWeight::Zero() &&
+//           final_state_check.find(final_pair) == final_state_check.end()) {
+//         MapT::const_iterator final_it = nnlm_scores.find(final_pair);
+//         double final_nnlm_score = 0.0;
+//         if (final_it != nnlm_scores.end())
+//           final_nnlm_score = final_it->second;
+//         // Add neural LM scores to the final weight.
+//         Weight final_weight(LatticeWeight(clat_final.Weight().Value1() +
+//                                           final_nnlm_score,
+//                                           clat_final.Weight().Value2()),
+//                                           clat_final.String());
+//         clat->SetFinal(arc.nextstate, final_weight);
+//         final_state_check[final_pair] = true;
+//       }
+//     } // end looping over arcs
+//   } // end looping over states
+// }
+//
+// void AddWordInsPenToCompactLattice(BaseFloat word_ins_penalty,
+//                                    CompactLattice *clat) {
+//   typedef CompactLatticeArc Arc;
+//   int32 num_states = clat->NumStates();
+//
+//   //scan the lattice
+//   for (int32 state = 0; state < num_states; state++) {
+//     for (fst::MutableArcIterator<CompactLattice> aiter(clat, state);
+//          !aiter.Done(); aiter.Next()) {
+//
+//       Arc arc(aiter.Value());
+//
+//       if (arc.ilabel != 0) { // if there is a word on this arc
+//         LatticeWeight weight = arc.weight.Weight();
+//         // add word insertion penalty to lattice
+//         weight.SetValue1( weight.Value1() + word_ins_penalty);
+//         arc.weight.SetWeight(weight);
+//         aiter.SetValue(arc);
+//       }
+//     } // end looping over arcs
+//   }  // end looping over states
+// }
+//
+// struct ClatRescoreTuple {
+//   ClatRescoreTuple(int32 state, int32 arc, int32 tid):
+//       state_id(state), arc_id(arc), tid(tid) { }
+//   int32 state_id;
+//   int32 arc_id;
+//   int32 tid;
+// };
+//
+// /** RescoreCompactLatticeInternal is the internal code for both
+//     RescoreCompactLattice and RescoreCompatLatticeSpeedup.  For
+//     RescoreCompactLattice, "tmodel" will be NULL and speedup_factor will be 1.0.
+//  */
+// bool RescoreCompactLatticeInternal(
+//     const TransitionModel *tmodel,
+//     BaseFloat speedup_factor,
+//     DecodableInterface *decodable,
+//     CompactLattice *clat) {
+//   KALDI_ASSERT(speedup_factor >= 1.0);
+//   if (clat->NumStates() == 0) {
+//     KALDI_WARN << "Rescoring empty lattice";
+//     return false;
+//   }
+//   if (!clat->Properties(fst::kTopSorted, true)) {
+//     if (fst::TopSort(clat) == false) {
+//       KALDI_WARN << "Cycles detected in lattice.";
+//       return false;
+//     }
+//   }
+//   std::vector<int32> state_times;
+//   int32 utt_len = kaldi::CompactLatticeStateTimes(*clat, &state_times);
+//
+//   std::vector<std::vector<ClatRescoreTuple> > time_to_state(utt_len);
+//
+//   int32 num_states = clat->NumStates();
+//   KALDI_ASSERT(num_states == state_times.size());
+//   for (size_t state = 0; state < num_states; state++) {
+//     KALDI_ASSERT(state_times[state] >= 0);
+//     int32 t = state_times[state];
+//     int32 arc_id = 0;
+//     for (fst::MutableArcIterator<CompactLattice> aiter(clat, state);
+//          !aiter.Done(); aiter.Next(), arc_id++) {
+//       CompactLatticeArc arc = aiter.Value();
+//       std::vector<int32> arc_string = arc.weight.String();
+//
+//       for (size_t offset = 0; offset < arc_string.size(); offset++) {
+//         if (t < utt_len) { // end state may be past this..
+//           int32 tid = arc_string[offset];
+//           time_to_state[t+offset].push_back(ClatRescoreTuple(state, arc_id, tid));
+//         } else {
+//           if (t != utt_len) {
+//             KALDI_WARN << "There appears to be lattice/feature mismatch, "
+//                        << "aborting.";
+//             return false;
+//           }
+//         }
+//       }
+//     }
+//     if (clat->Final(state) != CompactLatticeWeight::Zero()) {
+//       arc_id = -1;
+//       std::vector<int32> arc_string = clat->Final(state).String();
+//       for (size_t offset = 0; offset < arc_string.size(); offset++) {
+//         KALDI_ASSERT(t + offset < utt_len); // already checked in
+//         // CompactLatticeStateTimes, so would be code error.
+//         time_to_state[t+offset].push_back(
+//             ClatRescoreTuple(state, arc_id, arc_string[offset]));
+//       }
+//     }
+//   }
+//
+//   for (int32 t = 0; t < utt_len; t++) {
+//     if ((t < utt_len - 1) && decodable->IsLastFrame(t)) {
+//       KALDI_WARN << "Features are too short for lattice: utt-len is "
+//                  << utt_len << ", " << t << " is last frame";
+//       return false;
+//     }
+//     // frame_scale is the scale we put on the computed acoustic probs for this
+//     // frame.  It will always be 1.0 if tmodel == NULL (i.e. if we are not doing
+//     // the "speedup" code).  For frames with multiple pdf-ids it will be one.
+//     // For frames with only one pdf-id, it will equal speedup_factor (>=1.0)
+//     // with probability 1.0 / speedup_factor, and zero otherwise.  If it is zero,
+//     // we can avoid computing the probabilities.
+//     BaseFloat frame_scale = 1.0;
+//     KALDI_ASSERT(!time_to_state[t].empty());
+//     if (tmodel != NULL) {
+//       int32 pdf_id = tmodel->TransitionIdToPdf(time_to_state[t][0].tid);
+//       bool frame_has_multiple_pdfs = false;
+//       for (size_t i = 1; i < time_to_state[t].size(); i++) {
+//         if (tmodel->TransitionIdToPdf(time_to_state[t][i].tid) != pdf_id) {
+//           frame_has_multiple_pdfs = true;
+//           break;
+//         }
+//       }
+//       if (frame_has_multiple_pdfs) {
+//         frame_scale = 1.0;
+//       } else {
+//         if (WithProb(1.0 / speedup_factor)) {
+//           frame_scale = speedup_factor;
+//         } else {
+//           frame_scale = 0.0;
+//         }
+//       }
+//       if (frame_scale == 0.0)
+//         continue; // the code below would be pointless.
+//     }
+//
+//     for (size_t i = 0; i < time_to_state[t].size(); i++) {
+//       int32 state = time_to_state[t][i].state_id;
+//       int32 arc_id = time_to_state[t][i].arc_id;
+//       int32 tid = time_to_state[t][i].tid;
+//
+//       if (arc_id == -1) { // Final state
+//         // Access the trans_id
+//         CompactLatticeWeight curr_clat_weight = clat->Final(state);
+//
+//         // Calculate likelihood
+//         BaseFloat log_like = decodable->LogLikelihood(t, tid) * frame_scale;
+//         // update weight
+//         CompactLatticeWeight new_clat_weight = curr_clat_weight;
+//         LatticeWeight new_lat_weight = new_clat_weight.Weight();
+//         new_lat_weight.SetValue2(-log_like + curr_clat_weight.Weight().Value2());
+//         new_clat_weight.SetWeight(new_lat_weight);
+//         clat->SetFinal(state, new_clat_weight);
+//       } else {
+//         fst::MutableArcIterator<CompactLattice> aiter(clat, state);
+//
+//         aiter.Seek(arc_id);
+//         CompactLatticeArc arc = aiter.Value();
+//
+//         // Calculate likelihood
+//         BaseFloat log_like = decodable->LogLikelihood(t, tid) * frame_scale;
+//         // update weight
+//         LatticeWeight new_weight = arc.weight.Weight();
+//         new_weight.SetValue2(-log_like + arc.weight.Weight().Value2());
+//         arc.weight.SetWeight(new_weight);
+//         aiter.SetValue(arc);
+//       }
+//     }
+//   }
+//   return true;
+// }
+//
+//
+// bool RescoreCompactLatticeSpeedup(
+//     const TransitionModel &tmodel,
+//     BaseFloat speedup_factor,
+//     DecodableInterface *decodable,
+//     CompactLattice *clat) {
+//   return RescoreCompactLatticeInternal(&tmodel, speedup_factor, decodable, clat);
+// }
+//
+// bool RescoreCompactLattice(DecodableInterface *decodable,
+//                            CompactLattice *clat) {
+//   return RescoreCompactLatticeInternal(NULL, 1.0, decodable, clat);
+// }
+//
+//
+// bool RescoreLattice(DecodableInterface *decodable,
+//                     Lattice *lat) {
+//   if (lat->NumStates() == 0) {
+//     KALDI_WARN << "Rescoring empty lattice";
+//     return false;
+//   }
+//   if (!lat->Properties(fst::kTopSorted, true)) {
+//     if (fst::TopSort(lat) == false) {
+//       KALDI_WARN << "Cycles detected in lattice.";
+//       return false;
+//     }
+//   }
+//   std::vector<int32> state_times;
+//   int32 utt_len = kaldi::LatticeStateTimes(*lat, &state_times);
+//
+//   std::vector<std::vector<int32> > time_to_state(utt_len );
+//
+//   int32 num_states = lat->NumStates();
+//   KALDI_ASSERT(num_states == state_times.size());
+//   for (size_t state = 0; state < num_states; state++) {
+//     int32 t = state_times[state];
+//     // Don't check t >= 0 because non-accessible states could have t = -1.
+//     KALDI_ASSERT(t <= utt_len);
+//     if (t >= 0 && t < utt_len)
+//       time_to_state[t].push_back(state);
+//   }
+//
+//   for (int32 t = 0; t < utt_len; t++) {
+//     if ((t < utt_len - 1) && decodable->IsLastFrame(t)) {
+//       KALDI_WARN << "Features are too short for lattice: utt-len is "
+//                  << utt_len << ", " << t << " is last frame";
+//       return false;
+//     }
+//     for (size_t i = 0; i < time_to_state[t].size(); i++) {
+//       int32 state = time_to_state[t][i];
+//       for (fst::MutableArcIterator<Lattice> aiter(lat, state);
+//            !aiter.Done(); aiter.Next()) {
+//         LatticeArc arc = aiter.Value();
+//         if (arc.ilabel != 0) {
+//           int32 trans_id = arc.ilabel; // Note: it doesn't necessarily
+//           // have to be a transition-id, just whatever the Decodable
+//           // object is expecting, but it's normally a transition-id.
+//
+//           BaseFloat log_like = decodable->LogLikelihood(t, trans_id);
+//           arc.weight.SetValue2(-log_like + arc.weight.Value2());
+//           aiter.SetValue(arc);
+//         }
+//       }
+//     }
+//   }
+//   return true;
+// }
+//
+//
+// BaseFloat LatticeForwardBackwardMmi(
+//     const TransitionModel &tmodel,
+//     const Lattice &lat,
+//     const std::vector<int32> &num_ali,
+//     bool drop_frames,
+//     bool convert_to_pdf_ids,
+//     bool cancel,
+//     Posterior *post) {
+//   // First compute the MMI posteriors.
+//
+//   Posterior den_post;
+//   BaseFloat ans = LatticeForwardBackward(lat,
+//                                          &den_post,
+//                                          NULL);
+//
+//   Posterior num_post;
+//   AlignmentToPosterior(num_ali, &num_post);
+//
+//   // Now negate the MMI posteriors and add the numerator
+//   // posteriors.
+//   ScalePosterior(-1.0, &den_post);
+//
+//   if (convert_to_pdf_ids) {
+//     Posterior num_tmp;
+//     ConvertPosteriorToPdfs(tmodel, num_post, &num_tmp);
+//     num_tmp.swap(num_post);
+//     Posterior den_tmp;
+//     ConvertPosteriorToPdfs(tmodel, den_post, &den_tmp);
+//     den_tmp.swap(den_post);
+//   }
+//
+//   MergePosteriors(num_post, den_post,
+//                   cancel, drop_frames, post);
+//
+//   return ans;
+// }
+//
+//
+// int32 LongestSentenceLength(const Lattice &lat) {
+//   typedef Lattice::Arc Arc;
+//   typedef Arc::Label Label;
+//   typedef Arc::StateId StateId;
+//
+//   if (lat.Properties(fst::kTopSorted, true) == 0) {
+//     Lattice lat_copy(lat);
+//     if (!TopSort(&lat_copy))
+//       KALDI_ERR << "Was not able to topologically sort lattice (cycles found?)";
+//     return LongestSentenceLength(lat_copy);
+//   }
+//   std::vector<int32> max_length(lat.NumStates(), 0);
+//   int32 lattice_max_length = 0;
+//   for (StateId s = 0; s < lat.NumStates(); s++) {
+//     int32 this_max_length = max_length[s];
+//     for (fst::ArcIterator<Lattice> aiter(lat, s); !aiter.Done(); aiter.Next()) {
+//       const Arc &arc = aiter.Value();
+//       bool arc_has_word = (arc.olabel != 0);
+//       StateId nextstate = arc.nextstate;
+//       KALDI_ASSERT(static_cast<size_t>(nextstate) < max_length.size());
+//       if (arc_has_word) {
+//         // A lattice should ideally not have cycles anyway; a cycle with a word
+//         // on is something very bad.
+//         KALDI_ASSERT(nextstate > s && "Lattice has cycles with words on.");
+//         max_length[nextstate] = std::max(max_length[nextstate],
+//                                          this_max_length + 1);
+//       } else {
+//         max_length[nextstate] = std::max(max_length[nextstate],
+//                                          this_max_length);
+//       }
+//     }
+//     if (lat.Final(s) != LatticeWeight::Zero())
+//       lattice_max_length = std::max(lattice_max_length, max_length[s]);
+//   }
+//   return lattice_max_length;
+// }
+//
+// int32 LongestSentenceLength(const CompactLattice &clat) {
+//   typedef CompactLattice::Arc Arc;
+//   typedef Arc::Label Label;
+//   typedef Arc::StateId StateId;
+//
+//   if (clat.Properties(fst::kTopSorted, true) == 0) {
+//     CompactLattice clat_copy(clat);
+//     if (!TopSort(&clat_copy))
+//       KALDI_ERR << "Was not able to topologically sort lattice (cycles found?)";
+//     return LongestSentenceLength(clat_copy);
+//   }
+//   std::vector<int32> max_length(clat.NumStates(), 0);
+//   int32 lattice_max_length = 0;
+//   for (StateId s = 0; s < clat.NumStates(); s++) {
+//     int32 this_max_length = max_length[s];
+//     for (fst::ArcIterator<CompactLattice> aiter(clat, s);
+//          !aiter.Done(); aiter.Next()) {
+//       const Arc &arc = aiter.Value();
+//       bool arc_has_word = (arc.ilabel != 0); // note: olabel == ilabel.
+//       // also note: for normal CompactLattice, e.g. as produced by
+//       // determinization, all arcs will have nonzero labels, but the user might
+//       // decide to remplace some of the labels with zero for some reason, and we
+//       // want to support this.
+//       StateId nextstate = arc.nextstate;
+//       KALDI_ASSERT(static_cast<size_t>(nextstate) < max_length.size());
+//       KALDI_ASSERT(nextstate > s && "CompactLattice has cycles");
+//       if (arc_has_word)
+//         max_length[nextstate] = std::max(max_length[nextstate],
+//                                          this_max_length + 1);
+//       else
+//         max_length[nextstate] = std::max(max_length[nextstate],
+//                                          this_max_length);
+//     }
+//     if (clat.Final(s) != CompactLatticeWeight::Zero())
+//       lattice_max_length = std::max(lattice_max_length, max_length[s]);
+//   }
+//   return lattice_max_length;
+// }
+//
+// void ComposeCompactLatticeDeterministic(
+//     const CompactLattice& clat,
+//     fst::DeterministicOnDemandFst<fst::StdArc>* det_fst,
+//     CompactLattice* composed_clat) {
+//   // StdFst::Arc and CompactLatticeArc has the same StateId type.
+//   typedef fst::StdArc::StateId StateId;
+//   typedef fst::StdArc::Weight Weight1;
+//   typedef CompactLatticeArc::Weight Weight2;
+//   typedef std::pair<StateId, StateId> StatePair;
+//   typedef unordered_map<StatePair, StateId, PairHasher<StateId> > MapType;
+//   typedef MapType::iterator IterType;
+//
+//   // Empties the output FST.
+//   KALDI_ASSERT(composed_clat != NULL);
+//   composed_clat->DeleteStates();
+//
+//   MapType state_map;
+//   std::queue<StatePair> state_queue;
+//
+//   // Sets start state in <composed_clat>.
+//   StateId start_state = composed_clat->AddState();
+//   StatePair start_pair(clat.Start(), det_fst->Start());
+//   composed_clat->SetStart(start_state);
+//   state_queue.push(start_pair);
+//   std::pair<IterType, bool> result =
+//       state_map.insert(std::make_pair(start_pair, start_state));
+//   KALDI_ASSERT(result.second == true);
+//
+//   // Starts composition here.
+//   while (!state_queue.empty()) {
+//     // Gets the first state in the queue.
+//     StatePair s = state_queue.front();
+//     StateId s1 = s.first;
+//     StateId s2 = s.second;
+//     state_queue.pop();
+//
+//
+//     Weight2 clat_final = clat.Final(s1);
+//     if (clat_final.Weight().Value1() !=
+//         std::numeric_limits<BaseFloat>::infinity()) {
+//       // Test for whether the final-prob of state s1 was zero.
+//       Weight1 det_fst_final = det_fst->Final(s2);
+//       if (det_fst_final.Value() !=
+//           std::numeric_limits<BaseFloat>::infinity()) {
+//         // Test for whether the final-prob of state s2 was zero.  If neither
+//         // source-state final prob was zero, then we should create final state
+//         // in fst_composed. We compute the product manually since this is more
+//         // efficient.
+//         Weight2 final_weight(LatticeWeight(clat_final.Weight().Value1() +
+//                                            det_fst_final.Value(),
+//                                            clat_final.Weight().Value2()),
+//                              clat_final.String());
+//         // we can assume final_weight is not Zero(), since neither of
+//         // the sources was zero.
+//         KALDI_ASSERT(state_map.find(s) != state_map.end());
+//         composed_clat->SetFinal(state_map[s], final_weight);
+//       }
+//     }
+//
+//     // Loops over pair of edges at s1 and s2.
+//     for (fst::ArcIterator<CompactLattice> aiter(clat, s1);
+//          !aiter.Done(); aiter.Next()) {
+//       const CompactLatticeArc& arc1 = aiter.Value();
+//       fst::StdArc arc2;
+//       StateId next_state1 = arc1.nextstate, next_state2;
+//       bool matched = false;
+//
+//       if (arc1.olabel == 0) {
+//         // If the symbol on <arc1> is <epsilon>, we transit to the next state
+//         // for <clat>, but keep <det_fst> at the current state.
+//         matched = true;
+//         next_state2 = s2;
+//       } else {
+//         // Otherwise try to find the matched arc in <det_fst>.
+//         matched = det_fst->GetArc(s2, arc1.olabel, &arc2);
+//         if (matched) {
+//           next_state2 = arc2.nextstate;
+//         }
+//       }
+//
+//       // If matched arc is found in <det_fst>, then we have to add new arcs to
+//       // <composed_clat>.
+//       if (matched) {
+//         StatePair next_state_pair(next_state1, next_state2);
+//         IterType siter = state_map.find(next_state_pair);
+//         StateId next_state;
+//
+//         // Adds composed state to <state_map>.
+//         if (siter == state_map.end()) {
+//           // If the composed state has not been created yet, create it.
+//           next_state = composed_clat->AddState();
+//           std::pair<const StatePair, StateId> next_state_map(next_state_pair,
+//                                                              next_state);
+//           std::pair<IterType, bool> result = state_map.insert(next_state_map);
+//           KALDI_ASSERT(result.second);
+//           state_queue.push(next_state_pair);
+//         } else {
+//           // If the composed state is already in <state_map>, we can directly
+//           // use that.
+//           next_state = siter->second;
+//         }
+//
+//         // Adds arc to <composed_clat>.
+//         if (arc1.olabel == 0) {
+//           composed_clat->AddArc(state_map[s],
+//                                 CompactLatticeArc(arc1.ilabel, 0,
+//                                                   arc1.weight, next_state));
+//         } else {
+//           Weight2 composed_weight(
+//               LatticeWeight(arc1.weight.Weight().Value1() +
+//                             arc2.weight.Value(),
+//                             arc1.weight.Weight().Value2()),
+//               arc1.weight.String());
+//           composed_clat->AddArc(state_map[s],
+//                                 CompactLatticeArc(arc1.ilabel, arc2.olabel,
+//                                                   composed_weight, next_state));
+//         }
+//       }
+//     }
+//   }
+//   fst::Connect(composed_clat);
+// }
+//
+//
+// void ComputeAcousticScoresMap(
+//     const Lattice &lat,
+//     unordered_map<std::pair<int32, int32>, std::pair<BaseFloat, int32>,
+//                                         PairHasher<int32> > *acoustic_scores) {
+//   // typedef the arc, weight types
+//   typedef Lattice::Arc Arc;
+//   typedef Arc::Weight LatticeWeight;
+//   typedef Arc::StateId StateId;
+//
+//   acoustic_scores->clear();
+//
+//   std::vector<int32> state_times;
+//   LatticeStateTimes(lat, &state_times);   // Assumes the input is top sorted
+//
+//   KALDI_ASSERT(lat.Start() == 0);
+//
+//   for (StateId s = 0; s < lat.NumStates(); s++) {
+//     int32 t = state_times[s];
+//     for (fst::ArcIterator<Lattice> aiter(lat, s); !aiter.Done();
+//           aiter.Next()) {
+//       const Arc &arc = aiter.Value();
+//       const LatticeWeight &weight = arc.weight;
+//
+//       int32 tid = arc.ilabel;
+//
+//       if (tid != 0) {
+//         unordered_map<std::pair<int32, int32>, std::pair<BaseFloat, int32>,
+//           PairHasher<int32> >::iterator it = acoustic_scores->find(std::make_pair(t, tid));
+//         if (it == acoustic_scores->end()) {
+//           acoustic_scores->insert(std::make_pair(std::make_pair(t, tid),
+//                                           std::make_pair(weight.Value2(), 1)));
+//         } else {
+//           if (it->second.second == 2
+//                 && it->second.first / it->second.second != weight.Value2()) {
+//             KALDI_VLOG(2) << "Transitions on the same frame have different "
+//                           << "acoustic costs for tid " << tid << "; "
+//                           << it->second.first / it->second.second
+//                           << " vs " << weight.Value2();
+//           }
+//           it->second.first += weight.Value2();
+//           it->second.second++;
+//         }
+//       } else {
+//         // Arcs with epsilon input label (tid) must have 0 acoustic cost
+//         KALDI_ASSERT(weight.Value2() == 0);
+//       }
+//     }
+//
+//     LatticeWeight f = lat.Final(s);
+//     if (f != LatticeWeight::Zero()) {
+//       // Final acoustic cost must be 0 as we are reading from
+//       // non-determinized, non-compact lattice
+//       KALDI_ASSERT(f.Value2() == 0.0);
+//     }
+//   }
+// }
+//
+// void ReplaceAcousticScoresFromMap(
+//     const unordered_map<std::pair<int32, int32>, std::pair<BaseFloat, int32>,
+//                                         PairHasher<int32> > &acoustic_scores,
+//     Lattice *lat) {
+//   // typedef the arc, weight types
+//   typedef Lattice::Arc Arc;
+//   typedef Arc::Weight LatticeWeight;
+//   typedef Arc::StateId StateId;
+//
+//   TopSortLatticeIfNeeded(lat);
+//
+//   std::vector<int32> state_times;
+//   LatticeStateTimes(*lat, &state_times);
+//
+//   KALDI_ASSERT(lat->Start() == 0);
+//
+//   for (StateId s = 0; s < lat->NumStates(); s++) {
+//     int32 t = state_times[s];
+//     for (fst::MutableArcIterator<Lattice> aiter(lat, s);
+//          !aiter.Done(); aiter.Next()) {
+//       Arc arc(aiter.Value());
+//
+//       int32 tid = arc.ilabel;
+//       if (tid != 0) {
+//         unordered_map<std::pair<int32, int32>, std::pair<BaseFloat, int32>,
+//           PairHasher<int32> >::const_iterator it = acoustic_scores.find(std::make_pair(t, tid));
+//         if (it == acoustic_scores.end()) {
+//           KALDI_ERR << "Could not find tid " << tid << " at time " << t
+//                     << " in the acoustic scores map.";
+//         } else {
+//           arc.weight.SetValue2(it->second.first / it->second.second);
+//         }
+//       } else {
+//         // For epsilon arcs, set acoustic cost to 0.0
+//         arc.weight.SetValue2(0.0);
+//       }
+//       aiter.SetValue(arc);
+//     }
+//
+//     LatticeWeight f = lat->Final(s);
+//     if (f != LatticeWeight::Zero()) {
+//       // Set final acoustic cost to 0.0
+//       f.SetValue2(0.0);
+//       lat->SetFinal(s, f);
+//     }
+//   }
+// }
 
 }  // namespace kaldi
diff --git a/speechx/speechx/kaldi/lat/lattice-functions.h b/speechx/speechx/kaldi/lat/lattice-functions.h
index 2388dd98..6b1b6656 100644
--- a/speechx/speechx/kaldi/lat/lattice-functions.h
+++ b/speechx/speechx/kaldi/lat/lattice-functions.h
@@ -28,374 +28,427 @@
 #include <map>
 
 #include "base/kaldi-common.h"
+// #include "hmm/posterior.h"
 #include "fstext/fstext-lib.h"
-#include "itf/decodable-itf.h"
-#include "itf/transition-information.h"
+// #include "hmm/transition-model.h"
 #include "lat/kaldi-lattice.h"
+// #include "itf/decodable-itf.h"
 
 namespace kaldi {
 
-// Redundant with the typedef in hmm/posterior.h. We want functions
-// using the Posterior type to be usable without a dependency on the
-// hmm library.
-typedef std::vector<std::vector<std::pair<int32, BaseFloat> > > Posterior;
-
-/**
-   This function extracts the per-frame log likelihoods from a linear
-   lattice (which we refer to as an 'nbest' lattice elsewhere in Kaldi code).
-   The dimension of *per_frame_loglikes will be set to the
-   number of input symbols in 'nbest'.  The elements of
-   '*per_frame_loglikes' will be set to the .Value2() elements of the lattice
-   weights, which represent the acoustic costs; you may want to scale this
-   vector afterward by -1/acoustic_scale to get the original loglikes.
-   If there are acoustic costs on input-epsilon arcs or the final-prob in 'nbest'
-   (and this should not normally be the case in situations where it makes
-   sense to call this function), they will be included to the cost of the
-   preceding input symbol, or the following input symbol for input-epsilons
-   encountered prior to any input symbol.  If 'nbest' has no input symbols,
-   'per_frame_loglikes' will be set to the empty vector.
-**/
-void GetPerFrameAcousticCosts(const Lattice &nbest,
-                              Vector<BaseFloat> *per_frame_loglikes);
-
-/// This function iterates over the states of a topologically sorted lattice and
-/// counts the time instance corresponding to each state. The times are returned
-/// in a vector of integers 'times' which is resized to have a size equal to the
-/// number of states in the lattice. The function also returns the maximum time
-/// in the lattice (this will equal the number of frames in the file).
-int32 LatticeStateTimes(const Lattice &lat, std::vector<int32> *times);
-
-/// As LatticeStateTimes, but in the CompactLattice format.  Note: must
-/// be topologically sorted.  Returns length of the utterance in frames, which
-/// might not be the same as the maximum time in the lattice, due to frames
-/// in the final-prob.
-int32 CompactLatticeStateTimes(const CompactLattice &clat,
-                               std::vector<int32> *times);
-
-/// This function does the forward-backward over lattices and computes the
-/// posterior probabilities of the arcs. It returns the total log-probability
-/// of the lattice.  The Posterior quantities contain pairs of (transition-id, weight)
-/// on each frame.
-/// If the pointer "acoustic_like_sum" is provided, this value is set to
-/// the sum over the arcs, of the posterior of the arc times the
-/// acoustic likelihood [i.e. negated acoustic score] on that link.
-/// This is used in combination with other quantities to work out
-/// the objective function in MMI discriminative training.
-BaseFloat LatticeForwardBackward(const Lattice &lat,
-                                 Posterior *arc_post,
-                                 double *acoustic_like_sum = NULL);
-
-// This function is something similar to LatticeForwardBackward(), but it is on
-// the CompactLattice lattice format. Also we only need the alpha in the forward
-// path, not the posteriors.
-bool ComputeCompactLatticeAlphas(const CompactLattice &lat,
-                                 std::vector<double> *alpha);
-
-// A sibling of the function CompactLatticeAlphas()... We compute the beta from
-// the backward path here.
-bool ComputeCompactLatticeBetas(const CompactLattice &lat,
-                                std::vector<double> *beta);
-
-
-// Computes (normal or Viterbi) alphas and betas; returns (total-prob, or
-// best-path negated cost) Note: in either case, the alphas and betas are
-// negated costs.  Requires that lat be topologically sorted.  This code
-// will work for either CompactLattice or Lattice.
-template<typename LatticeType>
-double ComputeLatticeAlphasAndBetas(const LatticeType &lat,
-                                    bool viterbi,
-                                    std::vector<double> *alpha,
-                                    std::vector<double> *beta);
-
-
-/// Topologically sort the compact lattice if not already topologically sorted.
-/// Will crash if the lattice cannot be topologically sorted.
-void TopSortCompactLatticeIfNeeded(CompactLattice *clat);
-
-
-/// Topologically sort the lattice if not already topologically sorted.
-/// Will crash if lattice cannot be topologically sorted.
-void TopSortLatticeIfNeeded(Lattice *clat);
-
-/// Returns the depth of the lattice, defined as the average number of arcs (or
-/// final-prob strings) crossing any given frame.  Returns 1 for empty lattices.
-/// Requires that clat is topologically sorted!
-BaseFloat CompactLatticeDepth(const CompactLattice &clat,
-                              int32 *num_frames = NULL);
-
-/// This function returns, for each frame, the number of arcs crossing that
-/// frame.
-void CompactLatticeDepthPerFrame(const CompactLattice &clat,
-                                 std::vector<int32> *depth_per_frame);
-
-
-/// This function limits the depth of the lattice, per frame: that means, it
-/// does not allow more than a specified number of arcs active on any given
-/// frame.  This can be used to reduce the size of the "very deep" portions of
-/// the lattice.
-void CompactLatticeLimitDepth(int32 max_arcs_per_frame,
-                              CompactLattice *clat);
-
-
-/// Given a lattice, and a transition model to map pdf-ids to phones,
-/// outputs for each frame the set of phones active on that frame.  If
-/// sil_phones (which must be sorted and uniq) is nonempty, it excludes
-/// phones in this list.
-void LatticeActivePhones(const Lattice &lat, const TransitionInformation &trans,
-                         const std::vector<int32> &sil_phones,
-                         std::vector<std::set<int32> > *active_phones);
-
-/// Given a lattice, and a transition model to map pdf-ids to phones,
-/// replace the output symbols (presumably words), with phones; we
-/// use the TransitionModel to work out the phone sequence.  Note
-/// that the phone labels are not exactly aligned with the phone
-/// boundaries.  We put a phone label to coincide with any transition
-/// to the final, nonemitting state of a phone (this state always exists,
-/// we ensure this in HmmTopology::Check()).  This would be the last
-/// transition-id in the phone if reordering is not done (but typically
-/// we do reorder).
-/// Also see PhoneAlignLattice, in phone-align-lattice.h.
-void ConvertLatticeToPhones(const TransitionInformation &trans_model,
-                            Lattice *lat);
+// /**
+//    This function extracts the per-frame log likelihoods from a linear
+//    lattice (which we refer to as an 'nbest' lattice elsewhere in Kaldi code).
+//    The dimension of *per_frame_loglikes will be set to the
+//    number of input symbols in 'nbest'.  The elements of
+//    '*per_frame_loglikes' will be set to the .Value2() elements of the lattice
+//    weights, which represent the acoustic costs; you may want to scale this
+//    vector afterward by -1/acoustic_scale to get the original loglikes.
+//    If there are acoustic costs on input-epsilon arcs or the final-prob in 'nbest'
+//    (and this should not normally be the case in situations where it makes
+//    sense to call this function), they will be included to the cost of the
+//    preceding input symbol, or the following input symbol for input-epsilons
+//    encountered prior to any input symbol.  If 'nbest' has no input symbols,
+//    'per_frame_loglikes' will be set to the empty vector.
+// **/
+// void GetPerFrameAcousticCosts(const Lattice &nbest,
+//                               Vector<BaseFloat> *per_frame_loglikes);
+//
+// /// This function iterates over the states of a topologically sorted lattice and
+// /// counts the time instance corresponding to each state. The times are returned
+// /// in a vector of integers 'times' which is resized to have a size equal to the
+// /// number of states in the lattice. The function also returns the maximum time
+// /// in the lattice (this will equal the number of frames in the file).
+// int32 LatticeStateTimes(const Lattice &lat, std::vector<int32> *times);
+//
+// /// As LatticeStateTimes, but in the CompactLattice format.  Note: must
+// /// be topologically sorted.  Returns length of the utterance in frames, which
+// /// might not be the same as the maximum time in the lattice, due to frames
+// /// in the final-prob.
+// int32 CompactLatticeStateTimes(const CompactLattice &clat,
+//                                std::vector<int32> *times);
+//
+// /// This function does the forward-backward over lattices and computes the
+// /// posterior probabilities of the arcs. It returns the total log-probability
+// /// of the lattice.  The Posterior quantities contain pairs of (transition-id, weight)
+// /// on each frame.
+// /// If the pointer "acoustic_like_sum" is provided, this value is set to
+// /// the sum over the arcs, of the posterior of the arc times the
+// /// acoustic likelihood [i.e. negated acoustic score] on that link.
+// /// This is used in combination with other quantities to work out
+// /// the objective function in MMI discriminative training.
+// BaseFloat LatticeForwardBackward(const Lattice &lat,
+//                                  Posterior *arc_post,
+//                                  double *acoustic_like_sum = NULL);
+//
+// // This function is something similar to LatticeForwardBackward(), but it is on
+// // the CompactLattice lattice format. Also we only need the alpha in the forward
+// // path, not the posteriors.
+// bool ComputeCompactLatticeAlphas(const CompactLattice &lat,
+//                                  std::vector<double> *alpha);
+//
+// // A sibling of the function CompactLatticeAlphas()... We compute the beta from
+// // the backward path here.
+// bool ComputeCompactLatticeBetas(const CompactLattice &lat,
+//                                 std::vector<double> *beta);
+//
+//
+// // Computes (normal or Viterbi) alphas and betas; returns (total-prob, or
+// // best-path negated cost) Note: in either case, the alphas and betas are
+// // negated costs.  Requires that lat be topologically sorted.  This code
+// // will work for either CompactLattice or Latice.
+// template<typename LatticeType>
+// double ComputeLatticeAlphasAndBetas(const LatticeType &lat,
+//                                     bool viterbi,
+//                                     std::vector<double> *alpha,
+//                                     std::vector<double> *beta);
+//
+//
+// /// Topologically sort the compact lattice if not already topologically sorted.
+// /// Will crash if the lattice cannot be topologically sorted.
+// void TopSortCompactLatticeIfNeeded(CompactLattice *clat);
+//
+//
+// /// Topologically sort the lattice if not already topologically sorted.
+// /// Will crash if lattice cannot be topologically sorted.
+// void TopSortLatticeIfNeeded(Lattice *clat);
+//
+// /// Returns the depth of the lattice, defined as the average number of arcs (or
+// /// final-prob strings) crossing any given frame.  Returns 1 for empty lattices.
+// /// Requires that clat is topologically sorted!
+// BaseFloat CompactLatticeDepth(const CompactLattice &clat,
+//                               int32 *num_frames = NULL);
+//
+// /// This function returns, for each frame, the number of arcs crossing that
+// /// frame.
+// void CompactLatticeDepthPerFrame(const CompactLattice &clat,
+//                                  std::vector<int32> *depth_per_frame);
+//
+//
+// /// This function limits the depth of the lattice, per frame: that means, it
+// /// does not allow more than a specified number of arcs active on any given
+// /// frame.  This can be used to reduce the size of the "very deep" portions of
+// /// the lattice.
+// void CompactLatticeLimitDepth(int32 max_arcs_per_frame,
+//                               CompactLattice *clat);
+//
+//
+// /// Given a lattice, and a transition model to map pdf-ids to phones,
+// /// outputs for each frame the set of phones active on that frame.  If
+// /// sil_phones (which must be sorted and uniq) is nonempty, it excludes
+// /// phones in this list.
+// void LatticeActivePhones(const Lattice &lat, const TransitionModel &trans,
+//                          const std::vector<int32> &sil_phones,
+//                          std::vector<std::set<int32> > *active_phones);
+//
+// /// Given a lattice, and a transition model to map pdf-ids to phones,
+// /// replace the output symbols (presumably words), with phones; we
+// /// use the TransitionModel to work out the phone sequence.  Note
+// /// that the phone labels are not exactly aligned with the phone
+// /// boundaries.  We put a phone label to coincide with any transition
+// /// to the final, nonemitting state of a phone (this state always exists,
+// /// we ensure this in HmmTopology::Check()).  This would be the last
+// /// transition-id in the phone if reordering is not done (but typically
+// /// we do reorder).
+// /// Also see PhoneAlignLattice, in phone-align-lattice.h.
+// void ConvertLatticeToPhones(const TransitionModel &trans_model,
+//                             Lattice *lat);
 
 /// Prunes a lattice or compact lattice.  Returns true on success, false if
 /// there was some kind of failure.
 template<class LatticeType>
 bool PruneLattice(BaseFloat beam, LatticeType *lat);
 
-
-/// Given a lattice, and a transition model to map pdf-ids to phones,
-/// replace the sequences of transition-ids with sequences of phones.
-/// Note that this is different from ConvertLatticeToPhones, in that
-/// we replace the transition-ids not the words.
-void ConvertCompactLatticeToPhones(const TransitionInformation &trans_model,
-                                   CompactLattice *clat);
-
-/// Boosts LM probabilities by b * [number of frame errors]; equivalently, adds
-/// -b*[number of frame errors] to the graph-component of the cost of each arc/path.
-/// There is a frame error if a particular transition-id on a particular frame
-/// corresponds to a phone not matching transcription's alignment for that frame.
-/// This is used in "margin-inspired" discriminative training, esp. Boosted MMI.
-/// The TransitionInformation is used to map transition-ids in the lattice
-/// input-side to phones; the phones appearing in
-/// "silence_phones" are treated specially in that we replace the frame error f
-/// (either zero or 1) for a frame, with the minimum of f or max_silence_error.
-/// For the normal recipe, max_silence_error would be zero.
-/// Returns true on success, false if there was some kind of mismatch.
-/// At input, silence_phones must be sorted and unique.
-bool LatticeBoost(const TransitionInformation &trans,
-                  const std::vector<int32> &alignment,
-                  const std::vector<int32> &silence_phones,
-                  BaseFloat b,
-                  BaseFloat max_silence_error,
-                  Lattice *lat);
-
-
-/**
-   This function implements either the MPFE (minimum phone frame error) or SMBR
-   (state-level minimum bayes risk) forward-backward, depending on whether
-   "criterion" is "mpfe" or "smbr".  It returns the MPFE
-   criterion of SMBR criterion for this utterance, and outputs the posteriors (which
-   may be positive or negative) into "post".
-
-   @param [in] trans    The transition model. Used to map the
-                        transition-ids to phones or pdfs.
-   @param [in] silence_phones   A list of integer ids of silence phones. The
-                        silence frames i.e. the frames where num_ali
-                        corresponds to a silence phones are treated specially.
-                        The behavior is determined by 'one_silence_class'
-                        being false (traditional behavior) or true.
-                        Usually in our setup, several phones including
-                        the silence, vocalized noise, non-spoken noise
-                        and unk are treated as "silence phones"
-   @param [in] lat      The denominator lattice
-   @param [in] num_ali  The numerator alignment
-   @param [in] criterion    The objective function. Must be "mpfe" or "smbr"
-                        for MPFE (minimum phone frame error) or sMBR
-                        (state minimum bayes risk) training.
-   @param [in] one_silence_class   Determines how the silence frames are treated.
-                        Setting this to false gives the old traditional behavior,
-                        where the silence frames (according to num_ali) are
-                        treated as incorrect. However, this means that the
-                        insertions are not penalized by the objective.
-                        Setting this to true gives the new behaviour, where we
-                        treat silence as any other phone, except that all pdfs
-                        of silence phones are collapsed into a single class for
-                        the frame-error computation. This can possible reduce
-                        the insertions in the trained model. This is closer to
-                        the WER metric that we actually care about, since WER is
-                        generally computed after filtering out noises, but
-                        does penalize insertions.
-    @param [out] post   The "MBR posteriors" i.e. derivatives w.r.t to the
-                        pseudo log-likelihoods of states at each frame.
-*/
-BaseFloat LatticeForwardBackwardMpeVariants(
-    const TransitionInformation &trans,
-    const std::vector<int32> &silence_phones,
-    const Lattice &lat,
-    const std::vector<int32> &num_ali,
-    std::string criterion,
-    bool one_silence_class,
-    Posterior *post);
-
-/// This function takes a CompactLattice that should only contain a single
-/// linear sequence (e.g. derived from lattice-1best), and that should have been
-/// processed so that the arcs in the CompactLattice align correctly with the
-/// word boundaries (e.g. by lattice-align-words).  It outputs 3 vectors of the
-/// same size, which give, for each word in the lattice (in sequence), the word
-/// label and the begin time and length in frames.  This is done even for zero
-/// (epsilon) words, generally corresponding to optional silence-- if you don't
-/// want them, just ignore them in the output.
-/// This function will print a warning and return false, if the lattice
-/// did not have the correct format (e.g. if it is empty or it is not
-/// linear).
-bool CompactLatticeToWordAlignment(const CompactLattice &clat,
-                                   std::vector<int32> *words,
-                                   std::vector<int32> *begin_times,
-                                   std::vector<int32> *lengths);
-
-/// A form of the shortest-path/best-path algorithm that's specially coded for
-/// CompactLattice.  Requires that clat be acyclic.
-void CompactLatticeShortestPath(const CompactLattice &clat,
-                                CompactLattice *shortest_path);
-
-/// This function expands a CompactLattice to ensure high-probability paths
-/// have unique histories. Arcs with posteriors larger than epsilon get splitted.
-void ExpandCompactLattice(const CompactLattice &clat,
-                          double epsilon,
-                          CompactLattice *expand_clat);
-
-/// For each state, compute forward and backward best (viterbi) costs and its
-/// traceback states (for generating best paths later). The forward best cost
-/// for a state is the cost of the best path from the start state to the state.
-/// The traceback state of this state is its predecessor state in the best path.
-/// The backward best cost for a state is the cost of the best path from the
-/// state to a final one. Its traceback state is the successor state in the best
-/// path in the forward direction.
-/// Note: final weights of states are in backward_best_cost_and_pred.
-/// Requires the input CompactLattice clat be acyclic.
-typedef std::vector<std::pair<double,
-        CompactLatticeArc::StateId> > CostTraceType;
-void CompactLatticeBestCostsAndTracebacks(
-    const CompactLattice &clat,
-    CostTraceType *forward_best_cost_and_pred,
-    CostTraceType *backward_best_cost_and_pred);
-
-/// This function adds estimated neural language model scores of words in a
-/// minimal list of hypotheses that covers a lattice, to the graph scores on the
-/// arcs. The list of hypotheses are generated by latbin/lattice-path-cover.
-typedef unordered_map<std::pair<int32, int32>, double, PairHasher<int32> > MapT;
-void AddNnlmScoreToCompactLattice(const MapT &nnlm_scores,
-                                  CompactLattice *clat);
-
-/// This function add the word insertion penalty to graph score of each word
-/// in the compact lattice
-void AddWordInsPenToCompactLattice(BaseFloat word_ins_penalty,
-                                   CompactLattice *clat);
-
-/// This function *adds* the negated scores obtained from the Decodable object,
-/// to the acoustic scores on the arcs.  If you want to replace them, you should
-/// use ScaleCompactLattice to first set the acoustic scores to zero.  Returns
-/// true on success, false on error (typically some kind of mismatched inputs).
-bool RescoreCompactLattice(DecodableInterface *decodable,
-                           CompactLattice *clat);
-
-
-/// This function returns the number of words in the longest sentence in a
-/// CompactLattice (i.e. the the maximum of any path, of the count of
-/// olabels on that path).
-int32 LongestSentenceLength(const Lattice &lat);
-
-/// This function returns the number of words in the longest sentence in a
-/// CompactLattice, i.e. the the maximum of any path, of the count of
-/// labels on that path... note, in CompactLattice, the ilabels and olabels
-/// are identical because it is an acceptor.
-int32 LongestSentenceLength(const CompactLattice &lat);
-
-
-/// This function is like RescoreCompactLattice, but it is modified to avoid
-/// computing probabilities on most frames where all the pdf-ids are the same.
-/// (it needs the transition-model to work out whether two transition-ids map to
-/// the same pdf-id, and it assumes that the lattice has transition-ids on it).
-/// The naive thing would be to just set all probabilities to zero on frames
-/// where all the pdf-ids are the same (because this value won't affect the
-/// lattice posterior).  But this would become confusing when we compute
-/// corpus-level diagnostics such as the MMI objective function.  Instead,
-/// imagine speedup_factor = 100 (it must be >= 1.0)... with probability (1.0 /
-/// speedup_factor) we compute those likelihoods and multiply them by
-/// speedup_factor; otherwise we set them to zero.  This gives the right
-/// expected probability so our corpus-level diagnostics will be about right.
-bool RescoreCompactLatticeSpeedup(
-    const TransitionInformation &tmodel,
-    BaseFloat speedup_factor,
-    DecodableInterface *decodable,
-    CompactLattice *clat);
-
-
-/// This function *adds* the negated scores obtained from the Decodable object,
-/// to the acoustic scores on the arcs.  If you want to replace them, you should
-/// use ScaleCompactLattice to first set the acoustic scores to zero.  Returns
-/// true on success, false on error (e.g. some kind of mismatched inputs).
-/// The input labels, if nonzero, are interpreted as transition-ids or whatever
-/// other index the Decodable object expects.
-bool RescoreLattice(DecodableInterface *decodable,
-                    Lattice *lat);
-
-/// This function Composes a CompactLattice format lattice with a
-/// DeterministicOnDemandFst<fst::StdFst> format fst, and outputs another
-/// CompactLattice format lattice. The first element (the one that corresponds
-/// to LM weight) in CompactLatticeWeight is used for composition.
-///
-/// Note that the DeterministicOnDemandFst interface is not "const", therefore
-/// we cannot use "const" for <det_fst>.
-void ComposeCompactLatticeDeterministic(
-    const CompactLattice& clat,
-    fst::DeterministicOnDemandFst<fst::StdArc>* det_fst,
-    CompactLattice* composed_clat);
-
-/// This function computes the mapping from the pair 
-/// (frame-index, transition-id) to the pair 
-/// (sum-of-acoustic-scores, num-of-occurences) over all occurences of the 
-/// transition-id in that frame.
-/// frame-index in the lattice. 
-/// This function is useful for retaining the acoustic scores in a 
-/// non-compact lattice after a process like determinization where the 
-/// frame-level acoustic scores are typically lost.
-/// The function ReplaceAcousticScoresFromMap is used to restore the 
-/// acoustic scores computed by this function.
-///
-///   @param [in] lat   Input lattice. Expected to be top-sorted. Otherwise the 
-///                     function will crash. 
-///   @param [out] acoustic_scores  
-///                     Pointer to a map from the pair (frame-index,
-///                     transition-id) to a pair (sum-of-acoustic-scores,
-///                     num-of-occurences).
-///                     Usually the acoustic scores for a pdf-id (and hence
-///                     transition-id) on a frame will be the same for all the
-///                     occurences of the pdf-id in that frame. 
-///                     But if not, we will take the average of the acoustic
-///                     scores. Hence, we store both the sum-of-acoustic-scores
-///                     and the num-of-occurences of the transition-id in that
-///                     frame.
-void ComputeAcousticScoresMap(
-    const Lattice &lat,
-    unordered_map<std::pair<int32, int32>, std::pair<BaseFloat, int32>,
-                                        PairHasher<int32> > *acoustic_scores);
-
-/// This function restores acoustic scores computed using the function
-/// ComputeAcousticScoresMap into the lattice.
-///
-///   @param [in] acoustic_scores  
-///                      A map from the pair (frame-index, transition-id) to a
-///                      pair (sum-of-acoustic-scores, num-of-occurences) of 
-///                      the occurences of the transition-id in that frame.
-///                      See the comments for ComputeAcousticScoresMap for 
-///                      details.
-///   @param [out] lat   Pointer to the output lattice.
-void ReplaceAcousticScoresFromMap(
-    const unordered_map<std::pair<int32, int32>, std::pair<BaseFloat, int32>,
-                                        PairHasher<int32> > &acoustic_scores,
-    Lattice *lat);
+//
+// /// Given a lattice, and a transition model to map pdf-ids to phones,
+// /// replace the sequences of transition-ids with sequences of phones.
+// /// Note that this is different from ConvertLatticeToPhones, in that
+// /// we replace the transition-ids not the words.
+// void ConvertCompactLatticeToPhones(const TransitionModel &trans_model,
+//                                    CompactLattice *clat);
+//
+// /// Boosts LM probabilities by b * [number of frame errors]; equivalently, adds
+// /// -b*[number of frame errors] to the graph-component of the cost of each arc/path.
+// /// There is a frame error if a particular transition-id on a particular frame
+// /// corresponds to a phone not matching transcription's alignment for that frame.
+// /// This is used in "margin-inspired" discriminative training, esp. Boosted MMI.
+// /// The TransitionModel is used to map transition-ids in the lattice
+// /// input-side to phones; the phones appearing in
+// /// "silence_phones" are treated specially in that we replace the frame error f
+// /// (either zero or 1) for a frame, with the minimum of f or max_silence_error.
+// /// For the normal recipe, max_silence_error would be zero.
+// /// Returns true on success, false if there was some kind of mismatch.
+// /// At input, silence_phones must be sorted and unique.
+// bool LatticeBoost(const TransitionModel &trans,
+//                   const std::vector<int32> &alignment,
+//                   const std::vector<int32> &silence_phones,
+//                   BaseFloat b,
+//                   BaseFloat max_silence_error,
+//                   Lattice *lat);
+//
+//
+// /**
+//    This function implements either the MPFE (minimum phone frame error) or SMBR
+//    (state-level minimum bayes risk) forward-backward, depending on whether
+//    "criterion" is "mpfe" or "smbr".  It returns the MPFE
+//    criterion of SMBR criterion for this utterance, and outputs the posteriors (which
+//    may be positive or negative) into "post".
+//
+//    @param [in] trans    The transition model. Used to map the
+//                         transition-ids to phones or pdfs.
+//    @param [in] silence_phones   A list of integer ids of silence phones. The
+//                         silence frames i.e. the frames where num_ali
+//                         corresponds to a silence phones are treated specially.
+//                         The behavior is determined by 'one_silence_class'
+//                         being false (traditional behavior) or true.
+//                         Usually in our setup, several phones including
+//                         the silence, vocalized noise, non-spoken noise
+//                         and unk are treated as "silence phones"
+//    @param [in] lat      The denominator lattice
+//    @param [in] num_ali  The numerator alignment
+//    @param [in] criterion    The objective function. Must be "mpfe" or "smbr"
+//                         for MPFE (minimum phone frame error) or sMBR
+//                         (state minimum bayes risk) training.
+//    @param [in] one_silence_class   Determines how the silence frames are treated.
+//                         Setting this to false gives the old traditional behavior,
+//                         where the silence frames (according to num_ali) are
+//                         treated as incorrect. However, this means that the
+//                         insertions are not penalized by the objective.
+//                         Setting this to true gives the new behaviour, where we
+//                         treat silence as any other phone, except that all pdfs
+//                         of silence phones are collapsed into a single class for
+//                         the frame-error computation. This can possible reduce
+//                         the insertions in the trained model. This is closer to
+//                         the WER metric that we actually care about, since WER is
+//                         generally computed after filtering out noises, but
+//                         does penalize insertions.
+//     @param [out] post   The "MBR posteriors" i.e. derivatives w.r.t to the
+//                         pseudo log-likelihoods of states at each frame.
+// */
+// BaseFloat LatticeForwardBackwardMpeVariants(
+//     const TransitionModel &trans,
+//     const std::vector<int32> &silence_phones,
+//     const Lattice &lat,
+//     const std::vector<int32> &num_ali,
+//     std::string criterion,
+//     bool one_silence_class,
+//     Posterior *post);
+//
+// /**
+//    This function can be used to compute posteriors for MMI, with a positive contribution
+//    for the numerator and a negative one for the denominator.  This function is not actually
+//    used in our normal MMI training recipes, where it's instead done using various command
+//    line programs that each do a part of the job.  This function was written for use in
+//    neural-net MMI training.
+//
+//    @param [in] trans    The transition model. Used to map the
+//                         transition-ids to phones or pdfs.
+//    @param [in] lat      The denominator lattice
+//    @param [in] num_ali  The numerator alignment
+//    @param [in] drop_frames   If "drop_frames" is true, it will not compute any
+//                         posteriors on frames where the num and den have disjoint
+//                         pdf-ids.
+//    @param [in] convert_to_pdf_ids   If "convert_to_pdfs_ids" is true, it will
+//                         convert the output to be at the level of pdf-ids, not
+//                         transition-ids.
+//    @param [in] cancel   If "cancel" is true, it will cancel out any positive and
+//                         negative parts from the same transition-id (or pdf-id,
+//                         if convert_to_pdf_ids == true).
+//    @param [out] arc_post   The output MMI posteriors of transition-ids (or
+//                         pdf-ids if convert_to_pdf_ids == true) at each frame
+//                         i.e. the difference between the numerator
+//                         and denominator posteriors.
+//
+//    It returns the forward-backward likelihood of the lattice. */
+// BaseFloat LatticeForwardBackwardMmi(
+//     const TransitionModel &trans,
+//     const Lattice &lat,
+//     const std::vector<int32> &num_ali,
+//     bool drop_frames,
+//     bool convert_to_pdf_ids,
+//     bool cancel,
+//     Posterior *arc_post);
+//
+//
+// /// This function takes a CompactLattice that should only contain a single
+// /// linear sequence (e.g. derived from lattice-1best), and that should have been
+// /// processed so that the arcs in the CompactLattice align correctly with the
+// /// word boundaries (e.g. by lattice-align-words).  It outputs 3 vectors of the
+// /// same size, which give, for each word in the lattice (in sequence), the word
+// /// label and the begin time and length in frames.  This is done even for zero
+// /// (epsilon) words, generally corresponding to optional silence-- if you don't
+// /// want them, just ignore them in the output.
+// /// This function will print a warning and return false, if the lattice
+// /// did not have the correct format (e.g. if it is empty or it is not
+// /// linear).
+// bool CompactLatticeToWordAlignment(const CompactLattice &clat,
+//                                    std::vector<int32> *words,
+//                                    std::vector<int32> *begin_times,
+//                                    std::vector<int32> *lengths);
+//
+// /// This function takes a CompactLattice that should only contain a single
+// /// linear sequence (e.g. derived from lattice-1best), and that should have been
+// /// processed so that the arcs in the CompactLattice align correctly with the
+// /// word boundaries (e.g. by lattice-align-words).  It outputs 4 vectors of the
+// /// same size, which give, for each word in the lattice (in sequence), the word
+// /// label, the begin time and length in frames, and the pronunciation (sequence
+// /// of phones).  This is done even for zero words, corresponding to optional
+// /// silences -- if you don't want them, just ignore them in the output.
+// /// This function will print a warning and return false, if the lattice
+// /// did not have the correct format (e.g. if it is empty or it is not
+// /// linear).
+// bool CompactLatticeToWordProns(
+//     const TransitionModel &tmodel,
+//     const CompactLattice &clat,
+//     std::vector<int32> *words,
+//     std::vector<int32> *begin_times,
+//     std::vector<int32> *lengths,
+//     std::vector<std::vector<int32> > *prons,
+//     std::vector<std::vector<int32> > *phone_lengths);
+//
+//
+// /// A form of the shortest-path/best-path algorithm that's specially coded for
+// /// CompactLattice.  Requires that clat be acyclic.
+// void CompactLatticeShortestPath(const CompactLattice &clat,
+//                                 CompactLattice *shortest_path);
+//
+// /// This function expands a CompactLattice to ensure high-probability paths
+// /// have unique histories. Arcs with posteriors larger than epsilon get splitted.
+// void ExpandCompactLattice(const CompactLattice &clat,
+//                           double epsilon,
+//                           CompactLattice *expand_clat);
+//
+// /// For each state, compute forward and backward best (viterbi) costs and its
+// /// traceback states (for generating best paths later). The forward best cost
+// /// for a state is the cost of the best path from the start state to the state.
+// /// The traceback state of this state is its predecessor state in the best path.
+// /// The backward best cost for a state is the cost of the best path from the
+// /// state to a final one. Its traceback state is the successor state in the best
+// /// path in the forward direction.
+// /// Note: final weights of states are in backward_best_cost_and_pred.
+// /// Requires the input CompactLattice clat be acyclic.
+// typedef std::vector<std::pair<double,
+//         CompactLatticeArc::StateId> > CostTraceType;
+// void CompactLatticeBestCostsAndTracebacks(
+//     const CompactLattice &clat,
+//     CostTraceType *forward_best_cost_and_pred,
+//     CostTraceType *backward_best_cost_and_pred);
+//
+// /// This function adds estimated neural language model scores of words in a
+// /// minimal list of hypotheses that covers a lattice, to the graph scores on the
+// /// arcs. The list of hypotheses are generated by latbin/lattice-path-cover.
+// typedef unordered_map<std::pair<int32, int32>, double, PairHasher<int32> > MapT;
+// void AddNnlmScoreToCompactLattice(const MapT &nnlm_scores,
+//                                   CompactLattice *clat);
+//
+// /// This function add the word insertion penalty to graph score of each word
+// /// in the compact lattice
+// void AddWordInsPenToCompactLattice(BaseFloat word_ins_penalty,
+//                                    CompactLattice *clat);
+//
+// /// This function *adds* the negated scores obtained from the Decodable object,
+// /// to the acoustic scores on the arcs.  If you want to replace them, you should
+// /// use ScaleCompactLattice to first set the acoustic scores to zero.  Returns
+// /// true on success, false on error (typically some kind of mismatched inputs).
+// bool RescoreCompactLattice(DecodableInterface *decodable,
+//                            CompactLattice *clat);
+//
+//
+// /// This function returns the number of words in the longest sentence in a
+// /// CompactLattice (i.e. the the maximum of any path, of the count of
+// /// olabels on that path).
+// int32 LongestSentenceLength(const Lattice &lat);
+//
+// /// This function returns the number of words in the longest sentence in a
+// /// CompactLattice, i.e. the the maximum of any path, of the count of
+// /// labels on that path... note, in CompactLattice, the ilabels and olabels
+// /// are identical because it is an acceptor.
+// int32 LongestSentenceLength(const CompactLattice &lat);
+//
+//
+// /// This function is like RescoreCompactLattice, but it is modified to avoid
+// /// computing probabilities on most frames where all the pdf-ids are the same.
+// /// (it needs the transition-model to work out whether two transition-ids map to
+// /// the same pdf-id, and it assumes that the lattice has transition-ids on it).
+// /// The naive thing would be to just set all probabilities to zero on frames
+// /// where all the pdf-ids are the same (because this value won't affect the
+// /// lattice posterior).  But this would become confusing when we compute
+// /// corpus-level diagnostics such as the MMI objective function.  Instead,
+// /// imagine speedup_factor = 100 (it must be >= 1.0)... with probability (1.0 /
+// /// speedup_factor) we compute those likelihoods and multiply them by
+// /// speedup_factor; otherwise we set them to zero.  This gives the right
+// /// expected probability so our corpus-level diagnostics will be about right.
+// bool RescoreCompactLatticeSpeedup(
+//     const TransitionModel &tmodel,
+//     BaseFloat speedup_factor,
+//     DecodableInterface *decodable,
+//     CompactLattice *clat);
+//
+//
+// /// This function *adds* the negated scores obtained from the Decodable object,
+// /// to the acoustic scores on the arcs.  If you want to replace them, you should
+// /// use ScaleCompactLattice to first set the acoustic scores to zero.  Returns
+// /// true on success, false on error (e.g. some kind of mismatched inputs).
+// /// The input labels, if nonzero, are interpreted as transition-ids or whatever
+// /// other index the Decodable object expects.
+// bool RescoreLattice(DecodableInterface *decodable,
+//                     Lattice *lat);
+//
+// /// This function Composes a CompactLattice format lattice with a
+// /// DeterministicOnDemandFst<fst::StdFst> format fst, and outputs another
+// /// CompactLattice format lattice. The first element (the one that corresponds
+// /// to LM weight) in CompactLatticeWeight is used for composition.
+// ///
+// /// Note that the DeterministicOnDemandFst interface is not "const", therefore
+// /// we cannot use "const" for <det_fst>.
+// void ComposeCompactLatticeDeterministic(
+//     const CompactLattice& clat,
+//     fst::DeterministicOnDemandFst<fst::StdArc>* det_fst,
+//     CompactLattice* composed_clat);
+//
+// /// This function computes the mapping from the pair
+// /// (frame-index, transition-id) to the pair
+// /// (sum-of-acoustic-scores, num-of-occurences) over all occurences of the
+// /// transition-id in that frame.
+// /// frame-index in the lattice.
+// /// This function is useful for retaining the acoustic scores in a
+// /// non-compact lattice after a process like determinization where the
+// /// frame-level acoustic scores are typically lost.
+// /// The function ReplaceAcousticScoresFromMap is used to restore the
+// /// acoustic scores computed by this function.
+// ///
+// ///   @param [in] lat   Input lattice. Expected to be top-sorted. Otherwise the
+// ///                     function will crash.
+// ///   @param [out] acoustic_scores
+// ///                     Pointer to a map from the pair (frame-index,
+// ///                     transition-id) to a pair (sum-of-acoustic-scores,
+// ///                     num-of-occurences).
+// ///                     Usually the acoustic scores for a pdf-id (and hence
+// ///                     transition-id) on a frame will be the same for all the
+// ///                     occurences of the pdf-id in that frame.
+// ///                     But if not, we will take the average of the acoustic
+// ///                     scores. Hence, we store both the sum-of-acoustic-scores
+// ///                     and the num-of-occurences of the transition-id in that
+// ///                     frame.
+// void ComputeAcousticScoresMap(
+//     const Lattice &lat,
+//     unordered_map<std::pair<int32, int32>, std::pair<BaseFloat, int32>,
+//                                         PairHasher<int32> > *acoustic_scores);
+//
+// /// This function restores acoustic scores computed using the function
+// /// ComputeAcousticScoresMap into the lattice.
+// ///
+// ///   @param [in] acoustic_scores
+// ///                      A map from the pair (frame-index, transition-id) to a
+// ///                      pair (sum-of-acoustic-scores, num-of-occurences) of
+// ///                      the occurences of the transition-id in that frame.
+// ///                      See the comments for ComputeAcousticScoresMap for
+// ///                      details.
+// ///   @param [out] lat   Pointer to the output lattice.
+// void ReplaceAcousticScoresFromMap(
+//     const unordered_map<std::pair<int32, int32>, std::pair<BaseFloat, int32>,
+//                                         PairHasher<int32> > &acoustic_scores,
+//     Lattice *lat);
 
 }  // namespace kaldi
 
diff --git a/speechx/speechx/nnet/decodable.cc b/speechx/speechx/nnet/decodable.cc
index 6c0909ca..0a5e0a77 100644
--- a/speechx/speechx/nnet/decodable.cc
+++ b/speechx/speechx/nnet/decodable.cc
@@ -31,17 +31,23 @@ void Decodable::Acceptlikelihood(const Matrix<BaseFloat>& likelihood) {
 
 // Decodable::Init(DecodableConfig config) {
 //}
+int32 Decodable::NumFramesReady() const {
+    return frames_ready_;
+}
 
-bool Decodable::IsLastFrame(int32 frame) const {
+bool Decodable::IsLastFrame(int32 frame) {
     CHECK_LE(frame, frames_ready_);
-    return IsInputFinished() && (frame == frames_ready_ - 1);
+    bool flag = EnsureFrameHaveComputed(frame);
+    return (flag == false) && (frame == frames_ready_ - 1);
 }
 
 int32 Decodable::NumIndices() const { return 0; }
 
 BaseFloat Decodable::LogLikelihood(int32 frame, int32 index) {
     CHECK_LE(index, nnet_cache_.NumCols());
-    return 0;
+    CHECK_LE(frame, frames_ready_);
+    int32 frame_idx = frame - frame_offset_;
+    return nnet_cache_(frame_idx, index);
 }
 
 bool Decodable::EnsureFrameHaveComputed(int32 frame) {
diff --git a/speechx/speechx/nnet/decodable.h b/speechx/speechx/nnet/decodable.h
index 7938b582..9a480d21 100644
--- a/speechx/speechx/nnet/decodable.h
+++ b/speechx/speechx/nnet/decodable.h
@@ -15,7 +15,7 @@
 #include "base/common.h"
 #include "frontend/feature_extractor_interface.h"
 #include "kaldi/matrix/kaldi-matrix.h"
-#include "nnet/decodable-itf.h"
+#include "kaldi/decoder/decodable-itf.h"
 #include "nnet/nnet_interface.h"
 
 namespace ppspeech {
@@ -29,15 +29,17 @@ class Decodable : public kaldi::DecodableInterface {
         const std::shared_ptr<FeatureExtractorInterface>& frontend);
     // void Init(DecodableOpts config);
     virtual kaldi::BaseFloat LogLikelihood(int32 frame, int32 index);
-    virtual bool IsLastFrame(int32 frame) const;
+    virtual bool IsLastFrame(int32 frame);
     virtual int32 NumIndices() const;
     virtual bool FrameLogLikelihood(int32 frame,
                                     std::vector<kaldi::BaseFloat>* likelihood);
+    virtual int32 NumFramesReady() const;
     // for offline test
     void Acceptlikelihood(const kaldi::Matrix<kaldi::BaseFloat>& likelihood);
     void Reset();
     bool IsInputFinished() const { return frontend_->IsFinished(); }
     bool EnsureFrameHaveComputed(int32 frame);
+    
 
   private:
     bool AdvanceChunk();

From 15f434a5d3e20ab9535bc068b7fb60cf2eba655d Mon Sep 17 00:00:00 2001
From: Yang Zhou <goat.zhou@qq.com>
Date: Mon, 28 Mar 2022 19:57:55 +0800
Subject: [PATCH 2/9] add make TLG binary & script

---
 .../aishell/local/aishell_train_lms.sh        |  59 ++++++
 speechx/examples/aishell/mkgraph.sh           |  31 +++
 speechx/examples/aishell/path.sh              |  14 ++
 .../aishell/tools/fst/add_lex_disambig.pl     | 195 ++++++++++++++++++
 .../tools/fst/compile_lexicon_token_fst.sh    |  86 ++++++++
 .../aishell/tools/fst/ctc_token_fst.py        |  24 +++
 .../tools/fst/ctc_token_fst_compact.py        |  21 ++
 .../tools/fst/ctc_token_fst_corrected.py      |  55 +++++
 .../aishell/tools/fst/eps2disambig.pl         |  29 +++
 .../aishell/tools/fst/make_lexicon_fst.pl     | 155 ++++++++++++++
 .../examples/aishell/tools/fst/make_tlg.sh    |  38 ++++
 .../aishell/tools/fst/prepare_dict.py         |  64 ++++++
 .../examples/aishell/tools/fst/remove_oovs.pl |  43 ++++
 .../aishell/tools/fst/rnnt_token_fst.py       |  17 ++
 speechx/examples/aishell/tools/fst/s2eps.pl   |  27 +++
 .../examples/aishell/tools/parse_options.sh   |  97 +++++++++
 .../decoder/offline_wfst_decoder_main.cc      |   4 +
 speechx/speechx/frontend/normalizer.cc        |   2 +-
 speechx/speechx/nnet/decodable.cc             |   4 +-
 speechx/tools/fstbin/fstaddselfloops.cc       | 100 +++++++++
 speechx/tools/fstbin/fstdeterminizestar.cc    | 114 ++++++++++
 speechx/tools/fstbin/fstisstochastic.cc       |  91 ++++++++
 speechx/tools/fstbin/fstminimizeencoded.cc    |  74 +++++++
 speechx/tools/fstbin/fsttablecompose.cc       | 133 ++++++++++++
 speechx/tools/install_srilm.sh                |  97 +++++++++
 speechx/tools/lmbin/CMakeLists.txt            |   5 +
 speechx/tools/lmbin/arpa2fst.cc               | 145 +++++++++++++
 27 files changed, 1721 insertions(+), 3 deletions(-)
 create mode 100755 speechx/examples/aishell/local/aishell_train_lms.sh
 create mode 100644 speechx/examples/aishell/mkgraph.sh
 create mode 100644 speechx/examples/aishell/path.sh
 create mode 100755 speechx/examples/aishell/tools/fst/add_lex_disambig.pl
 create mode 100755 speechx/examples/aishell/tools/fst/compile_lexicon_token_fst.sh
 create mode 100755 speechx/examples/aishell/tools/fst/ctc_token_fst.py
 create mode 100755 speechx/examples/aishell/tools/fst/ctc_token_fst_compact.py
 create mode 100755 speechx/examples/aishell/tools/fst/ctc_token_fst_corrected.py
 create mode 100755 speechx/examples/aishell/tools/fst/eps2disambig.pl
 create mode 100755 speechx/examples/aishell/tools/fst/make_lexicon_fst.pl
 create mode 100755 speechx/examples/aishell/tools/fst/make_tlg.sh
 create mode 100755 speechx/examples/aishell/tools/fst/prepare_dict.py
 create mode 100755 speechx/examples/aishell/tools/fst/remove_oovs.pl
 create mode 100755 speechx/examples/aishell/tools/fst/rnnt_token_fst.py
 create mode 100755 speechx/examples/aishell/tools/fst/s2eps.pl
 create mode 100755 speechx/examples/aishell/tools/parse_options.sh
 create mode 100644 speechx/tools/fstbin/fstaddselfloops.cc
 create mode 100644 speechx/tools/fstbin/fstdeterminizestar.cc
 create mode 100644 speechx/tools/fstbin/fstisstochastic.cc
 create mode 100644 speechx/tools/fstbin/fstminimizeencoded.cc
 create mode 100644 speechx/tools/fstbin/fsttablecompose.cc
 create mode 100755 speechx/tools/install_srilm.sh
 create mode 100644 speechx/tools/lmbin/CMakeLists.txt
 create mode 100644 speechx/tools/lmbin/arpa2fst.cc

diff --git a/speechx/examples/aishell/local/aishell_train_lms.sh b/speechx/examples/aishell/local/aishell_train_lms.sh
new file mode 100755
index 00000000..30ffb797
--- /dev/null
+++ b/speechx/examples/aishell/local/aishell_train_lms.sh
@@ -0,0 +1,59 @@
+#!/bin/bash
+
+
+# To be run from one directory above this script.
+. ./path.sh
+
+text=data/local/lm/text
+lexicon=data/local/dict/lexicon.txt
+
+for f in "$text" "$lexicon"; do
+  [ ! -f $x ] && echo "$0: No such file $f" && exit 1;
+done
+
+# Check SRILM tools
+if ! which ngram-count > /dev/null; then
+    echo "srilm tools are not found, please download it and install it from: "
+    echo "http://www.speech.sri.com/projects/srilm/download.html"
+    echo "Then add the tools to your PATH"
+    exit 1
+fi
+
+# This script takes no arguments.  It assumes you have already run
+# aishell_data_prep.sh.
+# It takes as input the files
+# data/local/lm/text
+# data/local/dict/lexicon.txt
+dir=data/local/lm
+mkdir -p $dir
+
+
+cleantext=$dir/text.no_oov
+
+cat $text | awk -v lex=$lexicon 'BEGIN{while((getline<lex) >0){ seen[$1]=1; } }
+  {for(n=1; n<=NF;n++) {  if (seen[$n]) { printf("%s ", $n); } else {printf("<SPOKEN_NOISE> ");} } printf("\n");}' \
+  > $cleantext || exit 1;
+
+cat $cleantext | awk '{for(n=2;n<=NF;n++) print $n; }' | sort | uniq -c | \
+   sort -nr > $dir/word.counts || exit 1;
+
+# Get counts from acoustic training transcripts, and add  one-count
+# for each word in the lexicon (but not silence, we don't want it
+# in the LM-- we'll add it optionally later).
+cat $cleantext | awk '{for(n=2;n<=NF;n++) print $n; }' | \
+  cat - <(grep -w -v '!SIL' $lexicon | awk '{print $1}') | \
+   sort | uniq -c | sort -nr > $dir/unigram.counts || exit 1;
+
+cat $dir/unigram.counts | awk '{print $2}' | cat - <(echo "<s>"; echo "</s>" ) > $dir/wordlist
+
+heldout_sent=10000 # Don't change this if you want result to be comparable with
+    # kaldi_lm results
+mkdir -p $dir
+cat $cleantext | awk '{for(n=2;n<=NF;n++){ printf $n; if(n<NF) printf " "; else print ""; }}' | \
+  head -$heldout_sent > $dir/heldout
+cat $cleantext | awk '{for(n=2;n<=NF;n++){ printf $n; if(n<NF) printf " "; else print ""; }}' | \
+  tail -n +$heldout_sent > $dir/train
+
+ngram-count -text $dir/train -order 3 -limit-vocab -vocab $dir/wordlist -unk \
+  -map-unk "<UNK>" -kndiscount -interpolate -lm $dir/lm.arpa
+ngram -lm $dir/lm.arpa -ppl $dir/heldout
diff --git a/speechx/examples/aishell/mkgraph.sh b/speechx/examples/aishell/mkgraph.sh
new file mode 100644
index 00000000..f66cd4dc
--- /dev/null
+++ b/speechx/examples/aishell/mkgraph.sh
@@ -0,0 +1,31 @@
+#!/bin/bash
+
+. ./path.sh || exit 1;
+
+. tools/parse_options.sh || exit 1;
+
+data=/mnt/dataset/aishell
+
+# Optionally, you can add LM and test it with runtime.
+dir=./ds2_graph
+dict=$dir/vocab.txt
+if [ ${stage} -le 7 ] && [ ${stop_stage} -ge 7 ]; then
+  # 7.1 Prepare dict
+  unit_file=$dict
+  mkdir -p $dir/local/dict
+  cp $unit_file $dir/local/dict/units.txt
+  tools/fst/prepare_dict.py $unit_file ${data}/resource_aishell/lexicon.txt \
+    $dir/local/dict/lexicon.txt
+  # Train lm
+  lm=$dir/local/lm
+  mkdir -p $lm
+  tools/filter_scp.pl data/train/text \
+    $data/data_aishell/transcript/aishell_transcript_v0.8.txt > $lm/text
+  local/ds2_aishell_train_lms.sh
+  # Build decoding TLG
+  tools/fst/compile_lexicon_token_fst.sh \
+    $dir/local/dict $dir/local/tmp $dir/local/lang
+  tools/fst/make_tlg.sh $dir/local/lm $dir/local/lang $dir/lang_test || exit 1;
+fi
+
+
diff --git a/speechx/examples/aishell/path.sh b/speechx/examples/aishell/path.sh
new file mode 100644
index 00000000..8ab7ee29
--- /dev/null
+++ b/speechx/examples/aishell/path.sh
@@ -0,0 +1,14 @@
+# This contains the locations of binarys build required for running the examples.
+
+SPEECHX_ROOT=$PWD/../..
+SPEECHX_EXAMPLES=$SPEECHX_ROOT/build/examples
+
+SPEECHX_TOOLS=$SPEECHX_ROOT/tools
+TOOLS_BIN=$SPEECHX_TOOLS/valgrind/install/bin
+
+[ -d $SPEECHX_EXAMPLES ] || { echo "Error: 'build/examples' directory not found. please ensure that the project build successfully"; }
+
+export LC_AL=C
+
+SPEECHX_BIN=$SPEECHX_EXAMPLES/feat
+export PATH=$PATH:$SPEECHX_BIN:$TOOLS_BIN
diff --git a/speechx/examples/aishell/tools/fst/add_lex_disambig.pl b/speechx/examples/aishell/tools/fst/add_lex_disambig.pl
new file mode 100755
index 00000000..dd8a25de
--- /dev/null
+++ b/speechx/examples/aishell/tools/fst/add_lex_disambig.pl
@@ -0,0 +1,195 @@
+#!/usr/bin/env perl
+#  Copyright 2010-2011  Microsoft Corporation
+#            2013-2016  Johns Hopkins University (author: Daniel Povey)
+#                 2015  Hainan Xu
+#                 2015  Guoguo Chen
+
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#  http://www.apache.org/licenses/LICENSE-2.0
+#
+# THIS CODE IS PROVIDED *AS IS* BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY
+# KIND, EITHER EXPRESS OR IMPLIED, INCLUDING WITHOUT LIMITATION ANY IMPLIED
+# WARRANTIES OR CONDITIONS OF TITLE, FITNESS FOR A PARTICULAR PURPOSE,
+# MERCHANTABLITY OR NON-INFRINGEMENT.
+# See the Apache 2 License for the specific language governing permissions and
+# limitations under the License.
+
+
+# Adds disambiguation symbols to a lexicon.
+# Outputs still in the normal lexicon format.
+# Disambig syms are numbered #1, #2, #3, etc. (#0
+# reserved for symbol in grammar).
+# Outputs the number of disambig syms to the standard output.
+# With the --pron-probs option, expects the second field
+# of each lexicon line to be a pron-prob.
+# With the --sil-probs option, expects three additional
+# fields after the pron-prob, representing various components
+# of the silence probability model.
+
+$pron_probs = 0;
+$sil_probs = 0;
+$first_allowed_disambig = 1;
+
+for ($n = 1; $n <= 3 && @ARGV > 0; $n++) {
+  if ($ARGV[0] eq "--pron-probs") {
+    $pron_probs = 1;
+    shift @ARGV;
+  }
+  if ($ARGV[0] eq "--sil-probs") {
+    $sil_probs = 1;
+    shift @ARGV;
+  }
+  if ($ARGV[0] eq "--first-allowed-disambig") {
+    $first_allowed_disambig = 0 + $ARGV[1];
+    if ($first_allowed_disambig < 1) {
+      die "add_lex_disambig.pl: invalid --first-allowed-disambig option: $first_allowed_disambig\n";
+    }
+    shift @ARGV;
+    shift @ARGV;
+  }
+}
+
+if (@ARGV != 2) {
+  die "Usage: add_lex_disambig.pl [opts] <lexicon-in> <lexicon-out>\n" .
+    "This script adds disambiguation symbols to a lexicon in order to\n" .
+    "make decoding graphs determinizable; it adds pseudo-phone\n" .
+    "disambiguation symbols #1, #2 and so on at the ends of phones\n" .
+    "to ensure that all pronunciations are different, and that none\n" .
+    "is a prefix of another.\n" .
+    "It prints to the standard output the number of the largest-numbered" .
+    "disambiguation symbol that was used.\n" .
+    "\n" .
+    "Options:   --pron-probs       Expect pronunciation probabilities in the 2nd field\n" .
+    "           --sil-probs        [should be with --pron-probs option]\n" .
+    "                              Expect 3 extra fields after the pron-probs, for aspects of\n" .
+    "                              the silence probability model\n" .
+    "           --first-allowed-disambig <n>  The number of the first disambiguation symbol\n" .
+    "                              that this script is allowed to add.  By default this is\n" .
+    "                              #1, but you can set this to a larger value using this option.\n" .
+    "e.g.:\n" .
+    " add_lex_disambig.pl lexicon.txt lexicon_disambig.txt\n" .
+    " add_lex_disambig.pl --pron-probs lexiconp.txt lexiconp_disambig.txt\n" .
+    " add_lex_disambig.pl --pron-probs --sil-probs lexiconp_silprob.txt lexiconp_silprob_disambig.txt\n";
+}
+
+
+$lexfn = shift @ARGV;
+$lexoutfn = shift @ARGV;
+
+open(L, "<$lexfn") || die "Error opening lexicon $lexfn";
+
+# (1)  Read in the lexicon.
+@L = ( );
+while(<L>) {
+    @A = split(" ", $_);
+    push @L, join(" ", @A);
+}
+
+# (2) Work out the count of each phone-sequence in the
+# lexicon.
+
+foreach $l (@L) {
+    @A = split(" ", $l);
+    shift @A; # Remove word.
+    if ($pron_probs) {
+      $p = shift @A;
+      if (!($p > 0.0 && $p <= 1.0)) { die "Bad lexicon line $l (expecting pron-prob as second field)"; }
+    }
+    if ($sil_probs) {
+      $silp = shift @A;
+      if (!($silp > 0.0 && $silp <= 1.0)) { die "Bad lexicon line $l for silprobs"; }
+      $correction = shift @A;
+      if ($correction <= 0.0) { die "Bad lexicon line $l for silprobs"; }
+      $correction = shift @A;
+      if ($correction <= 0.0) { die "Bad lexicon line $l for silprobs"; }
+    }
+    if (!(@A)) {
+      die "Bad lexicon line $1, no phone in phone list";
+    }
+    $count{join(" ",@A)}++;
+}
+
+# (3) For each left sub-sequence of each phone-sequence, note down
+# that it exists (for identifying prefixes of longer strings).
+
+foreach $l (@L) {
+    @A = split(" ", $l);
+    shift @A; # Remove word.
+    if ($pron_probs) { shift @A; } # remove pron-prob.
+    if ($sil_probs) {
+      shift @A; # Remove silprob
+      shift @A; # Remove silprob
+    }
+    while(@A > 0) {
+        pop @A;  # Remove last phone
+        $issubseq{join(" ",@A)} = 1;
+    }
+}
+
+# (4) For each entry in the lexicon:
+#  if the phone sequence is unique and is not a
+#  prefix of another word, no diambig symbol.
+#  Else output #1, or #2, #3, ... if the same phone-seq
+#  has already been assigned a disambig symbol.
+
+
+open(O, ">$lexoutfn") || die "Opening lexicon file $lexoutfn for writing.\n";
+
+# max_disambig will always be the highest-numbered disambiguation symbol that
+# has been used so far.
+$max_disambig = $first_allowed_disambig - 1;
+
+foreach $l (@L) {
+  @A = split(" ", $l);
+  $word = shift @A;
+  if ($pron_probs) {
+    $pron_prob = shift @A;
+  }
+  if ($sil_probs) {
+    $sil_word_prob = shift @A;
+    $word_sil_correction = shift @A;
+    $prev_nonsil_correction = shift @A
+  }
+  $phnseq = join(" ", @A);
+  if (!defined $issubseq{$phnseq}
+      && $count{$phnseq} == 1) {
+    ;                           # Do nothing.
+  } else {
+    if ($phnseq eq "") {        # need disambig symbols for the empty string
+      # that are not use anywhere else.
+      $max_disambig++;
+      $reserved_for_the_empty_string{$max_disambig} = 1;
+      $phnseq = "#$max_disambig";
+    } else {
+      $cur_disambig = $last_used_disambig_symbol_of{$phnseq};
+      if (!defined $cur_disambig) {
+        $cur_disambig = $first_allowed_disambig;
+      } else {
+        $cur_disambig++;           # Get a number that has not been used yet for
+                                   # this phone sequence.
+      }
+      while (defined $reserved_for_the_empty_string{$cur_disambig}) {
+        $cur_disambig++;
+      }
+      if ($cur_disambig > $max_disambig) {
+        $max_disambig = $cur_disambig;
+      }
+      $last_used_disambig_symbol_of{$phnseq} = $cur_disambig;
+      $phnseq = $phnseq . " #" . $cur_disambig;
+    }
+  }
+  if ($pron_probs) {
+    if ($sil_probs) {
+      print O "$word\t$pron_prob\t$sil_word_prob\t$word_sil_correction\t$prev_nonsil_correction\t$phnseq\n";
+    } else {
+      print O "$word\t$pron_prob\t$phnseq\n";
+    }
+  } else {
+    print O "$word\t$phnseq\n";
+  }
+}
+
+print $max_disambig . "\n";
diff --git a/speechx/examples/aishell/tools/fst/compile_lexicon_token_fst.sh b/speechx/examples/aishell/tools/fst/compile_lexicon_token_fst.sh
new file mode 100755
index 00000000..fda971fe
--- /dev/null
+++ b/speechx/examples/aishell/tools/fst/compile_lexicon_token_fst.sh
@@ -0,0 +1,86 @@
+#!/bin/bash
+# Copyright 2015       Yajie Miao    (Carnegie Mellon University)
+
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#  http://www.apache.org/licenses/LICENSE-2.0
+#
+# THIS CODE IS PROVIDED *AS IS* BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY
+# KIND, EITHER EXPRESS OR IMPLIED, INCLUDING WITHOUT LIMITATION ANY IMPLIED
+# WARRANTIES OR CONDITIONS OF TITLE, FITNESS FOR A PARTICULAR PURPOSE,
+# MERCHANTABLITY OR NON-INFRINGEMENT.
+# See the Apache 2 License for the specific language governing permissions and
+# limitations under the License.
+
+# This script compiles the lexicon and CTC tokens into FSTs. FST compiling slightly differs between the
+# phoneme and character-based lexicons.
+set -eo pipefail
+. tools/parse_options.sh
+
+if [ $# -ne 3 ]; then
+  echo "usage: tools/fst/compile_lexicon_token_fst.sh <dict-src-dir> <tmp-dir> <lang-dir>"
+  echo "e.g.: tools/fst/compile_lexicon_token_fst.sh data/local/dict data/local/lang_tmp data/lang"
+  echo "<dict-src-dir> should contain the following files:"
+  echo "lexicon.txt lexicon_numbers.txt units.txt"
+  echo "options: "
+  exit 1;
+fi
+
+srcdir=$1
+tmpdir=$2
+dir=$3
+mkdir -p $dir $tmpdir
+
+[ -f path.sh ] && . ./path.sh
+
+cp $srcdir/units.txt $dir
+
+# Add probabilities to lexicon entries. There is in fact no point of doing this here since all the entries have 1.0.
+# But utils/make_lexicon_fst.pl requires a probabilistic version, so we just leave it as it is.
+perl -ape 's/(\S+\s+)(.+)/${1}1.0\t$2/;' < $srcdir/lexicon.txt > $tmpdir/lexiconp.txt || exit 1;
+
+# Add disambiguation symbols to the lexicon. This is necessary for determinizing the composition of L.fst and G.fst.
+# Without these symbols, determinization will fail.
+ndisambig=`tools/fst/add_lex_disambig.pl $tmpdir/lexiconp.txt $tmpdir/lexiconp_disambig.txt`
+ndisambig=$[$ndisambig+1];
+
+( for n in `seq 0 $ndisambig`; do echo '#'$n; done ) > $tmpdir/disambig.list
+
+# Get the full list of CTC tokens used in FST. These tokens include <eps>, the blank <blk>,
+# the actual model unit, and the disambiguation symbols.
+cat $srcdir/units.txt | awk '{print $1}' > $tmpdir/units.list
+(echo '<eps>';) | cat - $tmpdir/units.list $tmpdir/disambig.list | awk '{print $1 " " (NR-1)}' > $dir/tokens.txt
+
+# ctc_token_fst_corrected is too big and too slow for character based chinese modeling,
+# so here just use simple ctc_token_fst
+tools/fst/ctc_token_fst.py $dir/tokens.txt | \
+  fstcompile --isymbols=$dir/tokens.txt --osymbols=$dir/tokens.txt --keep_isymbols=false --keep_osymbols=false | \
+  fstarcsort --sort_type=olabel > $dir/T.fst || exit 1;
+
+# Encode the words with indices. Will be used in lexicon and language model FST compiling.
+cat $tmpdir/lexiconp.txt | awk '{print $1}' | sort | awk '
+  BEGIN {
+    print "<eps> 0";
+  }
+  {
+    printf("%s %d\n", $1, NR);
+  }
+  END {
+    printf("#0 %d\n", NR+1);
+    printf("<s> %d\n", NR+2);
+    printf("</s> %d\n", NR+3);
+  }' > $dir/words.txt || exit 1;
+
+# Now compile the lexicon FST. Depending on the size of your lexicon, it may take some time.
+token_disambig_symbol=`grep \#0 $dir/tokens.txt | awk '{print $2}'`
+word_disambig_symbol=`grep \#0 $dir/words.txt | awk '{print $2}'`
+
+tools/fst/make_lexicon_fst.pl --pron-probs $tmpdir/lexiconp_disambig.txt 0 "sil" '#'$ndisambig | \
+  fstcompile --isymbols=$dir/tokens.txt --osymbols=$dir/words.txt \
+  --keep_isymbols=false --keep_osymbols=false |   \
+  fstaddselfloops  "echo $token_disambig_symbol |" "echo $word_disambig_symbol |" | \
+  fstarcsort --sort_type=olabel > $dir/L.fst || exit 1;
+
+echo "Lexicon and token FSTs compiling succeeded"
diff --git a/speechx/examples/aishell/tools/fst/ctc_token_fst.py b/speechx/examples/aishell/tools/fst/ctc_token_fst.py
new file mode 100755
index 00000000..048734c8
--- /dev/null
+++ b/speechx/examples/aishell/tools/fst/ctc_token_fst.py
@@ -0,0 +1,24 @@
+#!/usr/bin/env python
+
+import sys
+
+print('0 1 <eps> <eps>')
+print('1 1 <blank> <eps>')
+print('2 2 <blank> <eps>')
+print('2 0 <eps> <eps>')
+
+with open(sys.argv[1], 'r') as fin:
+    node = 3
+    for entry in fin:
+        fields = entry.strip().split(' ')
+        phone = fields[0]
+        if phone == '<eps>' or phone == '<blank>':
+            continue
+        elif '#' in phone:  # disambiguous phone
+            print('{} {} {} {}'.format(0, 0, '<eps>', phone))
+        else:
+            print('{} {} {} {}'.format(1, node, phone, phone))
+            print('{} {} {} {}'.format(node, node, phone, '<eps>'))
+            print('{} {} {} {}'.format(node, 2, '<eps>', '<eps>'))
+        node += 1
+print('0')
diff --git a/speechx/examples/aishell/tools/fst/ctc_token_fst_compact.py b/speechx/examples/aishell/tools/fst/ctc_token_fst_compact.py
new file mode 100755
index 00000000..d3018d8b
--- /dev/null
+++ b/speechx/examples/aishell/tools/fst/ctc_token_fst_compact.py
@@ -0,0 +1,21 @@
+#!/usr/bin/env python
+
+import sys
+
+print('0 0 <blank> <eps>')
+
+with open(sys.argv[1], 'r', encoding='utf8') as fin:
+    node = 1
+    for entry in fin:
+        fields = entry.strip().split(' ')
+        phone = fields[0]
+        if phone == '<eps>' or phone == '<blank>':
+            continue
+        elif '#' in phone:  # disambiguous phone
+            print('{} {} {} {}'.format(0, 0, '<eps>', phone))
+        else:
+            print('{} {} {} {}'.format(0, node, phone, phone))
+            print('{} {} {} {}'.format(node, node, phone, '<eps>'))
+            print('{} {} {} {}'.format(node, 0, '<eps>', '<eps>'))
+        node += 1
+print('0')
diff --git a/speechx/examples/aishell/tools/fst/ctc_token_fst_corrected.py b/speechx/examples/aishell/tools/fst/ctc_token_fst_corrected.py
new file mode 100755
index 00000000..81f7079e
--- /dev/null
+++ b/speechx/examples/aishell/tools/fst/ctc_token_fst_corrected.py
@@ -0,0 +1,55 @@
+#!/usr/bin/env python
+
+import sys
+
+
+def il(n):
+    return n + 1
+
+
+def ol(n):
+    return n + 1
+
+
+def s(n):
+    return n
+
+
+if __name__ == "__main__":
+    with open(sys.argv[1]) as f:
+        lines = f.readlines()
+    phone_count = 0
+    disambig_count = 0
+    for line in lines:
+        sp = line.split()
+        phone = sp[0]
+        if phone == '<eps>' or phone == '<blank>':
+            continue
+        if phone.startswith('#'):
+            disambig_count += 1
+        else:
+            phone_count += 1
+
+    # 1. add start state
+    print('0 0 {} 0'.format(il(0)))
+
+    # 2. 0 -> i, i -> i, i -> 0
+    for i in range(1, phone_count + 1):
+        print('0 {} {} {}'.format(s(i), il(i), ol(i)))
+        print('{} {} {} 0'.format(s(i), s(i), il(i)))
+        print('{} 0 {} 0'.format(s(i), il(0)))
+
+    # 3. i -> other phone
+    for i in range(1, phone_count + 1):
+        for j in range(1, phone_count + 1):
+            if i != j:
+                print('{} {} {} {}'.format(s(i), s(j), il(j), ol(j)))
+
+    # 4. add disambiguous arcs on every final state
+    for i in range(0, phone_count + 1):
+        for j in range(phone_count + 2, phone_count + disambig_count + 2):
+            print('{} {} {} {}'.format(s(i), s(i), 0, j))
+
+    # 5. every i is final state
+    for i in range(0, phone_count + 1):
+        print(s(i))
diff --git a/speechx/examples/aishell/tools/fst/eps2disambig.pl b/speechx/examples/aishell/tools/fst/eps2disambig.pl
new file mode 100755
index 00000000..e1d84a6b
--- /dev/null
+++ b/speechx/examples/aishell/tools/fst/eps2disambig.pl
@@ -0,0 +1,29 @@
+#!/usr/bin/env perl
+# Copyright 2010-2011 Microsoft Corporation
+#                2015 Guoguo Chen
+
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#  http://www.apache.org/licenses/LICENSE-2.0
+#
+# THIS CODE IS PROVIDED *AS IS* BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY
+# KIND, EITHER EXPRESS OR IMPLIED, INCLUDING WITHOUT LIMITATION ANY IMPLIED
+# WARRANTIES OR CONDITIONS OF TITLE, FITNESS FOR A PARTICULAR PURPOSE,
+# MERCHANTABLITY OR NON-INFRINGEMENT.
+# See the Apache 2 License for the specific language governing permissions and
+# limitations under the License.
+
+# This script replaces epsilon with #0 on the input side only, of the G.fst
+# acceptor.
+
+while(<>){
+  if (/\s+#0\s+/) {
+    print STDERR "$0: ERROR: LM has word #0, " .
+                 "which is reserved as disambiguation symbol\n";
+    exit 1;
+  }
+  s:^(\d+\s+\d+\s+)\<eps\>(\s+):$1#0$2:;
+  print;
+}
diff --git a/speechx/examples/aishell/tools/fst/make_lexicon_fst.pl b/speechx/examples/aishell/tools/fst/make_lexicon_fst.pl
new file mode 100755
index 00000000..f97129c0
--- /dev/null
+++ b/speechx/examples/aishell/tools/fst/make_lexicon_fst.pl
@@ -0,0 +1,155 @@
+#!/usr/bin/env perl
+use warnings; #sed replacement for -w perl parameter
+# Copyright 2010-2011  Microsoft Corporation
+#                2013  Johns Hopkins University (author: Daniel Povey)
+
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#  http://www.apache.org/licenses/LICENSE-2.0
+#
+# THIS CODE IS PROVIDED *AS IS* BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY
+# KIND, EITHER EXPRESS OR IMPLIED, INCLUDING WITHOUT LIMITATION ANY IMPLIED
+# WARRANTIES OR CONDITIONS OF TITLE, FITNESS FOR A PARTICULAR PURPOSE,
+# MERCHANTABLITY OR NON-INFRINGEMENT.
+# See the Apache 2 License for the specific language governing permissions and
+# limitations under the License.
+
+
+# makes lexicon FST, in text form, from lexicon (pronunciation probabilities optional).
+
+$pron_probs = 0;
+
+if ((@ARGV > 0) && ($ARGV[0] eq "--pron-probs")) {
+  $pron_probs = 1;
+  shift @ARGV;
+}
+
+if (@ARGV != 1 && @ARGV != 3 && @ARGV != 4) {
+  print STDERR "Usage: make_lexicon_fst.pl [--pron-probs] lexicon.txt [silprob silphone [sil_disambig_sym]] >lexiconfst.txt\n\n";
+  print STDERR "Creates a lexicon FST that transduces phones to words, and may allow optional silence.\n\n";
+  print STDERR "Note: ordinarily, each line of lexicon.txt is:\n";
+  print STDERR "  word phone1 phone2 ... phoneN;\n";
+  print STDERR "if the --pron-probs option is used, each line is:\n";
+  print STDERR "  word pronunciation-probability phone1 phone2 ... phoneN.\n\n";
+  print STDERR "The probability 'prob' will typically be between zero and one, and note that\n";
+  print STDERR "it's generally helpful to normalize so the largest one for each word is 1.0, but\n";
+  print STDERR "this is your responsibility.\n\n";
+  print STDERR "The silence disambiguation symbol, e.g. something like #5, is used only\n";
+  print STDERR "when creating a lexicon with disambiguation symbols, e.g. L_disambig.fst,\n";
+  print STDERR "and was introduced to fix a particular case of non-determinism of decoding graphs.\n\n";
+  exit(1);
+}
+
+$lexfn = shift @ARGV;
+if (@ARGV == 0) {
+  $silprob = 0.0;
+} elsif (@ARGV == 2) {
+  ($silprob,$silphone) = @ARGV;
+} else {
+  ($silprob,$silphone,$sildisambig) = @ARGV;
+}
+if ($silprob != 0.0) {
+  $silprob < 1.0 || die "Sil prob cannot be >= 1.0";
+  $silcost = -log($silprob);
+  $nosilcost = -log(1.0 - $silprob);
+}
+
+
+open(L, "<$lexfn") || die "Error opening lexicon $lexfn";
+
+
+if ( $silprob == 0.0 ) { # No optional silences: just have one (loop+final) state which is numbered zero.
+  $loopstate = 0;
+  $nextstate = 1;               # next unallocated state.
+  while (<L>) {
+    @A = split(" ", $_);
+    @A == 0 && die "Empty lexicon line.";
+    foreach $a (@A) {
+      if ($a eq "<eps>") {
+        die "Bad lexicon line $_ (<eps> is forbidden)";
+      }
+    }
+    $w = shift @A;
+    if (! $pron_probs) {
+      $pron_cost = 0.0;
+    } else {
+      $pron_prob = shift @A;
+      if (! defined $pron_prob || !($pron_prob > 0.0 && $pron_prob <= 1.0)) {
+        die "Bad pronunciation probability in line $_";
+      }
+      $pron_cost = -log($pron_prob);
+    }
+    if ($pron_cost != 0.0) { $pron_cost_string = "\t$pron_cost"; } else { $pron_cost_string = ""; }
+
+    $s = $loopstate;
+    $word_or_eps = $w;
+    while (@A > 0) {
+      $p = shift @A;
+      if (@A > 0) {
+        $ns = $nextstate++;
+      } else {
+        $ns = $loopstate;
+      }
+      print "$s\t$ns\t$p\t$word_or_eps$pron_cost_string\n";
+      $word_or_eps = "<eps>";
+      $pron_cost_string = ""; # so we only print it on the first arc of the word.
+      $s = $ns;
+    }
+  }
+  print "$loopstate\t0\n";      # final-cost.
+} else {                        # have silence probs.
+  $startstate = 0;
+  $loopstate = 1;
+  $silstate = 2;   # state from where we go to loopstate after emitting silence.
+  print "$startstate\t$loopstate\t<eps>\t<eps>\t$nosilcost\n"; # no silence.
+  if (!defined $sildisambig) {
+    print "$startstate\t$loopstate\t$silphone\t<eps>\t$silcost\n"; # silence.
+    print "$silstate\t$loopstate\t$silphone\t<eps>\n";             # no cost.
+    $nextstate = 3;
+  } else {
+    $disambigstate = 3;
+    $nextstate = 4;
+    print "$startstate\t$disambigstate\t$silphone\t<eps>\t$silcost\n"; # silence.
+    print "$silstate\t$disambigstate\t$silphone\t<eps>\n"; # no cost.
+    print "$disambigstate\t$loopstate\t$sildisambig\t<eps>\n"; # silence disambiguation symbol.
+  }
+  while (<L>) {
+    @A = split(" ", $_);
+    $w = shift @A;
+    if (! $pron_probs) {
+      $pron_cost = 0.0;
+    } else {
+      $pron_prob = shift @A;
+      if (! defined $pron_prob || !($pron_prob > 0.0 && $pron_prob <= 1.0)) {
+        die "Bad pronunciation probability in line $_";
+      }
+      $pron_cost = -log($pron_prob);
+    }
+    if ($pron_cost != 0.0) { $pron_cost_string = "\t$pron_cost"; } else { $pron_cost_string = ""; }
+    $s = $loopstate;
+    $word_or_eps = $w;
+    while (@A > 0) {
+      $p = shift @A;
+      if (@A > 0) {
+        $ns = $nextstate++;
+        print "$s\t$ns\t$p\t$word_or_eps$pron_cost_string\n";
+        $word_or_eps = "<eps>";
+        $pron_cost_string = ""; $pron_cost = 0.0; # so we only print it the 1st time.
+        $s = $ns;
+      } elsif (!defined($silphone) || $p ne $silphone) {
+        # This is non-deterministic but relatively compact,
+        # and avoids epsilons.
+        $local_nosilcost = $nosilcost + $pron_cost;
+        $local_silcost = $silcost + $pron_cost;
+        print "$s\t$loopstate\t$p\t$word_or_eps\t$local_nosilcost\n";
+        print "$s\t$silstate\t$p\t$word_or_eps\t$local_silcost\n";
+      } else {
+        # no point putting opt-sil after silence word.
+        print "$s\t$loopstate\t$p\t$word_or_eps$pron_cost_string\n";
+      }
+    }
+  }
+  print "$loopstate\t0\n";      # final-cost.
+}
diff --git a/speechx/examples/aishell/tools/fst/make_tlg.sh b/speechx/examples/aishell/tools/fst/make_tlg.sh
new file mode 100755
index 00000000..98694e55
--- /dev/null
+++ b/speechx/examples/aishell/tools/fst/make_tlg.sh
@@ -0,0 +1,38 @@
+#!/bin/bash
+#
+
+if [ -f path.sh ]; then . path.sh; fi
+
+lm_dir=$1
+src_lang=$2
+tgt_lang=$3
+
+arpa_lm=${lm_dir}/lm.arpa
+[ ! -f $arpa_lm ] && echo No such file $arpa_lm && exit 1;
+
+rm -rf $tgt_lang
+cp -r $src_lang $tgt_lang
+
+# Compose the language model to FST
+cat $arpa_lm | \
+   grep -v '<s> <s>' | \
+   grep -v '</s> <s>' | \
+   grep -v '</s> </s>' | \
+   grep -v -i '<unk>' | \
+   grep -v -i '<spoken_noise>' | \
+   arpa2fst --read-symbol-table=$tgt_lang/words.txt --keep-symbols=true - | fstprint | \
+   tools/fst/eps2disambig.pl | tools/fst/s2eps.pl | fstcompile --isymbols=$tgt_lang/words.txt \
+     --osymbols=$tgt_lang/words.txt  --keep_isymbols=false --keep_osymbols=false | \
+    fstrmepsilon | fstarcsort --sort_type=ilabel > $tgt_lang/G.fst
+
+
+echo  "Checking how stochastic G is (the first of these numbers should be small):"
+fstisstochastic $tgt_lang/G.fst
+
+# Compose the token, lexicon and language-model FST into the final decoding graph
+fsttablecompose $tgt_lang/L.fst $tgt_lang/G.fst | fstdeterminizestar --use-log=true | \
+    fstminimizeencoded | fstarcsort --sort_type=ilabel > $tgt_lang/LG.fst || exit 1;
+fsttablecompose $tgt_lang/T.fst $tgt_lang/LG.fst > $tgt_lang/TLG.fst || exit 1;
+
+echo "Composing decoding graph TLG.fst succeeded"
+#rm -r $tgt_lang/LG.fst   # We don't need to keep this intermediate FST
diff --git a/speechx/examples/aishell/tools/fst/prepare_dict.py b/speechx/examples/aishell/tools/fst/prepare_dict.py
new file mode 100755
index 00000000..b6b92ea1
--- /dev/null
+++ b/speechx/examples/aishell/tools/fst/prepare_dict.py
@@ -0,0 +1,64 @@
+#!/usr/bin/env python3
+# encoding: utf-8
+
+import sys
+
+# sys.argv[1]: e2e model unit file(lang_char.txt)
+# sys.argv[2]: raw lexicon file
+# sys.argv[3]: output lexicon file
+# sys.argv[4]: bpemodel
+
+unit_table = set()
+with open(sys.argv[1], 'r', encoding='utf8') as fin:
+    for line in fin:
+        unit = line.split()[0]
+        unit_table.add(unit)
+
+
+def contain_oov(units):
+    for unit in units:
+        if unit not in unit_table:
+            return True
+    return False
+
+
+bpemode = len(sys.argv) > 4
+if bpemode:
+    import sentencepiece as spm
+    sp = spm.SentencePieceProcessor()
+    sp.Load(sys.argv[4])
+lexicon_table = set()
+with open(sys.argv[2], 'r', encoding='utf8') as fin, \
+        open(sys.argv[3], 'w', encoding='utf8') as fout:
+    for line in fin:
+        word = line.split()[0]
+        if word == 'SIL' and not bpemode:  # `sil` might be a valid piece in bpemodel
+            continue
+        elif word == '<SPOKEN_NOISE>':
+            continue
+        else:
+            # each word only has one pronunciation for e2e system
+            if word in lexicon_table:
+                continue
+            if bpemode:
+                pieces = sp.EncodeAsPieces(word)
+                if contain_oov(pieces):
+                    print(
+                        'Ignoring words {}, which contains oov unit'.format(
+                            ''.join(word).strip('▁'))
+                    )
+                    continue
+                chars = ' '.join(
+                    [p if p in unit_table else '<unk>' for p in pieces])
+            else:
+                # ignore words with OOV
+                if contain_oov(word):
+                    print('Ignoring words {}, which contains oov unit'.format(word))
+                    continue
+                # Optional, append ▁ in front of english word
+                # we assume the model unit of our e2e system is char now.
+                if word.encode('utf8').isalpha() and '▁' in unit_table:
+                    word = '▁' + word
+                chars = ' '.join(word)  # word is a char list
+            fout.write('{} {}\n'.format(word, chars))
+            lexicon_table.add(word)
diff --git a/speechx/examples/aishell/tools/fst/remove_oovs.pl b/speechx/examples/aishell/tools/fst/remove_oovs.pl
new file mode 100755
index 00000000..ac914c3b
--- /dev/null
+++ b/speechx/examples/aishell/tools/fst/remove_oovs.pl
@@ -0,0 +1,43 @@
+#!/usr/bin/env perl
+# Copyright 2010-2011 Microsoft Corporation
+
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#  http://www.apache.org/licenses/LICENSE-2.0
+#
+# THIS CODE IS PROVIDED *AS IS* BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY
+# KIND, EITHER EXPRESS OR IMPLIED, INCLUDING WITHOUT LIMITATION ANY IMPLIED
+# WARRANTIES OR CONDITIONS OF TITLE, FITNESS FOR A PARTICULAR PURPOSE,
+# MERCHANTABLITY OR NON-INFRINGEMENT.
+# See the Apache 2 License for the specific language governing permissions and
+# limitations under the License.
+
+# This script removes lines that contain these OOVs on either the
+# third or fourth fields  of the line.  It is intended to remove arcs
+# with OOVs on, from FSTs (probably compiled from ARPAs with OOVs in).
+
+if (  @ARGV < 1 && @ARGV > 2) {
+    die "Usage: remove_oovs.pl unk_list.txt [ printed-fst ]\n";
+}
+
+$unklist = shift @ARGV;
+open(S, "<$unklist") || die "Failed opening unknown-symbol list $unklist\n";
+while(<S>){
+    @A = split(" ", $_);
+    @A == 1 || die "Bad line in unknown-symbol list: $_";
+    $unk{$A[0]} = 1;
+}
+
+$num_removed = 0;
+while(<>){
+    @A = split(" ", $_);
+    if(defined $unk{$A[2]} || defined $unk{$A[3]}) {
+        $num_removed++;
+    } else {
+        print;
+    }
+}
+print STDERR "remove_oovs.pl: removed $num_removed lines.\n";
+
diff --git a/speechx/examples/aishell/tools/fst/rnnt_token_fst.py b/speechx/examples/aishell/tools/fst/rnnt_token_fst.py
new file mode 100755
index 00000000..cc6def17
--- /dev/null
+++ b/speechx/examples/aishell/tools/fst/rnnt_token_fst.py
@@ -0,0 +1,17 @@
+#!/usr/bin/env python
+
+import sys
+
+print('0 0 <blank> <eps>')
+
+with open(sys.argv[1], 'r', encoding='utf8') as fin:
+    for entry in fin:
+        fields = entry.strip().split(' ')
+        phone = fields[0]
+        if phone == '<eps>' or phone == '<blank>':
+            continue
+        elif '#' in phone:  # disambiguous phone
+            print('{} {} {} {}'.format(0, 0, '<eps>', phone))
+        else:
+            print('{} {} {} {}'.format(0, 0, phone, phone))
+print('0')
diff --git a/speechx/examples/aishell/tools/fst/s2eps.pl b/speechx/examples/aishell/tools/fst/s2eps.pl
new file mode 100755
index 00000000..ffeeb8eb
--- /dev/null
+++ b/speechx/examples/aishell/tools/fst/s2eps.pl
@@ -0,0 +1,27 @@
+#!/usr/bin/env perl
+# Copyright 2010-2011 Microsoft Corporation
+
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#  http://www.apache.org/licenses/LICENSE-2.0
+#
+# THIS CODE IS PROVIDED *AS IS* BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY
+# KIND, EITHER EXPRESS OR IMPLIED, INCLUDING WITHOUT LIMITATION ANY IMPLIED
+# WARRANTIES OR CONDITIONS OF TITLE, FITNESS FOR A PARTICULAR PURPOSE,
+# MERCHANTABLITY OR NON-INFRINGEMENT.
+# See the Apache 2 License for the specific language governing permissions and
+# limitations under the License.
+
+# This script replaces <s> and </s> with <eps> (on both input and output sides),
+# for the G.fst acceptor.
+
+while(<>){
+    @A = split(" ", $_);
+    if ( @A >= 4 ) {
+        if ($A[2] eq "<s>" || $A[2] eq "</s>") { $A[2] = "<eps>"; }
+        if ($A[3] eq "<s>" || $A[3] eq "</s>") { $A[3] = "<eps>"; }
+    }
+    print join("\t", @A) . "\n";
+}
diff --git a/speechx/examples/aishell/tools/parse_options.sh b/speechx/examples/aishell/tools/parse_options.sh
new file mode 100755
index 00000000..34476fdb
--- /dev/null
+++ b/speechx/examples/aishell/tools/parse_options.sh
@@ -0,0 +1,97 @@
+#!/bin/bash
+
+# Copyright 2012  Johns Hopkins University (Author: Daniel Povey);
+#                 Arnab Ghoshal, Karel Vesely
+
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#  http://www.apache.org/licenses/LICENSE-2.0
+#
+# THIS CODE IS PROVIDED *AS IS* BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY
+# KIND, EITHER EXPRESS OR IMPLIED, INCLUDING WITHOUT LIMITATION ANY IMPLIED
+# WARRANTIES OR CONDITIONS OF TITLE, FITNESS FOR A PARTICULAR PURPOSE,
+# MERCHANTABLITY OR NON-INFRINGEMENT.
+# See the Apache 2 License for the specific language governing permissions and
+# limitations under the License.
+
+
+# Parse command-line options.
+# To be sourced by another script (as in ". parse_options.sh").
+# Option format is: --option-name arg
+# and shell variable "option_name" gets set to value "arg."
+# The exception is --help, which takes no arguments, but prints the
+# $help_message variable (if defined).
+
+
+###
+### The --config file options have lower priority to command line
+### options, so we need to import them first...
+###
+
+# Now import all the configs specified by command-line, in left-to-right order
+for ((argpos=1; argpos<$#; argpos++)); do
+  if [ "${!argpos}" == "--config" ]; then
+    argpos_plus1=$((argpos+1))
+    config=${!argpos_plus1}
+    [ ! -r $config ] && echo "$0: missing config '$config'" && exit 1
+    . $config  # source the config file.
+  fi
+done
+
+
+###
+### No we process the command line options
+###
+while true; do
+  [ -z "${1:-}" ] && break;  # break if there are no arguments
+  case "$1" in
+    # If the enclosing script is called with --help option, print the help
+    # message and exit.  Scripts should put help messages in $help_message
+    --help|-h) if [ -z "$help_message" ]; then echo "No help found." 1>&2;
+      else printf "$help_message\n" 1>&2 ; fi;
+      exit 0 ;;
+    --*=*) echo "$0: options to scripts must be of the form --name value, got '$1'"
+      exit 1 ;;
+    # If the first command-line argument begins with "--" (e.g. --foo-bar),
+    # then work out the variable name as $name, which will equal "foo_bar".
+    --*) name=`echo "$1" | sed s/^--// | sed s/-/_/g`;
+      # Next we test whether the variable in question is undefned-- if so it's
+      # an invalid option and we die.  Note: $0 evaluates to the name of the
+      # enclosing script.
+      # The test [ -z ${foo_bar+xxx} ] will return true if the variable foo_bar
+      # is undefined.  We then have to wrap this test inside "eval" because
+      # foo_bar is itself inside a variable ($name).
+      eval '[ -z "${'$name'+xxx}" ]' && echo "$0: invalid option $1" 1>&2 && exit 1;
+
+      oldval="`eval echo \\$$name`";
+      # Work out whether we seem to be expecting a Boolean argument.
+      if [ "$oldval" == "true" ] || [ "$oldval" == "false" ]; then
+        was_bool=true;
+      else
+        was_bool=false;
+      fi
+
+      # Set the variable to the right value-- the escaped quotes make it work if
+      # the option had spaces, like --cmd "queue.pl -sync y"
+      eval $name=\"$2\";
+
+      # Check that Boolean-valued arguments are really Boolean.
+      if $was_bool && [[ "$2" != "true" && "$2" != "false" ]]; then
+        echo "$0: expected \"true\" or \"false\": $1 $2" 1>&2
+        exit 1;
+      fi
+      shift 2;
+      ;;
+  *) break;
+  esac
+done
+
+
+# Check for an empty argument to the --cmd option, which can easily occur as a
+# result of scripting errors.
+[ ! -z "${cmd+xxx}" ] && [ -z "$cmd" ] && echo "$0: empty argument to --cmd option" 1>&2 && exit 1;
+
+
+true; # so this script returns exit code 0.
diff --git a/speechx/examples/decoder/offline_wfst_decoder_main.cc b/speechx/examples/decoder/offline_wfst_decoder_main.cc
index 7c1f6226..f0b9cc4f 100644
--- a/speechx/examples/decoder/offline_wfst_decoder_main.cc
+++ b/speechx/examples/decoder/offline_wfst_decoder_main.cc
@@ -27,6 +27,8 @@ DEFINE_string(model_path, "avg_1.jit.pdmodel", "paddle nnet model");
 DEFINE_string(param_path, "avg_1.jit.pdiparams", "paddle nnet model param");
 DEFINE_string(word_symbol_table, "vocab.txt", "word symbol table");
 DEFINE_string(graph_path, "TLG", "decoder graph");
+DEFINE_double(acoustic_scale, 1.0, "acoustic scale");
+DEFINE_int32(max_active, 5000, "decoder graph");
 
 
 using kaldi::BaseFloat;
@@ -49,11 +51,13 @@ int main(int argc, char* argv[]) {
     ppspeech::TLGDecoderOptions opts;
     opts.word_symbol_table = word_symbol_table;
     opts.fst_path = graph_path;
+    opts.opts.max_active = FLAGS_max_active;
     ppspeech::TLGDecoder decoder(opts);
 
     ppspeech::ModelOptions model_opts;
     model_opts.model_path = model_graph;
     model_opts.params_path = model_params;
+    model_opts.cache_shape = "5-1-1024,5-1-1024";
     std::shared_ptr<ppspeech::PaddleNnet> nnet(
         new ppspeech::PaddleNnet(model_opts));
     std::shared_ptr<ppspeech::RawDataCache> raw_data(
diff --git a/speechx/speechx/frontend/normalizer.cc b/speechx/speechx/frontend/normalizer.cc
index 1adddb40..52412561 100644
--- a/speechx/speechx/frontend/normalizer.cc
+++ b/speechx/speechx/frontend/normalizer.cc
@@ -107,7 +107,7 @@ void CMVN::Accept(const kaldi::VectorBase<kaldi::BaseFloat>& inputs) {
 }
 
 bool CMVN::Read(kaldi::Vector<BaseFloat>* feats) {
-    if (base_extractor_->Read(feats) == false) {
+    if (base_extractor_->Read(feats) == false || feats->Dim() == 0) {
         return false;
     }
     Compute(feats);
diff --git a/speechx/speechx/nnet/decodable.cc b/speechx/speechx/nnet/decodable.cc
index 0a5e0a77..b0f82e0c 100644
--- a/speechx/speechx/nnet/decodable.cc
+++ b/speechx/speechx/nnet/decodable.cc
@@ -36,9 +36,9 @@ int32 Decodable::NumFramesReady() const {
 }
 
 bool Decodable::IsLastFrame(int32 frame) {
-    CHECK_LE(frame, frames_ready_);
     bool flag = EnsureFrameHaveComputed(frame);
-    return (flag == false) && (frame == frames_ready_ - 1);
+    //CHECK_LE(frame, frames_ready_);
+    return (flag == false) && (frame == frames_ready_);
 }
 
 int32 Decodable::NumIndices() const { return 0; }
diff --git a/speechx/tools/fstbin/fstaddselfloops.cc b/speechx/tools/fstbin/fstaddselfloops.cc
new file mode 100644
index 00000000..145bf006
--- /dev/null
+++ b/speechx/tools/fstbin/fstaddselfloops.cc
@@ -0,0 +1,100 @@
+// fstbin/fstaddselfloops.cc
+
+// Copyright 2009-2011  Microsoft Corporation
+
+// See ../../COPYING for clarification regarding multiple authors
+//
+// Licensed under the Apache License, Version 2.0 (the "License");
+// you may not use this file except in compliance with the License.
+// You may obtain a copy of the License at
+//
+//  http://www.apache.org/licenses/LICENSE-2.0
+//
+// THIS CODE IS PROVIDED *AS IS* BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY
+// KIND, EITHER EXPRESS OR IMPLIED, INCLUDING WITHOUT LIMITATION ANY IMPLIED
+// WARRANTIES OR CONDITIONS OF TITLE, FITNESS FOR A PARTICULAR PURPOSE,
+// MERCHANTABLITY OR NON-INFRINGEMENT.
+// See the Apache 2 License for the specific language governing permissions and
+// limitations under the License.
+
+#include "base/kaldi-common.h"
+#include "fst/fstlib.h"
+#include "fstext/determinize-star.h"
+#include "fstext/fstext-utils.h"
+#include "fstext/kaldi-fst-io.h"
+#include "util/parse-options.h"
+#include "util/simple-io-funcs.h"
+
+/* some test examples:
+  pushd ~/tmpdir
+ ( echo 3; echo  4) > in.list
+ ( echo 5; echo  6) > out.list
+ ( echo "0 0 0 0"; echo "0 0" ) | fstcompile | fstaddselfloops in.list out.list
+ | fstprint ( echo "0 1 0 1"; echo " 0 2 1 0"; echo "1 0"; echo "2 0"; ) |
+ fstcompile | fstaddselfloops in.list out.list | fstprint
+*/
+
+int main(int argc, char *argv[]) {
+  try {
+    using namespace kaldi;  // NOLINT
+    using namespace fst;  // NOLINT
+    using kaldi::int32;
+
+    const char *usage =
+        "Adds self-loops to states of an FST to propagate disambiguation "
+        "symbols through it\n"
+        "They are added on each final state and each state with non-epsilon "
+        "output symbols\n"
+        "on at least one arc out of the state.  Useful in conjunction with "
+        "predeterminize\n"
+        "\n"
+        "Usage:  fstaddselfloops in-disambig-list out-disambig-list  [in.fst "
+        "[out.fst] ]\n"
+        "E.g:  fstaddselfloops in.list out.list < in.fst > withloops.fst\n"
+        "in.list and out.list are lists of integers, one per line, of the\n"
+        "same length.\n";
+
+    ParseOptions po(usage);
+    po.Read(argc, argv);
+
+    if (po.NumArgs() < 2 || po.NumArgs() > 4) {
+      po.PrintUsage();
+      exit(1);
+    }
+
+    std::string disambig_in_rxfilename = po.GetArg(1),
+                disambig_out_rxfilename = po.GetArg(2),
+                fst_in_filename = po.GetOptArg(3),
+                fst_out_filename = po.GetOptArg(4);
+
+    VectorFst<StdArc> *fst = ReadFstKaldi(fst_in_filename);
+
+    std::vector<int32> disambig_in;
+    if (!ReadIntegerVectorSimple(disambig_in_rxfilename, &disambig_in))
+      KALDI_ERR
+          << "fstaddselfloops: Could not read disambiguation symbols from "
+          << kaldi::PrintableRxfilename(disambig_in_rxfilename);
+
+    std::vector<int32> disambig_out;
+    if (!ReadIntegerVectorSimple(disambig_out_rxfilename, &disambig_out))
+      KALDI_ERR
+          << "fstaddselfloops: Could not read disambiguation symbols from "
+          << kaldi::PrintableRxfilename(disambig_out_rxfilename);
+
+    if (disambig_in.size() != disambig_out.size())
+      KALDI_ERR
+          << "fstaddselfloops: mismatch in size of disambiguation symbols";
+
+    AddSelfLoops(fst, disambig_in, disambig_out);
+
+    WriteFstKaldi(*fst, fst_out_filename);
+
+    delete fst;
+
+    return 0;
+  } catch (const std::exception &e) {
+    std::cerr << e.what();
+    return -1;
+  }
+  return 0;
+}
diff --git a/speechx/tools/fstbin/fstdeterminizestar.cc b/speechx/tools/fstbin/fstdeterminizestar.cc
new file mode 100644
index 00000000..e8181430
--- /dev/null
+++ b/speechx/tools/fstbin/fstdeterminizestar.cc
@@ -0,0 +1,114 @@
+// fstbin/fstdeterminizestar.cc
+
+// Copyright 2009-2011  Microsoft Corporation
+
+// See ../../COPYING for clarification regarding multiple authors
+//
+// Licensed under the Apache License, Version 2.0 (the "License");
+// you may not use this file except in compliance with the License.
+// You may obtain a copy of the License at
+//
+//  http://www.apache.org/licenses/LICENSE-2.0
+//
+// THIS CODE IS PROVIDED *AS IS* BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY
+// KIND, EITHER EXPRESS OR IMPLIED, INCLUDING WITHOUT LIMITATION ANY IMPLIED
+// WARRANTIES OR CONDITIONS OF TITLE, FITNESS FOR A PARTICULAR PURPOSE,
+// MERCHANTABLITY OR NON-INFRINGEMENT.
+// See the Apache 2 License for the specific language governing permissions and
+// limitations under the License.
+
+#include "base/kaldi-common.h"
+#include "fst/fstlib.h"
+#include "fstext/determinize-star.h"
+#include "fstext/fstext-utils.h"
+#include "fstext/kaldi-fst-io.h"
+#include "util/parse-options.h"
+#if !defined(_MSC_VER) && !defined(__APPLE__)
+#include <signal.h>  // Comment this line and the call to signal below if
+// it causes compilation problems.  It is only to enable a debugging procedure
+// when determinization does not terminate.  We are disabling this code if
+// compiling on Windows because signal.h is not available there, and on
+// MacOS due to a problem with <signal.h> in the initial release of Sierra.
+#endif
+
+/* some test  examples:
+ ( echo "0 0 0 0"; echo "0 0" ) | fstcompile | fstdeterminizestar | fstprint
+ ( echo "0 0 1 0"; echo "0 0" ) | fstcompile | fstdeterminizestar | fstprint
+ ( echo "0 0 1 0"; echo "0 1 1 0"; echo "0 0" ) | fstcompile |
+ fstdeterminizestar | fstprint # this last one fails [correctly]: ( echo "0 0 0
+ 1"; echo "0 0" ) | fstcompile | fstdeterminizestar | fstprint
+
+  cd ~/tmpdir
+  while true; do
+    fstrand > 1.fst
+    fstpredeterminize out.lst 1.fst | fstdeterminizestar | fstrmsymbols out.lst
+ > 2.fst fstequivalent --random=true 1.fst 2.fst || echo "Test failed" echo -n
+ "." done
+
+ Test of debugging [with non-determinizable input]:
+ ( echo " 0 0 1 0 1.0"; echo "0 1 1 0"; echo "1 1 1 0 0"; echo "0 2 2 0"; echo
+ "2"; echo "1" ) | fstcompile | fstdeterminizestar kill -SIGUSR1 [the process-id
+ of fstdeterminizestar] # prints out a bunch of debugging output showing the
+ mess it got itself into.
+*/
+
+bool debug_location = false;
+void signal_handler(int) { debug_location = true; }
+
+int main(int argc, char *argv[]) {
+  try {
+    using namespace kaldi;  // NOLINT
+    using namespace fst;  // NOLINT
+    using kaldi::int32;
+
+    const char *usage =
+        "Removes epsilons and determinizes in one step\n"
+        "\n"
+        "Usage:  fstdeterminizestar [in.fst [out.fst] ]\n"
+        "\n"
+        "See also: fstdeterminizelog, lattice-determinize\n";
+
+    float delta = kDelta;
+    int max_states = -1;
+    bool use_log = false;
+    ParseOptions po(usage);
+    po.Register("use-log", &use_log, "Determinize in log semiring.");
+    po.Register("delta", &delta,
+                "Delta value used to determine equivalence of weights.");
+    po.Register(
+        "max-states", &max_states,
+        "Maximum number of states in determinized FST before it will abort.");
+    po.Read(argc, argv);
+
+    if (po.NumArgs() > 2) {
+      po.PrintUsage();
+      exit(1);
+    }
+
+    std::string fst_in_str = po.GetOptArg(1), fst_out_str = po.GetOptArg(2);
+
+    // This enables us to get traceback info from determinization that is
+    // not seeming to terminate.
+#if !defined(_MSC_VER) && !defined(__APPLE__)
+    signal(SIGUSR1, signal_handler);
+#endif
+    // Normal case: just files.
+    VectorFst<StdArc> *fst = ReadFstKaldi(fst_in_str);
+
+    ArcSort(fst, ILabelCompare<StdArc>());  // improves speed.
+    if (use_log) {
+      DeterminizeStarInLog(fst, delta, &debug_location, max_states);
+    } else {
+      VectorFst<StdArc> det_fst;
+      DeterminizeStar(*fst, &det_fst, delta, &debug_location, max_states);
+      *fst = det_fst;  // will do shallow copy and then det_fst goes
+      // out of scope anyway.
+    }
+    WriteFstKaldi(*fst, fst_out_str);
+    delete fst;
+    return 0;
+  } catch (const std::exception &e) {
+    std::cerr << e.what();
+    return -1;
+  }
+}
diff --git a/speechx/tools/fstbin/fstisstochastic.cc b/speechx/tools/fstbin/fstisstochastic.cc
new file mode 100644
index 00000000..468ed0da
--- /dev/null
+++ b/speechx/tools/fstbin/fstisstochastic.cc
@@ -0,0 +1,91 @@
+// fstbin/fstisstochastic.cc
+
+// Copyright 2009-2011  Microsoft Corporation
+
+// See ../../COPYING for clarification regarding multiple authors
+//
+// Licensed under the Apache License, Version 2.0 (the "License");
+// you may not use this file except in compliance with the License.
+// You may obtain a copy of the License at
+//
+//  http://www.apache.org/licenses/LICENSE-2.0
+//
+// THIS CODE IS PROVIDED *AS IS* BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY
+// KIND, EITHER EXPRESS OR IMPLIED, INCLUDING WITHOUT LIMITATION ANY IMPLIED
+// WARRANTIES OR CONDITIONS OF TITLE, FITNESS FOR A PARTICULAR PURPOSE,
+// MERCHANTABLITY OR NON-INFRINGEMENT.
+// See the Apache 2 License for the specific language governing permissions and
+// limitations under the License.
+
+#include "base/kaldi-common.h"
+#include "fst/fstlib.h"
+#include "fstext/fstext-utils.h"
+#include "fstext/kaldi-fst-io.h"
+#include "util/kaldi-io.h"
+#include "util/parse-options.h"
+
+// e.g. of test:
+// echo " 0 0" | fstcompile | fstisstochastic
+// should return 0 and print "0 0" [meaning, min and
+// max weight are one = exp(0)]
+// echo " 0 1" | fstcompile | fstisstochastic
+// should  return 1, not stochastic, and print 1 1
+// (echo "0 0 0 0 0.693147 "; echo "0 1 0 0 0.693147 "; echo "1 0" ) |
+// fstcompile | fstisstochastic should return 0, stochastic; it prints "0
+// -1.78e-07" for me (echo "0 0 0 0 0.693147 "; echo "0 1 0 0 0.693147 "; echo
+// "1 0" ) | fstcompile | fstisstochastic --test-in-log=false should return 1,
+// not stochastic in tropical; it prints "0 0.693147" for me (echo "0 0 0 0 0 ";
+// echo "0 1 0 0 0 "; echo "1 0" ) | fstcompile | fstisstochastic
+// --test-in-log=false should return 0, stochastic in tropical; it prints "0 0"
+// for me (echo "0 0 0 0 0.693147 "; echo "0 1 0 0 0.693147 "; echo "1 0" ) |
+// fstcompile | fstisstochastic --test-in-log=false --delta=1 returns 0 even
+// though not stochastic because we gave it an absurdly large delta.
+
+int main(int argc, char *argv[]) {
+  try {
+    using namespace kaldi;  // NOLINT
+    using namespace fst;  // NOLINT
+    using kaldi::int32;
+
+    const char *usage =
+        "Checks whether an FST is stochastic and exits with success if so.\n"
+        "Prints out maximum error (in log units).\n"
+        "\n"
+        "Usage:  fstisstochastic [ in.fst ]\n";
+
+    float delta = 0.01;
+    bool test_in_log = true;
+
+    ParseOptions po(usage);
+    po.Register("delta", &delta, "Maximum error to accept.");
+    po.Register("test-in-log", &test_in_log,
+                "Test stochasticity in log semiring.");
+    po.Read(argc, argv);
+
+    if (po.NumArgs() > 1) {
+      po.PrintUsage();
+      exit(1);
+    }
+
+    std::string fst_in_filename = po.GetOptArg(1);
+
+    Fst<StdArc> *fst = ReadFstKaldiGeneric(fst_in_filename);
+
+    bool ans;
+    StdArc::Weight min, max;
+    if (test_in_log)
+      ans = IsStochasticFstInLog(*fst, delta, &min, &max);
+    else
+      ans = IsStochasticFst(*fst, delta, &min, &max);
+
+    std::cout << min.Value() << " " << max.Value() << '\n';
+    delete fst;
+    if (ans)
+      return 0;  // success;
+    else
+      return 1;
+  } catch (const std::exception &e) {
+    std::cerr << e.what();
+    return -1;
+  }
+}
diff --git a/speechx/tools/fstbin/fstminimizeencoded.cc b/speechx/tools/fstbin/fstminimizeencoded.cc
new file mode 100644
index 00000000..ae9ca6d7
--- /dev/null
+++ b/speechx/tools/fstbin/fstminimizeencoded.cc
@@ -0,0 +1,74 @@
+// fstbin/fstminimizeencoded.cc
+
+// Copyright 2009-2011  Microsoft Corporation
+
+// See ../../COPYING for clarification regarding multiple authors
+//
+// Licensed under the Apache License, Version 2.0 (the "License");
+// you may not use this file except in compliance with the License.
+// You may obtain a copy of the License at
+//
+//  http://www.apache.org/licenses/LICENSE-2.0
+//
+// THIS CODE IS PROVIDED *AS IS* BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY
+// KIND, EITHER EXPRESS OR IMPLIED, INCLUDING WITHOUT LIMITATION ANY IMPLIED
+// WARRANTIES OR CONDITIONS OF TITLE, FITNESS FOR A PARTICULAR PURPOSE,
+// MERCHANTABLITY OR NON-INFRINGEMENT.
+// See the Apache 2 License for the specific language governing permissions and
+// limitations under the License.
+
+#include "base/kaldi-common.h"
+#include "fst/fstlib.h"
+#include "fstext/determinize-star.h"
+#include "fstext/fstext-utils.h"
+#include "fstext/kaldi-fst-io.h"
+#include "util/kaldi-io.h"
+#include "util/parse-options.h"
+#include "util/text-utils.h"
+
+/* some test  examples:
+ ( echo "0 0 0 0"; echo "0 0" ) | fstcompile | fstminimizeencoded | fstprint
+ ( echo "0 1 0 0"; echo " 0 2 0 0"; echo "1 0"; echo "2 0"; ) | fstcompile |
+ fstminimizeencoded | fstprint
+*/
+
+int main(int argc, char *argv[]) {
+  try {
+    using namespace kaldi;  // NOLINT
+    using namespace fst;  // NOLINT
+    using kaldi::int32;
+
+    const char *usage =
+        "Minimizes FST after encoding [similar to fstminimize, but no "
+        "weight-pushing]\n"
+        "\n"
+        "Usage:  fstminimizeencoded [in.fst [out.fst] ]\n";
+
+    float delta = kDelta;
+    ParseOptions po(usage);
+    po.Register("delta", &delta,
+                "Delta likelihood used for quantization of weights");
+    po.Read(argc, argv);
+
+    if (po.NumArgs() > 2) {
+      po.PrintUsage();
+      exit(1);
+    }
+
+    std::string fst_in_filename = po.GetOptArg(1),
+                fst_out_filename = po.GetOptArg(2);
+
+    VectorFst<StdArc> *fst = ReadFstKaldi(fst_in_filename);
+
+    MinimizeEncoded(fst, delta);
+
+    WriteFstKaldi(*fst, fst_out_filename);
+
+    delete fst;
+    return 0;
+  } catch (const std::exception &e) {
+    std::cerr << e.what();
+    return -1;
+  }
+  return 0;
+}
diff --git a/speechx/tools/fstbin/fsttablecompose.cc b/speechx/tools/fstbin/fsttablecompose.cc
new file mode 100644
index 00000000..bdd476da
--- /dev/null
+++ b/speechx/tools/fstbin/fsttablecompose.cc
@@ -0,0 +1,133 @@
+// fstbin/fsttablecompose.cc
+
+// Copyright 2009-2011  Microsoft Corporation
+//                2013  Johns Hopkins University (author: Daniel Povey)
+
+// See ../../COPYING for clarification regarding multiple authors
+//
+// Licensed under the Apache License, Version 2.0 (the "License");
+// you may not use this file except in compliance with the License.
+// You may obtain a copy of the License at
+//
+//  http://www.apache.org/licenses/LICENSE-2.0
+//
+// THIS CODE IS PROVIDED *AS IS* BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY
+// KIND, EITHER EXPRESS OR IMPLIED, INCLUDING WITHOUT LIMITATION ANY IMPLIED
+// WARRANTIES OR CONDITIONS OF TITLE, FITNESS FOR A PARTICULAR PURPOSE,
+// MERCHANTABLITY OR NON-INFRINGEMENT.
+// See the Apache 2 License for the specific language governing permissions and
+// limitations under the License.
+
+#include "base/kaldi-common.h"
+#include "fst/fstlib.h"
+#include "fstext/fstext-utils.h"
+#include "fstext/kaldi-fst-io.h"
+#include "fstext/table-matcher.h"
+#include "util/parse-options.h"
+
+/*
+  cd ~/tmpdir
+  while true; do
+    fstrand  | fstarcsort --sort_type=olabel > 1.fst; fstrand | fstarcsort
+  > 2.fst fstcompose 1.fst 2.fst > 3a.fst fsttablecompose 1.fst 2.fst > 3b.fst
+    fstequivalent --random=true 3a.fst 3b.fst || echo "Test failed"
+    echo -n "."
+  done
+
+*/
+
+int main(int argc, char *argv[]) {
+  try {
+    using namespace kaldi;  // NOLINT
+    using namespace fst;  // NOLINT
+    using kaldi::int32;
+    /*
+      fsttablecompose should always give equivalent results to compose,
+      but it is more efficient for certain kinds of inputs.
+      In particular, it is useful when, say, the left FST has states
+      that typically either have epsilon olabels, or
+      one transition out for each of the possible symbols (as the
+      olabel).  The same with the input symbols of the right-hand FST
+      is possible.
+    */
+
+    const char *usage =
+        "Composition algorithm [between two FSTs of standard type, in "
+        "tropical\n"
+        "semiring] that is more efficient for certain cases-- in particular,\n"
+        "where one of the FSTs (the left one, if --match-side=left) has large\n"
+        "out-degree\n"
+        "\n"
+        "Usage:  fsttablecompose (fst1-rxfilename|fst1-rspecifier) "
+        "(fst2-rxfilename|fst2-rspecifier) [(out-rxfilename|out-rspecifier)]\n";
+
+    ParseOptions po(usage);
+
+    TableComposeOptions opts;
+    std::string match_side = "left";
+    std::string compose_filter = "sequence";
+
+    po.Register("connect", &opts.connect, "If true, trim FST before output.");
+    po.Register("match-side", &match_side,
+                "Side of composition to do table "
+                "match, one of: \"left\" or \"right\".");
+    po.Register("compose-filter", &compose_filter,
+                "Composition filter to use, "
+                "one of: \"alt_sequence\", \"auto\", \"match\", \"sequence\"");
+
+    po.Read(argc, argv);
+
+    if (match_side == "left") {
+      opts.table_match_type = MATCH_OUTPUT;
+    } else if (match_side == "right") {
+      opts.table_match_type = MATCH_INPUT;
+    } else {
+      KALDI_ERR << "Invalid match-side option: " << match_side;
+    }
+
+    if (compose_filter == "alt_sequence") {
+      opts.filter_type = ALT_SEQUENCE_FILTER;
+    } else if (compose_filter == "auto") {
+      opts.filter_type = AUTO_FILTER;
+    } else if (compose_filter == "match") {
+      opts.filter_type = MATCH_FILTER;
+    } else if (compose_filter == "sequence") {
+      opts.filter_type = SEQUENCE_FILTER;
+    } else {
+      KALDI_ERR << "Invalid compose-filter option: " << compose_filter;
+    }
+
+    if (po.NumArgs() < 2 || po.NumArgs() > 3) {
+      po.PrintUsage();
+      exit(1);
+    }
+
+    std::string fst1_in_str = po.GetArg(1), fst2_in_str = po.GetArg(2),
+                fst_out_str = po.GetOptArg(3);
+
+    VectorFst<StdArc> *fst1 = ReadFstKaldi(fst1_in_str);
+
+    VectorFst<StdArc> *fst2 = ReadFstKaldi(fst2_in_str);
+
+    // Checks if <fst1> is olabel sorted and <fst2> is ilabel sorted.
+    if (fst1->Properties(fst::kOLabelSorted, true) == 0) {
+      KALDI_WARN << "The first FST is not olabel sorted.";
+    }
+    if (fst2->Properties(fst::kILabelSorted, true) == 0) {
+      KALDI_WARN << "The second FST is not ilabel sorted.";
+    }
+
+    VectorFst<StdArc> composed_fst;
+
+    TableCompose(*fst1, *fst2, &composed_fst, opts);
+
+    delete fst1;
+    delete fst2;
+
+    WriteFstKaldi(composed_fst, fst_out_str);
+    return 0;
+  } catch (const std::exception &e) {
+    std::cerr << e.what();
+    return -1;
+  }
+}
diff --git a/speechx/tools/install_srilm.sh b/speechx/tools/install_srilm.sh
new file mode 100755
index 00000000..813109db
--- /dev/null
+++ b/speechx/tools/install_srilm.sh
@@ -0,0 +1,97 @@
+#!/usr/bin/env bash
+
+current_path=`pwd`
+current_dir=`basename "$current_path"`
+
+if [ "tools" != "$current_dir" ]; then
+    echo "You should run this script in tools/ directory!!"
+    exit 1
+fi
+
+if [ ! -d liblbfgs-1.10 ]; then
+    echo Installing libLBFGS library to support MaxEnt LMs
+    bash extras/install_liblbfgs.sh || exit 1
+fi
+
+! command -v gawk > /dev/null && \
+   echo "GNU awk is not installed so SRILM will probably not work correctly: refusing to install" && exit 1;
+
+if [ $# -ne 3 ]; then
+    echo "SRILM download requires some information about you"
+    echo
+    echo "Usage: $0 <name> <organization> <email>"
+    exit 1
+fi
+
+srilm_url="http://www.speech.sri.com/projects/srilm/srilm_download.php"
+post_data="WWW_file=srilm-1.7.3.tar.gz&WWW_name=$1&WWW_org=$2&WWW_email=$3"
+
+if ! wget --post-data "$post_data" -O ./srilm.tar.gz "$srilm_url"; then
+    echo 'There was a problem downloading the file.'
+    echo 'Check you internet connection and try again.'
+    exit 1
+fi
+
+mkdir -p srilm
+cd srilm
+
+
+if [ -f ../srilm.tgz ]; then
+    tar -xvzf ../srilm.tgz # Old SRILM format
+elif [  -f ../srilm.tar.gz ]; then
+    tar -xvzf ../srilm.tar.gz # Changed format type from tgz to tar.gz
+fi
+
+major=`gawk -F. '{ print $1 }' RELEASE`
+minor=`gawk -F. '{ print $2 }' RELEASE`
+micro=`gawk -F. '{ print $3 }' RELEASE`
+
+if [ $major -le 1 ] && [ $minor -le 7 ] && [ $micro -le 1 ]; then
+  echo "Detected version 1.7.1 or earlier. Applying patch."
+  patch -p0 < ../extras/srilm.patch
+fi
+
+# set the SRILM variable in the top-level Makefile to this directory.
+cp Makefile tmpf
+
+cat tmpf | gawk -v pwd=`pwd` '/SRILM =/{printf("SRILM = %s\n", pwd); next;} {print;}' \
+  > Makefile || exit 1
+rm tmpf
+
+mtype=`sbin/machine-type`
+
+echo HAVE_LIBLBFGS=1 >> common/Makefile.machine.$mtype
+grep ADDITIONAL_INCLUDES common/Makefile.machine.$mtype | \
+    sed 's|$| -I$(SRILM)/../liblbfgs-1.10/include|' \
+    >> common/Makefile.machine.$mtype
+
+grep ADDITIONAL_LDFLAGS common/Makefile.machine.$mtype | \
+    sed 's|$| -L$(SRILM)/../liblbfgs-1.10/lib/ -Wl,-rpath -Wl,$(SRILM)/../liblbfgs-1.10/lib/|' \
+    >> common/Makefile.machine.$mtype
+
+make || exit
+
+cd ..
+(
+  [ ! -z "${SRILM}" ] && \
+    echo >&2 "SRILM variable is aleady defined. Undefining..." && \
+    unset SRILM
+
+  [ -f ./env.sh ] && . ./env.sh
+
+  [ ! -z "${SRILM}" ] && \
+    echo >&2 "SRILM config is already in env.sh" && exit
+
+  wd=`pwd`
+  wd=`readlink -f $wd || pwd`
+
+  echo "export SRILM=$wd/srilm"
+  dirs="\${PATH}"
+  for directory in $(cd srilm && find bin -type d ) ; do
+    dirs="$dirs:\${SRILM}/$directory"
+  done
+  echo "export PATH=$dirs"
+) >> env.sh
+
+echo >&2 "Installation of SRILM finished successfully"
+echo >&2 "Please source the tools/env.sh in your path.sh to enable it"
diff --git a/speechx/tools/lmbin/CMakeLists.txt b/speechx/tools/lmbin/CMakeLists.txt
new file mode 100644
index 00000000..277e2077
--- /dev/null
+++ b/speechx/tools/lmbin/CMakeLists.txt
@@ -0,0 +1,5 @@
+cmake_minimum_required(VERSION 3.14 FATAL_ERROR)
+
+add_executable(arpa2fst ${CMAKE_CURRENT_SOURCE_DIR}/arpa2fst.cc)
+target_include_directories(arpa2fst PRIVATE ${SPEECHX_ROOT} ${SPEECHX_ROOT}/kaldi)
+target_link_libraries(arpa2fst )
diff --git a/speechx/tools/lmbin/arpa2fst.cc b/speechx/tools/lmbin/arpa2fst.cc
new file mode 100644
index 00000000..881a45c5
--- /dev/null
+++ b/speechx/tools/lmbin/arpa2fst.cc
@@ -0,0 +1,145 @@
+// bin/arpa2fst.cc
+//
+// Copyright 2009-2011  Gilles Boulianne.
+//
+// See ../../COPYING for clarification regarding multiple authors
+//
+// Licensed under the Apache License, Version 2.0 (the "License");
+// you may not use this file except in compliance with the License.
+// You may obtain a copy of the License at
+//
+//  http://www.apache.org/licenses/LICENSE-2.0
+//
+// THIS CODE IS PROVIDED *AS IS* BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY
+// KIND, EITHER EXPRESS OR IMPLIED, INCLUDING WITHOUT LIMITATION ANY IMPLIED
+// WARRANTIES OR CONDITIONS OF TITLE, FITNESS FOR A PARTICULAR PURPOSE,
+// MERCHANTABILITY OR NON-INFRINGEMENT.
+// See the Apache 2 License for the specific language governing permissions and
+// limitations under the License.
+
+#include <string>
+
+#include "lm/arpa-lm-compiler.h"
+#include "util/kaldi-io.h"
+#include "util/parse-options.h"
+
+int main(int argc, char *argv[]) {
+  using namespace kaldi;  // NOLINT
+  try {
+    const char *usage =
+        "Convert an ARPA format language model into an FST\n"
+        "Usage: arpa2fst [opts] <input-arpa> <output-fst>\n"
+        " e.g.: arpa2fst --disambig-symbol=#0 --read-symbol-table="
+        "data/lang/words.txt lm/input.arpa G.fst\n\n"
+        "Note: When called without switches, the output G.fst will contain\n"
+        "an embedded symbol table. This is compatible with the way a previous\n"
+        "version of arpa2fst worked.\n";
+
+    ParseOptions po(usage);
+
+    ArpaParseOptions options;
+    options.Register(&po);
+
+    // Option flags.
+    std::string bos_symbol = "<s>";
+    std::string eos_symbol = "</s>";
+    std::string disambig_symbol;
+    std::string read_syms_filename;
+    std::string write_syms_filename;
+    bool keep_symbols = false;
+    bool ilabel_sort = true;
+
+    po.Register("bos-symbol", &bos_symbol, "Beginning of sentence symbol");
+    po.Register("eos-symbol", &eos_symbol, "End of sentence symbol");
+    po.Register("disambig-symbol", &disambig_symbol,
+                "Disambiguator. If provided (e. g. #0), used on input side of "
+                "backoff links, and <s> and </s> are replaced with epsilons");
+    po.Register("read-symbol-table", &read_syms_filename,
+                "Use existing symbol table");
+    po.Register("write-symbol-table", &write_syms_filename,
+                "Write generated symbol table to a file");
+    po.Register("keep-symbols", &keep_symbols,
+                "Store symbol table with FST. Symbols always saved to FST if "
+                "symbol tables are neither read or written (otherwise symbols "
+                "would be lost entirely)");
+    po.Register("ilabel-sort", &ilabel_sort, "Ilabel-sort the output FST");
+
+    po.Read(argc, argv);
+
+    if (po.NumArgs() != 1 && po.NumArgs() != 2) {
+      po.PrintUsage();
+      exit(1);
+    }
+    std::string arpa_rxfilename = po.GetArg(1),
+                fst_wxfilename = po.GetOptArg(2);
+
+    int64 disambig_symbol_id = 0;
+
+    fst::SymbolTable *symbols;
+    if (!read_syms_filename.empty()) {
+      // Use existing symbols. Required symbols must be in the table.
+      kaldi::Input kisym(read_syms_filename);
+      symbols = fst::SymbolTable::ReadText(
+          kisym.Stream(), PrintableWxfilename(read_syms_filename));
+      if (symbols == NULL)
+        KALDI_ERR << "Could not read symbol table from file "
+                  << read_syms_filename;
+
+      options.oov_handling = ArpaParseOptions::kSkipNGram;
+      if (!disambig_symbol.empty()) {
+        disambig_symbol_id = symbols->Find(disambig_symbol);
+        if (disambig_symbol_id == -1)  // fst::kNoSymbol
+          KALDI_ERR << "Symbol table " << read_syms_filename
+                    << " has no symbol for " << disambig_symbol;
+      }
+    } else {
+      // Create a new symbol table and populate it from ARPA file.
+      symbols = new fst::SymbolTable(PrintableWxfilename(fst_wxfilename));
+      options.oov_handling = ArpaParseOptions::kAddToSymbols;
+      symbols->AddSymbol("<eps>", 0);
+      if (!disambig_symbol.empty()) {
+        disambig_symbol_id = symbols->AddSymbol(disambig_symbol);
+      }
+    }
+
+    // Add or use existing BOS and EOS.
+    options.bos_symbol = symbols->AddSymbol(bos_symbol);
+    options.eos_symbol = symbols->AddSymbol(eos_symbol);
+
+    // If producing new (not reading existing) symbols and not saving them,
+    // need to keep symbols with FST, otherwise they would be lost.
+    if (read_syms_filename.empty() && write_syms_filename.empty())
+      keep_symbols = true;
+
+    // Actually compile LM.
+    KALDI_ASSERT(symbols != NULL);
+    ArpaLmCompiler lm_compiler(options, disambig_symbol_id, symbols);
+    {
+      Input ki(arpa_rxfilename);
+      lm_compiler.Read(ki.Stream());
+    }
+
+    // Sort the FST in-place if requested by options.
+    if (ilabel_sort) {
+      fst::ArcSort(lm_compiler.MutableFst(), fst::StdILabelCompare());
+    }
+
+    // Write symbols if requested.
+    if (!write_syms_filename.empty()) {
+      kaldi::Output kosym(write_syms_filename, false);
+      symbols->WriteText(kosym.Stream());
+    }
+
+    // Write LM FST.
+    bool write_binary = true, write_header = false;
+    kaldi::Output kofst(fst_wxfilename, write_binary, write_header);
+    fst::FstWriteOptions wopts(PrintableWxfilename(fst_wxfilename));
+    wopts.write_isymbols = wopts.write_osymbols = keep_symbols;
+    lm_compiler.Fst().Write(kofst.Stream(), wopts);
+
+    delete symbols;
+  } catch (const std::exception &e) {
+    std::cerr << e.what();
+    return -1;
+  }
+}

From cc434566a121d0a6e7cce413d7092e5d1d1daa19 Mon Sep 17 00:00:00 2001
From: Yang Zhou <goat.zhou@qq.com>
Date: Mon, 28 Mar 2022 20:17:09 +0800
Subject: [PATCH 3/9] rm tools/fst

---
 .../aishell/tools/fst/add_lex_disambig.pl     | 195 ------------------
 .../tools/fst/compile_lexicon_token_fst.sh    |  86 --------
 .../aishell/tools/fst/ctc_token_fst.py        |  24 ---
 .../tools/fst/ctc_token_fst_compact.py        |  21 --
 .../tools/fst/ctc_token_fst_corrected.py      |  55 -----
 .../aishell/tools/fst/eps2disambig.pl         |  29 ---
 .../aishell/tools/fst/make_lexicon_fst.pl     | 155 --------------
 .../examples/aishell/tools/fst/make_tlg.sh    |  38 ----
 .../aishell/tools/fst/prepare_dict.py         |  64 ------
 .../examples/aishell/tools/fst/remove_oovs.pl |  43 ----
 .../aishell/tools/fst/rnnt_token_fst.py       |  17 --
 speechx/examples/aishell/tools/fst/s2eps.pl   |  27 ---
 .../examples/aishell/tools/parse_options.sh   |  97 ---------
 13 files changed, 851 deletions(-)
 delete mode 100755 speechx/examples/aishell/tools/fst/add_lex_disambig.pl
 delete mode 100755 speechx/examples/aishell/tools/fst/compile_lexicon_token_fst.sh
 delete mode 100755 speechx/examples/aishell/tools/fst/ctc_token_fst.py
 delete mode 100755 speechx/examples/aishell/tools/fst/ctc_token_fst_compact.py
 delete mode 100755 speechx/examples/aishell/tools/fst/ctc_token_fst_corrected.py
 delete mode 100755 speechx/examples/aishell/tools/fst/eps2disambig.pl
 delete mode 100755 speechx/examples/aishell/tools/fst/make_lexicon_fst.pl
 delete mode 100755 speechx/examples/aishell/tools/fst/make_tlg.sh
 delete mode 100755 speechx/examples/aishell/tools/fst/prepare_dict.py
 delete mode 100755 speechx/examples/aishell/tools/fst/remove_oovs.pl
 delete mode 100755 speechx/examples/aishell/tools/fst/rnnt_token_fst.py
 delete mode 100755 speechx/examples/aishell/tools/fst/s2eps.pl
 delete mode 100755 speechx/examples/aishell/tools/parse_options.sh

diff --git a/speechx/examples/aishell/tools/fst/add_lex_disambig.pl b/speechx/examples/aishell/tools/fst/add_lex_disambig.pl
deleted file mode 100755
index dd8a25de..00000000
--- a/speechx/examples/aishell/tools/fst/add_lex_disambig.pl
+++ /dev/null
@@ -1,195 +0,0 @@
-#!/usr/bin/env perl
-#  Copyright 2010-2011  Microsoft Corporation
-#            2013-2016  Johns Hopkins University (author: Daniel Povey)
-#                 2015  Hainan Xu
-#                 2015  Guoguo Chen
-
-# Licensed under the Apache License, Version 2.0 (the "License");
-# you may not use this file except in compliance with the License.
-# You may obtain a copy of the License at
-#
-#  http://www.apache.org/licenses/LICENSE-2.0
-#
-# THIS CODE IS PROVIDED *AS IS* BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY
-# KIND, EITHER EXPRESS OR IMPLIED, INCLUDING WITHOUT LIMITATION ANY IMPLIED
-# WARRANTIES OR CONDITIONS OF TITLE, FITNESS FOR A PARTICULAR PURPOSE,
-# MERCHANTABLITY OR NON-INFRINGEMENT.
-# See the Apache 2 License for the specific language governing permissions and
-# limitations under the License.
-
-
-# Adds disambiguation symbols to a lexicon.
-# Outputs still in the normal lexicon format.
-# Disambig syms are numbered #1, #2, #3, etc. (#0
-# reserved for symbol in grammar).
-# Outputs the number of disambig syms to the standard output.
-# With the --pron-probs option, expects the second field
-# of each lexicon line to be a pron-prob.
-# With the --sil-probs option, expects three additional
-# fields after the pron-prob, representing various components
-# of the silence probability model.
-
-$pron_probs = 0;
-$sil_probs = 0;
-$first_allowed_disambig = 1;
-
-for ($n = 1; $n <= 3 && @ARGV > 0; $n++) {
-  if ($ARGV[0] eq "--pron-probs") {
-    $pron_probs = 1;
-    shift @ARGV;
-  }
-  if ($ARGV[0] eq "--sil-probs") {
-    $sil_probs = 1;
-    shift @ARGV;
-  }
-  if ($ARGV[0] eq "--first-allowed-disambig") {
-    $first_allowed_disambig = 0 + $ARGV[1];
-    if ($first_allowed_disambig < 1) {
-      die "add_lex_disambig.pl: invalid --first-allowed-disambig option: $first_allowed_disambig\n";
-    }
-    shift @ARGV;
-    shift @ARGV;
-  }
-}
-
-if (@ARGV != 2) {
-  die "Usage: add_lex_disambig.pl [opts] <lexicon-in> <lexicon-out>\n" .
-    "This script adds disambiguation symbols to a lexicon in order to\n" .
-    "make decoding graphs determinizable; it adds pseudo-phone\n" .
-    "disambiguation symbols #1, #2 and so on at the ends of phones\n" .
-    "to ensure that all pronunciations are different, and that none\n" .
-    "is a prefix of another.\n" .
-    "It prints to the standard output the number of the largest-numbered" .
-    "disambiguation symbol that was used.\n" .
-    "\n" .
-    "Options:   --pron-probs       Expect pronunciation probabilities in the 2nd field\n" .
-    "           --sil-probs        [should be with --pron-probs option]\n" .
-    "                              Expect 3 extra fields after the pron-probs, for aspects of\n" .
-    "                              the silence probability model\n" .
-    "           --first-allowed-disambig <n>  The number of the first disambiguation symbol\n" .
-    "                              that this script is allowed to add.  By default this is\n" .
-    "                              #1, but you can set this to a larger value using this option.\n" .
-    "e.g.:\n" .
-    " add_lex_disambig.pl lexicon.txt lexicon_disambig.txt\n" .
-    " add_lex_disambig.pl --pron-probs lexiconp.txt lexiconp_disambig.txt\n" .
-    " add_lex_disambig.pl --pron-probs --sil-probs lexiconp_silprob.txt lexiconp_silprob_disambig.txt\n";
-}
-
-
-$lexfn = shift @ARGV;
-$lexoutfn = shift @ARGV;
-
-open(L, "<$lexfn") || die "Error opening lexicon $lexfn";
-
-# (1)  Read in the lexicon.
-@L = ( );
-while(<L>) {
-    @A = split(" ", $_);
-    push @L, join(" ", @A);
-}
-
-# (2) Work out the count of each phone-sequence in the
-# lexicon.
-
-foreach $l (@L) {
-    @A = split(" ", $l);
-    shift @A; # Remove word.
-    if ($pron_probs) {
-      $p = shift @A;
-      if (!($p > 0.0 && $p <= 1.0)) { die "Bad lexicon line $l (expecting pron-prob as second field)"; }
-    }
-    if ($sil_probs) {
-      $silp = shift @A;
-      if (!($silp > 0.0 && $silp <= 1.0)) { die "Bad lexicon line $l for silprobs"; }
-      $correction = shift @A;
-      if ($correction <= 0.0) { die "Bad lexicon line $l for silprobs"; }
-      $correction = shift @A;
-      if ($correction <= 0.0) { die "Bad lexicon line $l for silprobs"; }
-    }
-    if (!(@A)) {
-      die "Bad lexicon line $1, no phone in phone list";
-    }
-    $count{join(" ",@A)}++;
-}
-
-# (3) For each left sub-sequence of each phone-sequence, note down
-# that it exists (for identifying prefixes of longer strings).
-
-foreach $l (@L) {
-    @A = split(" ", $l);
-    shift @A; # Remove word.
-    if ($pron_probs) { shift @A; } # remove pron-prob.
-    if ($sil_probs) {
-      shift @A; # Remove silprob
-      shift @A; # Remove silprob
-    }
-    while(@A > 0) {
-        pop @A;  # Remove last phone
-        $issubseq{join(" ",@A)} = 1;
-    }
-}
-
-# (4) For each entry in the lexicon:
-#  if the phone sequence is unique and is not a
-#  prefix of another word, no diambig symbol.
-#  Else output #1, or #2, #3, ... if the same phone-seq
-#  has already been assigned a disambig symbol.
-
-
-open(O, ">$lexoutfn") || die "Opening lexicon file $lexoutfn for writing.\n";
-
-# max_disambig will always be the highest-numbered disambiguation symbol that
-# has been used so far.
-$max_disambig = $first_allowed_disambig - 1;
-
-foreach $l (@L) {
-  @A = split(" ", $l);
-  $word = shift @A;
-  if ($pron_probs) {
-    $pron_prob = shift @A;
-  }
-  if ($sil_probs) {
-    $sil_word_prob = shift @A;
-    $word_sil_correction = shift @A;
-    $prev_nonsil_correction = shift @A
-  }
-  $phnseq = join(" ", @A);
-  if (!defined $issubseq{$phnseq}
-      && $count{$phnseq} == 1) {
-    ;                           # Do nothing.
-  } else {
-    if ($phnseq eq "") {        # need disambig symbols for the empty string
-      # that are not use anywhere else.
-      $max_disambig++;
-      $reserved_for_the_empty_string{$max_disambig} = 1;
-      $phnseq = "#$max_disambig";
-    } else {
-      $cur_disambig = $last_used_disambig_symbol_of{$phnseq};
-      if (!defined $cur_disambig) {
-        $cur_disambig = $first_allowed_disambig;
-      } else {
-        $cur_disambig++;           # Get a number that has not been used yet for
-                                   # this phone sequence.
-      }
-      while (defined $reserved_for_the_empty_string{$cur_disambig}) {
-        $cur_disambig++;
-      }
-      if ($cur_disambig > $max_disambig) {
-        $max_disambig = $cur_disambig;
-      }
-      $last_used_disambig_symbol_of{$phnseq} = $cur_disambig;
-      $phnseq = $phnseq . " #" . $cur_disambig;
-    }
-  }
-  if ($pron_probs) {
-    if ($sil_probs) {
-      print O "$word\t$pron_prob\t$sil_word_prob\t$word_sil_correction\t$prev_nonsil_correction\t$phnseq\n";
-    } else {
-      print O "$word\t$pron_prob\t$phnseq\n";
-    }
-  } else {
-    print O "$word\t$phnseq\n";
-  }
-}
-
-print $max_disambig . "\n";
diff --git a/speechx/examples/aishell/tools/fst/compile_lexicon_token_fst.sh b/speechx/examples/aishell/tools/fst/compile_lexicon_token_fst.sh
deleted file mode 100755
index fda971fe..00000000
--- a/speechx/examples/aishell/tools/fst/compile_lexicon_token_fst.sh
+++ /dev/null
@@ -1,86 +0,0 @@
-#!/bin/bash
-# Copyright 2015       Yajie Miao    (Carnegie Mellon University)
-
-# Licensed under the Apache License, Version 2.0 (the "License");
-# you may not use this file except in compliance with the License.
-# You may obtain a copy of the License at
-#
-#  http://www.apache.org/licenses/LICENSE-2.0
-#
-# THIS CODE IS PROVIDED *AS IS* BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY
-# KIND, EITHER EXPRESS OR IMPLIED, INCLUDING WITHOUT LIMITATION ANY IMPLIED
-# WARRANTIES OR CONDITIONS OF TITLE, FITNESS FOR A PARTICULAR PURPOSE,
-# MERCHANTABLITY OR NON-INFRINGEMENT.
-# See the Apache 2 License for the specific language governing permissions and
-# limitations under the License.
-
-# This script compiles the lexicon and CTC tokens into FSTs. FST compiling slightly differs between the
-# phoneme and character-based lexicons.
-set -eo pipefail
-. tools/parse_options.sh
-
-if [ $# -ne 3 ]; then
-  echo "usage: tools/fst/compile_lexicon_token_fst.sh <dict-src-dir> <tmp-dir> <lang-dir>"
-  echo "e.g.: tools/fst/compile_lexicon_token_fst.sh data/local/dict data/local/lang_tmp data/lang"
-  echo "<dict-src-dir> should contain the following files:"
-  echo "lexicon.txt lexicon_numbers.txt units.txt"
-  echo "options: "
-  exit 1;
-fi
-
-srcdir=$1
-tmpdir=$2
-dir=$3
-mkdir -p $dir $tmpdir
-
-[ -f path.sh ] && . ./path.sh
-
-cp $srcdir/units.txt $dir
-
-# Add probabilities to lexicon entries. There is in fact no point of doing this here since all the entries have 1.0.
-# But utils/make_lexicon_fst.pl requires a probabilistic version, so we just leave it as it is.
-perl -ape 's/(\S+\s+)(.+)/${1}1.0\t$2/;' < $srcdir/lexicon.txt > $tmpdir/lexiconp.txt || exit 1;
-
-# Add disambiguation symbols to the lexicon. This is necessary for determinizing the composition of L.fst and G.fst.
-# Without these symbols, determinization will fail.
-ndisambig=`tools/fst/add_lex_disambig.pl $tmpdir/lexiconp.txt $tmpdir/lexiconp_disambig.txt`
-ndisambig=$[$ndisambig+1];
-
-( for n in `seq 0 $ndisambig`; do echo '#'$n; done ) > $tmpdir/disambig.list
-
-# Get the full list of CTC tokens used in FST. These tokens include <eps>, the blank <blk>,
-# the actual model unit, and the disambiguation symbols.
-cat $srcdir/units.txt | awk '{print $1}' > $tmpdir/units.list
-(echo '<eps>';) | cat - $tmpdir/units.list $tmpdir/disambig.list | awk '{print $1 " " (NR-1)}' > $dir/tokens.txt
-
-# ctc_token_fst_corrected is too big and too slow for character based chinese modeling,
-# so here just use simple ctc_token_fst
-tools/fst/ctc_token_fst.py $dir/tokens.txt | \
-  fstcompile --isymbols=$dir/tokens.txt --osymbols=$dir/tokens.txt --keep_isymbols=false --keep_osymbols=false | \
-  fstarcsort --sort_type=olabel > $dir/T.fst || exit 1;
-
-# Encode the words with indices. Will be used in lexicon and language model FST compiling.
-cat $tmpdir/lexiconp.txt | awk '{print $1}' | sort | awk '
-  BEGIN {
-    print "<eps> 0";
-  }
-  {
-    printf("%s %d\n", $1, NR);
-  }
-  END {
-    printf("#0 %d\n", NR+1);
-    printf("<s> %d\n", NR+2);
-    printf("</s> %d\n", NR+3);
-  }' > $dir/words.txt || exit 1;
-
-# Now compile the lexicon FST. Depending on the size of your lexicon, it may take some time.
-token_disambig_symbol=`grep \#0 $dir/tokens.txt | awk '{print $2}'`
-word_disambig_symbol=`grep \#0 $dir/words.txt | awk '{print $2}'`
-
-tools/fst/make_lexicon_fst.pl --pron-probs $tmpdir/lexiconp_disambig.txt 0 "sil" '#'$ndisambig | \
-  fstcompile --isymbols=$dir/tokens.txt --osymbols=$dir/words.txt \
-  --keep_isymbols=false --keep_osymbols=false |   \
-  fstaddselfloops  "echo $token_disambig_symbol |" "echo $word_disambig_symbol |" | \
-  fstarcsort --sort_type=olabel > $dir/L.fst || exit 1;
-
-echo "Lexicon and token FSTs compiling succeeded"
diff --git a/speechx/examples/aishell/tools/fst/ctc_token_fst.py b/speechx/examples/aishell/tools/fst/ctc_token_fst.py
deleted file mode 100755
index 048734c8..00000000
--- a/speechx/examples/aishell/tools/fst/ctc_token_fst.py
+++ /dev/null
@@ -1,24 +0,0 @@
-#!/usr/bin/env python
-
-import sys
-
-print('0 1 <eps> <eps>')
-print('1 1 <blank> <eps>')
-print('2 2 <blank> <eps>')
-print('2 0 <eps> <eps>')
-
-with open(sys.argv[1], 'r') as fin:
-    node = 3
-    for entry in fin:
-        fields = entry.strip().split(' ')
-        phone = fields[0]
-        if phone == '<eps>' or phone == '<blank>':
-            continue
-        elif '#' in phone:  # disambiguous phone
-            print('{} {} {} {}'.format(0, 0, '<eps>', phone))
-        else:
-            print('{} {} {} {}'.format(1, node, phone, phone))
-            print('{} {} {} {}'.format(node, node, phone, '<eps>'))
-            print('{} {} {} {}'.format(node, 2, '<eps>', '<eps>'))
-        node += 1
-print('0')
diff --git a/speechx/examples/aishell/tools/fst/ctc_token_fst_compact.py b/speechx/examples/aishell/tools/fst/ctc_token_fst_compact.py
deleted file mode 100755
index d3018d8b..00000000
--- a/speechx/examples/aishell/tools/fst/ctc_token_fst_compact.py
+++ /dev/null
@@ -1,21 +0,0 @@
-#!/usr/bin/env python
-
-import sys
-
-print('0 0 <blank> <eps>')
-
-with open(sys.argv[1], 'r', encoding='utf8') as fin:
-    node = 1
-    for entry in fin:
-        fields = entry.strip().split(' ')
-        phone = fields[0]
-        if phone == '<eps>' or phone == '<blank>':
-            continue
-        elif '#' in phone:  # disambiguous phone
-            print('{} {} {} {}'.format(0, 0, '<eps>', phone))
-        else:
-            print('{} {} {} {}'.format(0, node, phone, phone))
-            print('{} {} {} {}'.format(node, node, phone, '<eps>'))
-            print('{} {} {} {}'.format(node, 0, '<eps>', '<eps>'))
-        node += 1
-print('0')
diff --git a/speechx/examples/aishell/tools/fst/ctc_token_fst_corrected.py b/speechx/examples/aishell/tools/fst/ctc_token_fst_corrected.py
deleted file mode 100755
index 81f7079e..00000000
--- a/speechx/examples/aishell/tools/fst/ctc_token_fst_corrected.py
+++ /dev/null
@@ -1,55 +0,0 @@
-#!/usr/bin/env python
-
-import sys
-
-
-def il(n):
-    return n + 1
-
-
-def ol(n):
-    return n + 1
-
-
-def s(n):
-    return n
-
-
-if __name__ == "__main__":
-    with open(sys.argv[1]) as f:
-        lines = f.readlines()
-    phone_count = 0
-    disambig_count = 0
-    for line in lines:
-        sp = line.split()
-        phone = sp[0]
-        if phone == '<eps>' or phone == '<blank>':
-            continue
-        if phone.startswith('#'):
-            disambig_count += 1
-        else:
-            phone_count += 1
-
-    # 1. add start state
-    print('0 0 {} 0'.format(il(0)))
-
-    # 2. 0 -> i, i -> i, i -> 0
-    for i in range(1, phone_count + 1):
-        print('0 {} {} {}'.format(s(i), il(i), ol(i)))
-        print('{} {} {} 0'.format(s(i), s(i), il(i)))
-        print('{} 0 {} 0'.format(s(i), il(0)))
-
-    # 3. i -> other phone
-    for i in range(1, phone_count + 1):
-        for j in range(1, phone_count + 1):
-            if i != j:
-                print('{} {} {} {}'.format(s(i), s(j), il(j), ol(j)))
-
-    # 4. add disambiguous arcs on every final state
-    for i in range(0, phone_count + 1):
-        for j in range(phone_count + 2, phone_count + disambig_count + 2):
-            print('{} {} {} {}'.format(s(i), s(i), 0, j))
-
-    # 5. every i is final state
-    for i in range(0, phone_count + 1):
-        print(s(i))
diff --git a/speechx/examples/aishell/tools/fst/eps2disambig.pl b/speechx/examples/aishell/tools/fst/eps2disambig.pl
deleted file mode 100755
index e1d84a6b..00000000
--- a/speechx/examples/aishell/tools/fst/eps2disambig.pl
+++ /dev/null
@@ -1,29 +0,0 @@
-#!/usr/bin/env perl
-# Copyright 2010-2011 Microsoft Corporation
-#                2015 Guoguo Chen
-
-# Licensed under the Apache License, Version 2.0 (the "License");
-# you may not use this file except in compliance with the License.
-# You may obtain a copy of the License at
-#
-#  http://www.apache.org/licenses/LICENSE-2.0
-#
-# THIS CODE IS PROVIDED *AS IS* BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY
-# KIND, EITHER EXPRESS OR IMPLIED, INCLUDING WITHOUT LIMITATION ANY IMPLIED
-# WARRANTIES OR CONDITIONS OF TITLE, FITNESS FOR A PARTICULAR PURPOSE,
-# MERCHANTABLITY OR NON-INFRINGEMENT.
-# See the Apache 2 License for the specific language governing permissions and
-# limitations under the License.
-
-# This script replaces epsilon with #0 on the input side only, of the G.fst
-# acceptor.
-
-while(<>){
-  if (/\s+#0\s+/) {
-    print STDERR "$0: ERROR: LM has word #0, " .
-                 "which is reserved as disambiguation symbol\n";
-    exit 1;
-  }
-  s:^(\d+\s+\d+\s+)\<eps\>(\s+):$1#0$2:;
-  print;
-}
diff --git a/speechx/examples/aishell/tools/fst/make_lexicon_fst.pl b/speechx/examples/aishell/tools/fst/make_lexicon_fst.pl
deleted file mode 100755
index f97129c0..00000000
--- a/speechx/examples/aishell/tools/fst/make_lexicon_fst.pl
+++ /dev/null
@@ -1,155 +0,0 @@
-#!/usr/bin/env perl
-use warnings; #sed replacement for -w perl parameter
-# Copyright 2010-2011  Microsoft Corporation
-#                2013  Johns Hopkins University (author: Daniel Povey)
-
-# Licensed under the Apache License, Version 2.0 (the "License");
-# you may not use this file except in compliance with the License.
-# You may obtain a copy of the License at
-#
-#  http://www.apache.org/licenses/LICENSE-2.0
-#
-# THIS CODE IS PROVIDED *AS IS* BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY
-# KIND, EITHER EXPRESS OR IMPLIED, INCLUDING WITHOUT LIMITATION ANY IMPLIED
-# WARRANTIES OR CONDITIONS OF TITLE, FITNESS FOR A PARTICULAR PURPOSE,
-# MERCHANTABLITY OR NON-INFRINGEMENT.
-# See the Apache 2 License for the specific language governing permissions and
-# limitations under the License.
-
-
-# makes lexicon FST, in text form, from lexicon (pronunciation probabilities optional).
-
-$pron_probs = 0;
-
-if ((@ARGV > 0) && ($ARGV[0] eq "--pron-probs")) {
-  $pron_probs = 1;
-  shift @ARGV;
-}
-
-if (@ARGV != 1 && @ARGV != 3 && @ARGV != 4) {
-  print STDERR "Usage: make_lexicon_fst.pl [--pron-probs] lexicon.txt [silprob silphone [sil_disambig_sym]] >lexiconfst.txt\n\n";
-  print STDERR "Creates a lexicon FST that transduces phones to words, and may allow optional silence.\n\n";
-  print STDERR "Note: ordinarily, each line of lexicon.txt is:\n";
-  print STDERR "  word phone1 phone2 ... phoneN;\n";
-  print STDERR "if the --pron-probs option is used, each line is:\n";
-  print STDERR "  word pronunciation-probability phone1 phone2 ... phoneN.\n\n";
-  print STDERR "The probability 'prob' will typically be between zero and one, and note that\n";
-  print STDERR "it's generally helpful to normalize so the largest one for each word is 1.0, but\n";
-  print STDERR "this is your responsibility.\n\n";
-  print STDERR "The silence disambiguation symbol, e.g. something like #5, is used only\n";
-  print STDERR "when creating a lexicon with disambiguation symbols, e.g. L_disambig.fst,\n";
-  print STDERR "and was introduced to fix a particular case of non-determinism of decoding graphs.\n\n";
-  exit(1);
-}
-
-$lexfn = shift @ARGV;
-if (@ARGV == 0) {
-  $silprob = 0.0;
-} elsif (@ARGV == 2) {
-  ($silprob,$silphone) = @ARGV;
-} else {
-  ($silprob,$silphone,$sildisambig) = @ARGV;
-}
-if ($silprob != 0.0) {
-  $silprob < 1.0 || die "Sil prob cannot be >= 1.0";
-  $silcost = -log($silprob);
-  $nosilcost = -log(1.0 - $silprob);
-}
-
-
-open(L, "<$lexfn") || die "Error opening lexicon $lexfn";
-
-
-if ( $silprob == 0.0 ) { # No optional silences: just have one (loop+final) state which is numbered zero.
-  $loopstate = 0;
-  $nextstate = 1;               # next unallocated state.
-  while (<L>) {
-    @A = split(" ", $_);
-    @A == 0 && die "Empty lexicon line.";
-    foreach $a (@A) {
-      if ($a eq "<eps>") {
-        die "Bad lexicon line $_ (<eps> is forbidden)";
-      }
-    }
-    $w = shift @A;
-    if (! $pron_probs) {
-      $pron_cost = 0.0;
-    } else {
-      $pron_prob = shift @A;
-      if (! defined $pron_prob || !($pron_prob > 0.0 && $pron_prob <= 1.0)) {
-        die "Bad pronunciation probability in line $_";
-      }
-      $pron_cost = -log($pron_prob);
-    }
-    if ($pron_cost != 0.0) { $pron_cost_string = "\t$pron_cost"; } else { $pron_cost_string = ""; }
-
-    $s = $loopstate;
-    $word_or_eps = $w;
-    while (@A > 0) {
-      $p = shift @A;
-      if (@A > 0) {
-        $ns = $nextstate++;
-      } else {
-        $ns = $loopstate;
-      }
-      print "$s\t$ns\t$p\t$word_or_eps$pron_cost_string\n";
-      $word_or_eps = "<eps>";
-      $pron_cost_string = ""; # so we only print it on the first arc of the word.
-      $s = $ns;
-    }
-  }
-  print "$loopstate\t0\n";      # final-cost.
-} else {                        # have silence probs.
-  $startstate = 0;
-  $loopstate = 1;
-  $silstate = 2;   # state from where we go to loopstate after emitting silence.
-  print "$startstate\t$loopstate\t<eps>\t<eps>\t$nosilcost\n"; # no silence.
-  if (!defined $sildisambig) {
-    print "$startstate\t$loopstate\t$silphone\t<eps>\t$silcost\n"; # silence.
-    print "$silstate\t$loopstate\t$silphone\t<eps>\n";             # no cost.
-    $nextstate = 3;
-  } else {
-    $disambigstate = 3;
-    $nextstate = 4;
-    print "$startstate\t$disambigstate\t$silphone\t<eps>\t$silcost\n"; # silence.
-    print "$silstate\t$disambigstate\t$silphone\t<eps>\n"; # no cost.
-    print "$disambigstate\t$loopstate\t$sildisambig\t<eps>\n"; # silence disambiguation symbol.
-  }
-  while (<L>) {
-    @A = split(" ", $_);
-    $w = shift @A;
-    if (! $pron_probs) {
-      $pron_cost = 0.0;
-    } else {
-      $pron_prob = shift @A;
-      if (! defined $pron_prob || !($pron_prob > 0.0 && $pron_prob <= 1.0)) {
-        die "Bad pronunciation probability in line $_";
-      }
-      $pron_cost = -log($pron_prob);
-    }
-    if ($pron_cost != 0.0) { $pron_cost_string = "\t$pron_cost"; } else { $pron_cost_string = ""; }
-    $s = $loopstate;
-    $word_or_eps = $w;
-    while (@A > 0) {
-      $p = shift @A;
-      if (@A > 0) {
-        $ns = $nextstate++;
-        print "$s\t$ns\t$p\t$word_or_eps$pron_cost_string\n";
-        $word_or_eps = "<eps>";
-        $pron_cost_string = ""; $pron_cost = 0.0; # so we only print it the 1st time.
-        $s = $ns;
-      } elsif (!defined($silphone) || $p ne $silphone) {
-        # This is non-deterministic but relatively compact,
-        # and avoids epsilons.
-        $local_nosilcost = $nosilcost + $pron_cost;
-        $local_silcost = $silcost + $pron_cost;
-        print "$s\t$loopstate\t$p\t$word_or_eps\t$local_nosilcost\n";
-        print "$s\t$silstate\t$p\t$word_or_eps\t$local_silcost\n";
-      } else {
-        # no point putting opt-sil after silence word.
-        print "$s\t$loopstate\t$p\t$word_or_eps$pron_cost_string\n";
-      }
-    }
-  }
-  print "$loopstate\t0\n";      # final-cost.
-}
diff --git a/speechx/examples/aishell/tools/fst/make_tlg.sh b/speechx/examples/aishell/tools/fst/make_tlg.sh
deleted file mode 100755
index 98694e55..00000000
--- a/speechx/examples/aishell/tools/fst/make_tlg.sh
+++ /dev/null
@@ -1,38 +0,0 @@
-#!/bin/bash
-#
-
-if [ -f path.sh ]; then . path.sh; fi
-
-lm_dir=$1
-src_lang=$2
-tgt_lang=$3
-
-arpa_lm=${lm_dir}/lm.arpa
-[ ! -f $arpa_lm ] && echo No such file $arpa_lm && exit 1;
-
-rm -rf $tgt_lang
-cp -r $src_lang $tgt_lang
-
-# Compose the language model to FST
-cat $arpa_lm | \
-   grep -v '<s> <s>' | \
-   grep -v '</s> <s>' | \
-   grep -v '</s> </s>' | \
-   grep -v -i '<unk>' | \
-   grep -v -i '<spoken_noise>' | \
-   arpa2fst --read-symbol-table=$tgt_lang/words.txt --keep-symbols=true - | fstprint | \
-   tools/fst/eps2disambig.pl | tools/fst/s2eps.pl | fstcompile --isymbols=$tgt_lang/words.txt \
-     --osymbols=$tgt_lang/words.txt  --keep_isymbols=false --keep_osymbols=false | \
-    fstrmepsilon | fstarcsort --sort_type=ilabel > $tgt_lang/G.fst
-
-
-echo  "Checking how stochastic G is (the first of these numbers should be small):"
-fstisstochastic $tgt_lang/G.fst
-
-# Compose the token, lexicon and language-model FST into the final decoding graph
-fsttablecompose $tgt_lang/L.fst $tgt_lang/G.fst | fstdeterminizestar --use-log=true | \
-    fstminimizeencoded | fstarcsort --sort_type=ilabel > $tgt_lang/LG.fst || exit 1;
-fsttablecompose $tgt_lang/T.fst $tgt_lang/LG.fst > $tgt_lang/TLG.fst || exit 1;
-
-echo "Composing decoding graph TLG.fst succeeded"
-#rm -r $tgt_lang/LG.fst   # We don't need to keep this intermediate FST
diff --git a/speechx/examples/aishell/tools/fst/prepare_dict.py b/speechx/examples/aishell/tools/fst/prepare_dict.py
deleted file mode 100755
index b6b92ea1..00000000
--- a/speechx/examples/aishell/tools/fst/prepare_dict.py
+++ /dev/null
@@ -1,64 +0,0 @@
-#!/usr/bin/env python3
-# encoding: utf-8
-
-import sys
-
-# sys.argv[1]: e2e model unit file(lang_char.txt)
-# sys.argv[2]: raw lexicon file
-# sys.argv[3]: output lexicon file
-# sys.argv[4]: bpemodel
-
-unit_table = set()
-with open(sys.argv[1], 'r', encoding='utf8') as fin:
-    for line in fin:
-        unit = line.split()[0]
-        unit_table.add(unit)
-
-
-def contain_oov(units):
-    for unit in units:
-        if unit not in unit_table:
-            return True
-    return False
-
-
-bpemode = len(sys.argv) > 4
-if bpemode:
-    import sentencepiece as spm
-    sp = spm.SentencePieceProcessor()
-    sp.Load(sys.argv[4])
-lexicon_table = set()
-with open(sys.argv[2], 'r', encoding='utf8') as fin, \
-        open(sys.argv[3], 'w', encoding='utf8') as fout:
-    for line in fin:
-        word = line.split()[0]
-        if word == 'SIL' and not bpemode:  # `sil` might be a valid piece in bpemodel
-            continue
-        elif word == '<SPOKEN_NOISE>':
-            continue
-        else:
-            # each word only has one pronunciation for e2e system
-            if word in lexicon_table:
-                continue
-            if bpemode:
-                pieces = sp.EncodeAsPieces(word)
-                if contain_oov(pieces):
-                    print(
-                        'Ignoring words {}, which contains oov unit'.format(
-                            ''.join(word).strip('▁'))
-                    )
-                    continue
-                chars = ' '.join(
-                    [p if p in unit_table else '<unk>' for p in pieces])
-            else:
-                # ignore words with OOV
-                if contain_oov(word):
-                    print('Ignoring words {}, which contains oov unit'.format(word))
-                    continue
-                # Optional, append ▁ in front of english word
-                # we assume the model unit of our e2e system is char now.
-                if word.encode('utf8').isalpha() and '▁' in unit_table:
-                    word = '▁' + word
-                chars = ' '.join(word)  # word is a char list
-            fout.write('{} {}\n'.format(word, chars))
-            lexicon_table.add(word)
diff --git a/speechx/examples/aishell/tools/fst/remove_oovs.pl b/speechx/examples/aishell/tools/fst/remove_oovs.pl
deleted file mode 100755
index ac914c3b..00000000
--- a/speechx/examples/aishell/tools/fst/remove_oovs.pl
+++ /dev/null
@@ -1,43 +0,0 @@
-#!/usr/bin/env perl
-# Copyright 2010-2011 Microsoft Corporation
-
-# Licensed under the Apache License, Version 2.0 (the "License");
-# you may not use this file except in compliance with the License.
-# You may obtain a copy of the License at
-#
-#  http://www.apache.org/licenses/LICENSE-2.0
-#
-# THIS CODE IS PROVIDED *AS IS* BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY
-# KIND, EITHER EXPRESS OR IMPLIED, INCLUDING WITHOUT LIMITATION ANY IMPLIED
-# WARRANTIES OR CONDITIONS OF TITLE, FITNESS FOR A PARTICULAR PURPOSE,
-# MERCHANTABLITY OR NON-INFRINGEMENT.
-# See the Apache 2 License for the specific language governing permissions and
-# limitations under the License.
-
-# This script removes lines that contain these OOVs on either the
-# third or fourth fields  of the line.  It is intended to remove arcs
-# with OOVs on, from FSTs (probably compiled from ARPAs with OOVs in).
-
-if (  @ARGV < 1 && @ARGV > 2) {
-    die "Usage: remove_oovs.pl unk_list.txt [ printed-fst ]\n";
-}
-
-$unklist = shift @ARGV;
-open(S, "<$unklist") || die "Failed opening unknown-symbol list $unklist\n";
-while(<S>){
-    @A = split(" ", $_);
-    @A == 1 || die "Bad line in unknown-symbol list: $_";
-    $unk{$A[0]} = 1;
-}
-
-$num_removed = 0;
-while(<>){
-    @A = split(" ", $_);
-    if(defined $unk{$A[2]} || defined $unk{$A[3]}) {
-        $num_removed++;
-    } else {
-        print;
-    }
-}
-print STDERR "remove_oovs.pl: removed $num_removed lines.\n";
-
diff --git a/speechx/examples/aishell/tools/fst/rnnt_token_fst.py b/speechx/examples/aishell/tools/fst/rnnt_token_fst.py
deleted file mode 100755
index cc6def17..00000000
--- a/speechx/examples/aishell/tools/fst/rnnt_token_fst.py
+++ /dev/null
@@ -1,17 +0,0 @@
-#!/usr/bin/env python
-
-import sys
-
-print('0 0 <blank> <eps>')
-
-with open(sys.argv[1], 'r', encoding='utf8') as fin:
-    for entry in fin:
-        fields = entry.strip().split(' ')
-        phone = fields[0]
-        if phone == '<eps>' or phone == '<blank>':
-            continue
-        elif '#' in phone:  # disambiguous phone
-            print('{} {} {} {}'.format(0, 0, '<eps>', phone))
-        else:
-            print('{} {} {} {}'.format(0, 0, phone, phone))
-print('0')
diff --git a/speechx/examples/aishell/tools/fst/s2eps.pl b/speechx/examples/aishell/tools/fst/s2eps.pl
deleted file mode 100755
index ffeeb8eb..00000000
--- a/speechx/examples/aishell/tools/fst/s2eps.pl
+++ /dev/null
@@ -1,27 +0,0 @@
-#!/usr/bin/env perl
-# Copyright 2010-2011 Microsoft Corporation
-
-# Licensed under the Apache License, Version 2.0 (the "License");
-# you may not use this file except in compliance with the License.
-# You may obtain a copy of the License at
-#
-#  http://www.apache.org/licenses/LICENSE-2.0
-#
-# THIS CODE IS PROVIDED *AS IS* BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY
-# KIND, EITHER EXPRESS OR IMPLIED, INCLUDING WITHOUT LIMITATION ANY IMPLIED
-# WARRANTIES OR CONDITIONS OF TITLE, FITNESS FOR A PARTICULAR PURPOSE,
-# MERCHANTABLITY OR NON-INFRINGEMENT.
-# See the Apache 2 License for the specific language governing permissions and
-# limitations under the License.
-
-# This script replaces <s> and </s> with <eps> (on both input and output sides),
-# for the G.fst acceptor.
-
-while(<>){
-    @A = split(" ", $_);
-    if ( @A >= 4 ) {
-        if ($A[2] eq "<s>" || $A[2] eq "</s>") { $A[2] = "<eps>"; }
-        if ($A[3] eq "<s>" || $A[3] eq "</s>") { $A[3] = "<eps>"; }
-    }
-    print join("\t", @A) . "\n";
-}
diff --git a/speechx/examples/aishell/tools/parse_options.sh b/speechx/examples/aishell/tools/parse_options.sh
deleted file mode 100755
index 34476fdb..00000000
--- a/speechx/examples/aishell/tools/parse_options.sh
+++ /dev/null
@@ -1,97 +0,0 @@
-#!/bin/bash
-
-# Copyright 2012  Johns Hopkins University (Author: Daniel Povey);
-#                 Arnab Ghoshal, Karel Vesely
-
-# Licensed under the Apache License, Version 2.0 (the "License");
-# you may not use this file except in compliance with the License.
-# You may obtain a copy of the License at
-#
-#  http://www.apache.org/licenses/LICENSE-2.0
-#
-# THIS CODE IS PROVIDED *AS IS* BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY
-# KIND, EITHER EXPRESS OR IMPLIED, INCLUDING WITHOUT LIMITATION ANY IMPLIED
-# WARRANTIES OR CONDITIONS OF TITLE, FITNESS FOR A PARTICULAR PURPOSE,
-# MERCHANTABLITY OR NON-INFRINGEMENT.
-# See the Apache 2 License for the specific language governing permissions and
-# limitations under the License.
-
-
-# Parse command-line options.
-# To be sourced by another script (as in ". parse_options.sh").
-# Option format is: --option-name arg
-# and shell variable "option_name" gets set to value "arg."
-# The exception is --help, which takes no arguments, but prints the
-# $help_message variable (if defined).
-
-
-###
-### The --config file options have lower priority to command line
-### options, so we need to import them first...
-###
-
-# Now import all the configs specified by command-line, in left-to-right order
-for ((argpos=1; argpos<$#; argpos++)); do
-  if [ "${!argpos}" == "--config" ]; then
-    argpos_plus1=$((argpos+1))
-    config=${!argpos_plus1}
-    [ ! -r $config ] && echo "$0: missing config '$config'" && exit 1
-    . $config  # source the config file.
-  fi
-done
-
-
-###
-### No we process the command line options
-###
-while true; do
-  [ -z "${1:-}" ] && break;  # break if there are no arguments
-  case "$1" in
-    # If the enclosing script is called with --help option, print the help
-    # message and exit.  Scripts should put help messages in $help_message
-    --help|-h) if [ -z "$help_message" ]; then echo "No help found." 1>&2;
-      else printf "$help_message\n" 1>&2 ; fi;
-      exit 0 ;;
-    --*=*) echo "$0: options to scripts must be of the form --name value, got '$1'"
-      exit 1 ;;
-    # If the first command-line argument begins with "--" (e.g. --foo-bar),
-    # then work out the variable name as $name, which will equal "foo_bar".
-    --*) name=`echo "$1" | sed s/^--// | sed s/-/_/g`;
-      # Next we test whether the variable in question is undefned-- if so it's
-      # an invalid option and we die.  Note: $0 evaluates to the name of the
-      # enclosing script.
-      # The test [ -z ${foo_bar+xxx} ] will return true if the variable foo_bar
-      # is undefined.  We then have to wrap this test inside "eval" because
-      # foo_bar is itself inside a variable ($name).
-      eval '[ -z "${'$name'+xxx}" ]' && echo "$0: invalid option $1" 1>&2 && exit 1;
-
-      oldval="`eval echo \\$$name`";
-      # Work out whether we seem to be expecting a Boolean argument.
-      if [ "$oldval" == "true" ] || [ "$oldval" == "false" ]; then
-        was_bool=true;
-      else
-        was_bool=false;
-      fi
-
-      # Set the variable to the right value-- the escaped quotes make it work if
-      # the option had spaces, like --cmd "queue.pl -sync y"
-      eval $name=\"$2\";
-
-      # Check that Boolean-valued arguments are really Boolean.
-      if $was_bool && [[ "$2" != "true" && "$2" != "false" ]]; then
-        echo "$0: expected \"true\" or \"false\": $1 $2" 1>&2
-        exit 1;
-      fi
-      shift 2;
-      ;;
-  *) break;
-  esac
-done
-
-
-# Check for an empty argument to the --cmd option, which can easily occur as a
-# result of scripting errors.
-[ ! -z "${cmd+xxx}" ] && [ -z "$cmd" ] && echo "$0: empty argument to --cmd option" 1>&2 && exit 1;
-
-
-true; # so this script returns exit code 0.

From 5170ccf00db57264fd9a4eda6ef38c8096705a3b Mon Sep 17 00:00:00 2001
From: Yang Zhou <goat.zhou@qq.com>
Date: Mon, 28 Mar 2022 20:26:35 +0800
Subject: [PATCH 4/9] rm example/aishell

---
 .../aishell/local/aishell_train_lms.sh        | 59 -------------------
 speechx/examples/aishell/mkgraph.sh           | 31 ----------
 speechx/examples/aishell/path.sh              | 14 -----
 3 files changed, 104 deletions(-)
 delete mode 100755 speechx/examples/aishell/local/aishell_train_lms.sh
 delete mode 100644 speechx/examples/aishell/mkgraph.sh
 delete mode 100644 speechx/examples/aishell/path.sh

diff --git a/speechx/examples/aishell/local/aishell_train_lms.sh b/speechx/examples/aishell/local/aishell_train_lms.sh
deleted file mode 100755
index 30ffb797..00000000
--- a/speechx/examples/aishell/local/aishell_train_lms.sh
+++ /dev/null
@@ -1,59 +0,0 @@
-#!/bin/bash
-
-
-# To be run from one directory above this script.
-. ./path.sh
-
-text=data/local/lm/text
-lexicon=data/local/dict/lexicon.txt
-
-for f in "$text" "$lexicon"; do
-  [ ! -f $x ] && echo "$0: No such file $f" && exit 1;
-done
-
-# Check SRILM tools
-if ! which ngram-count > /dev/null; then
-    echo "srilm tools are not found, please download it and install it from: "
-    echo "http://www.speech.sri.com/projects/srilm/download.html"
-    echo "Then add the tools to your PATH"
-    exit 1
-fi
-
-# This script takes no arguments.  It assumes you have already run
-# aishell_data_prep.sh.
-# It takes as input the files
-# data/local/lm/text
-# data/local/dict/lexicon.txt
-dir=data/local/lm
-mkdir -p $dir
-
-
-cleantext=$dir/text.no_oov
-
-cat $text | awk -v lex=$lexicon 'BEGIN{while((getline<lex) >0){ seen[$1]=1; } }
-  {for(n=1; n<=NF;n++) {  if (seen[$n]) { printf("%s ", $n); } else {printf("<SPOKEN_NOISE> ");} } printf("\n");}' \
-  > $cleantext || exit 1;
-
-cat $cleantext | awk '{for(n=2;n<=NF;n++) print $n; }' | sort | uniq -c | \
-   sort -nr > $dir/word.counts || exit 1;
-
-# Get counts from acoustic training transcripts, and add  one-count
-# for each word in the lexicon (but not silence, we don't want it
-# in the LM-- we'll add it optionally later).
-cat $cleantext | awk '{for(n=2;n<=NF;n++) print $n; }' | \
-  cat - <(grep -w -v '!SIL' $lexicon | awk '{print $1}') | \
-   sort | uniq -c | sort -nr > $dir/unigram.counts || exit 1;
-
-cat $dir/unigram.counts | awk '{print $2}' | cat - <(echo "<s>"; echo "</s>" ) > $dir/wordlist
-
-heldout_sent=10000 # Don't change this if you want result to be comparable with
-    # kaldi_lm results
-mkdir -p $dir
-cat $cleantext | awk '{for(n=2;n<=NF;n++){ printf $n; if(n<NF) printf " "; else print ""; }}' | \
-  head -$heldout_sent > $dir/heldout
-cat $cleantext | awk '{for(n=2;n<=NF;n++){ printf $n; if(n<NF) printf " "; else print ""; }}' | \
-  tail -n +$heldout_sent > $dir/train
-
-ngram-count -text $dir/train -order 3 -limit-vocab -vocab $dir/wordlist -unk \
-  -map-unk "<UNK>" -kndiscount -interpolate -lm $dir/lm.arpa
-ngram -lm $dir/lm.arpa -ppl $dir/heldout
diff --git a/speechx/examples/aishell/mkgraph.sh b/speechx/examples/aishell/mkgraph.sh
deleted file mode 100644
index f66cd4dc..00000000
--- a/speechx/examples/aishell/mkgraph.sh
+++ /dev/null
@@ -1,31 +0,0 @@
-#!/bin/bash
-
-. ./path.sh || exit 1;
-
-. tools/parse_options.sh || exit 1;
-
-data=/mnt/dataset/aishell
-
-# Optionally, you can add LM and test it with runtime.
-dir=./ds2_graph
-dict=$dir/vocab.txt
-if [ ${stage} -le 7 ] && [ ${stop_stage} -ge 7 ]; then
-  # 7.1 Prepare dict
-  unit_file=$dict
-  mkdir -p $dir/local/dict
-  cp $unit_file $dir/local/dict/units.txt
-  tools/fst/prepare_dict.py $unit_file ${data}/resource_aishell/lexicon.txt \
-    $dir/local/dict/lexicon.txt
-  # Train lm
-  lm=$dir/local/lm
-  mkdir -p $lm
-  tools/filter_scp.pl data/train/text \
-    $data/data_aishell/transcript/aishell_transcript_v0.8.txt > $lm/text
-  local/ds2_aishell_train_lms.sh
-  # Build decoding TLG
-  tools/fst/compile_lexicon_token_fst.sh \
-    $dir/local/dict $dir/local/tmp $dir/local/lang
-  tools/fst/make_tlg.sh $dir/local/lm $dir/local/lang $dir/lang_test || exit 1;
-fi
-
-
diff --git a/speechx/examples/aishell/path.sh b/speechx/examples/aishell/path.sh
deleted file mode 100644
index 8ab7ee29..00000000
--- a/speechx/examples/aishell/path.sh
+++ /dev/null
@@ -1,14 +0,0 @@
-# This contains the locations of binarys build required for running the examples.
-
-SPEECHX_ROOT=$PWD/../..
-SPEECHX_EXAMPLES=$SPEECHX_ROOT/build/examples
-
-SPEECHX_TOOLS=$SPEECHX_ROOT/tools
-TOOLS_BIN=$SPEECHX_TOOLS/valgrind/install/bin
-
-[ -d $SPEECHX_EXAMPLES ] || { echo "Error: 'build/examples' directory not found. please ensure that the project build successfully"; }
-
-export LC_AL=C
-
-SPEECHX_BIN=$SPEECHX_EXAMPLES/feat
-export PATH=$PATH:$SPEECHX_BIN:$TOOLS_BIN

From 642e0840b4e3d9affcb6ca5a07b3c6308546cc84 Mon Sep 17 00:00:00 2001
From: Yang Zhou <goat.zhou@qq.com>
Date: Thu, 7 Apr 2022 15:08:01 +0800
Subject: [PATCH 5/9] dev ing

---
 .../decoder/offline_wfst_decoder_main.cc        |  9 +++++----
 speechx/speechx/nnet/decodable.cc               | 17 ++++++++++++++---
 2 files changed, 19 insertions(+), 7 deletions(-)

diff --git a/speechx/examples/decoder/offline_wfst_decoder_main.cc b/speechx/examples/decoder/offline_wfst_decoder_main.cc
index f0b9cc4f..758942b5 100644
--- a/speechx/examples/decoder/offline_wfst_decoder_main.cc
+++ b/speechx/examples/decoder/offline_wfst_decoder_main.cc
@@ -17,7 +17,7 @@
 #include "base/flags.h"
 #include "base/log.h"
 #include "decoder/ctc_tlg_decoder.h"
-#include "frontend/raw_audio.h"
+#include "frontend/audio/data_cache.h"
 #include "kaldi/util/table-types.h"
 #include "nnet/decodable.h"
 #include "nnet/paddle_nnet.h"
@@ -27,7 +27,7 @@ DEFINE_string(model_path, "avg_1.jit.pdmodel", "paddle nnet model");
 DEFINE_string(param_path, "avg_1.jit.pdiparams", "paddle nnet model param");
 DEFINE_string(word_symbol_table, "vocab.txt", "word symbol table");
 DEFINE_string(graph_path, "TLG", "decoder graph");
-DEFINE_double(acoustic_scale, 1.0, "acoustic scale");
+DEFINE_double(acoustic_scale, 10.0, "acoustic scale");
 DEFINE_int32(max_active, 5000, "decoder graph");
 
 
@@ -52,6 +52,7 @@ int main(int argc, char* argv[]) {
     opts.word_symbol_table = word_symbol_table;
     opts.fst_path = graph_path;
     opts.opts.max_active = FLAGS_max_active;
+    opts.opts.beam = 
     ppspeech::TLGDecoder decoder(opts);
 
     ppspeech::ModelOptions model_opts;
@@ -60,8 +61,8 @@ int main(int argc, char* argv[]) {
     model_opts.cache_shape = "5-1-1024,5-1-1024";
     std::shared_ptr<ppspeech::PaddleNnet> nnet(
         new ppspeech::PaddleNnet(model_opts));
-    std::shared_ptr<ppspeech::RawDataCache> raw_data(
-        new ppspeech::RawDataCache());
+    std::shared_ptr<ppspeech::DataCache> raw_data(
+        new ppspeech::DataCache());
     std::shared_ptr<ppspeech::Decodable> decodable(
         new ppspeech::Decodable(nnet, raw_data));
 
diff --git a/speechx/speechx/nnet/decodable.cc b/speechx/speechx/nnet/decodable.cc
index de15c4f8..805c0dca 100644
--- a/speechx/speechx/nnet/decodable.cc
+++ b/speechx/speechx/nnet/decodable.cc
@@ -48,7 +48,7 @@ BaseFloat Decodable::LogLikelihood(int32 frame, int32 index) {
     CHECK_LE(index, nnet_cache_.NumCols());
     CHECK_LE(frame, frames_ready_);
     int32 frame_idx = frame - frame_offset_;
-    return nnet_cache_(frame_idx, index);
+    return std::log(nnet_cache_(frame_idx, index) + std::numeric_limits<float>::min());
 }
 
 bool Decodable::EnsureFrameHaveComputed(int32 frame) {
@@ -65,9 +65,20 @@ bool Decodable::AdvanceChunk() {
     }
     int32 nnet_dim = 0;
     Vector<BaseFloat> inferences;
+    Matrix<BaseFloat> nnet_cache_tmp;
     nnet_->FeedForward(features, frontend_->Dim(), &inferences, &nnet_dim);
-    nnet_cache_.Resize(inferences.Dim() / nnet_dim, nnet_dim);
-    nnet_cache_.CopyRowsFromVec(inferences);
+    nnet_cache_tmp.Resize(inferences.Dim() / nnet_dim, nnet_dim);
+    nnet_cache_tmp.CopyRowsFromVec(inferences);
+    // skip blank
+    vector<int> no_blank_record;
+    BaseFloat blank_threshold = 0.98;
+    for (int32 idx = 0; idx < nnet_cache_.NumRows(); ++idx) {
+      if (nnet_cache_(idx, 0) > blank_threshold) {
+             
+      }
+    }
+    
+    
     frame_offset_ = frames_ready_;
     frames_ready_ += nnet_cache_.NumRows();
     return true;

From 2455d8899710ad5322fa0ebc2281e0e650cd80fb Mon Sep 17 00:00:00 2001
From: Yang Zhou <goat.zhou@qq.com>
Date: Mon, 11 Apr 2022 15:43:21 +0800
Subject: [PATCH 6/9] make wfst work & align frame

---
 .../decoder/offline_wfst_decoder_main.cc      | 56 ++++++++++++----
 speechx/speechx/decoder/ctc_tlg_decoder.cc    | 64 ++++++++++++-------
 speechx/speechx/decoder/ctc_tlg_decoder.h     | 42 +++++++-----
 speechx/speechx/nnet/decodable.cc             | 41 ++++++------
 speechx/speechx/nnet/decodable.h              |  8 +--
 5 files changed, 135 insertions(+), 76 deletions(-)

diff --git a/speechx/examples/decoder/offline_wfst_decoder_main.cc b/speechx/examples/decoder/offline_wfst_decoder_main.cc
index 758942b5..90dc8840 100644
--- a/speechx/examples/decoder/offline_wfst_decoder_main.cc
+++ b/speechx/examples/decoder/offline_wfst_decoder_main.cc
@@ -27,14 +27,20 @@ DEFINE_string(model_path, "avg_1.jit.pdmodel", "paddle nnet model");
 DEFINE_string(param_path, "avg_1.jit.pdiparams", "paddle nnet model param");
 DEFINE_string(word_symbol_table, "vocab.txt", "word symbol table");
 DEFINE_string(graph_path, "TLG", "decoder graph");
-DEFINE_double(acoustic_scale, 10.0, "acoustic scale");
-DEFINE_int32(max_active, 5000, "decoder graph");
-
+DEFINE_double(acoustic_scale, 1.0, "acoustic scale");
+DEFINE_int32(max_active, 7500, "decoder graph");
+DEFINE_int32(receptive_field_length,
+             7,
+             "receptive field of two CNN(kernel=5) downsampling module.");
+DEFINE_int32(downsampling_rate,
+             4,
+             "two CNN(kernel=5) module downsampling rate.");
 
 using kaldi::BaseFloat;
 using kaldi::Matrix;
 using std::vector;
 
+// test clg decoder by feeding speech feature.
 int main(int argc, char* argv[]) {
     gflags::ParseCommandLineFlags(&argc, &argv, false);
     google::InitGoogleLogging(argv[0]);
@@ -52,7 +58,8 @@ int main(int argc, char* argv[]) {
     opts.word_symbol_table = word_symbol_table;
     opts.fst_path = graph_path;
     opts.opts.max_active = FLAGS_max_active;
-    opts.opts.beam = 
+    opts.opts.beam = 15.0;
+    opts.opts.lattice_beam = 7.5;
     ppspeech::TLGDecoder decoder(opts);
 
     ppspeech::ModelOptions model_opts;
@@ -61,30 +68,55 @@ int main(int argc, char* argv[]) {
     model_opts.cache_shape = "5-1-1024,5-1-1024";
     std::shared_ptr<ppspeech::PaddleNnet> nnet(
         new ppspeech::PaddleNnet(model_opts));
-    std::shared_ptr<ppspeech::DataCache> raw_data(
-        new ppspeech::DataCache());
+    std::shared_ptr<ppspeech::DataCache> raw_data(new ppspeech::DataCache());
     std::shared_ptr<ppspeech::Decodable> decodable(
-        new ppspeech::Decodable(nnet, raw_data));
+        new ppspeech::Decodable(nnet, raw_data, FLAGS_acoustic_scale));
 
-    int32 chunk_size = 35;
+    int32 chunk_size = FLAGS_receptive_field_length;
+    int32 chunk_stride = FLAGS_downsampling_rate;
+    int32 receptive_field_length = FLAGS_receptive_field_length;
+    LOG(INFO) << "chunk size (frame): " << chunk_size;
+    LOG(INFO) << "chunk stride (frame): " << chunk_stride;
+    LOG(INFO) << "receptive field (frame): " << receptive_field_length;
     decoder.InitDecoder();
 
     for (; !feature_reader.Done(); feature_reader.Next()) {
         string utt = feature_reader.Key();
-        const kaldi::Matrix<BaseFloat> feature = feature_reader.Value();
+        kaldi::Matrix<BaseFloat> feature = feature_reader.Value();
         raw_data->SetDim(feature.NumCols());
+        LOG(INFO) << "process utt: " << utt;
+        LOG(INFO) << "rows: " << feature.NumRows();
+        LOG(INFO) << "cols: " << feature.NumCols();
+
         int32 row_idx = 0;
-        int32 num_chunks = feature.NumRows() / chunk_size;
+        int32 padding_len = 0;
+        int32 ori_feature_len = feature.NumRows();
+        if ((feature.NumRows() - chunk_size) % chunk_stride != 0) {
+            padding_len =
+                chunk_stride - (feature.NumRows() - chunk_size) % chunk_stride;
+            feature.Resize(feature.NumRows() + padding_len,
+                           feature.NumCols(),
+                           kaldi::kCopyData);
+        }
+        int32 num_chunks = (feature.NumRows() - chunk_size) / chunk_stride + 1;
         for (int chunk_idx = 0; chunk_idx < num_chunks; ++chunk_idx) {
             kaldi::Vector<kaldi::BaseFloat> feature_chunk(chunk_size *
                                                           feature.NumCols());
+            int32 feature_chunk_size = 0;
+            if (ori_feature_len > chunk_idx * chunk_stride) {
+                feature_chunk_size = std::min(
+                    ori_feature_len - chunk_idx * chunk_stride, chunk_size);
+            }
+            if (feature_chunk_size < receptive_field_length) break;
+
+            int32 start = chunk_idx * chunk_stride;
             for (int row_id = 0; row_id < chunk_size; ++row_id) {
-                kaldi::SubVector<kaldi::BaseFloat> tmp(feature, row_idx);
+                kaldi::SubVector<kaldi::BaseFloat> tmp(feature, start);
                 kaldi::SubVector<kaldi::BaseFloat> f_chunk_tmp(
                     feature_chunk.Data() + row_id * feature.NumCols(),
                     feature.NumCols());
                 f_chunk_tmp.CopyFromVec(tmp);
-                row_idx++;
+                ++start;
             }
             raw_data->Accept(feature_chunk);
             if (chunk_idx == num_chunks - 1) {
diff --git a/speechx/speechx/decoder/ctc_tlg_decoder.cc b/speechx/speechx/decoder/ctc_tlg_decoder.cc
index c08a7d5b..5365e709 100644
--- a/speechx/speechx/decoder/ctc_tlg_decoder.cc
+++ b/speechx/speechx/decoder/ctc_tlg_decoder.cc
@@ -1,50 +1,66 @@
+// Copyright (c) 2022 PaddlePaddle Authors. All Rights Reserved.
+//
+// Licensed under the Apache License, Version 2.0 (the "License");
+// you may not use this file except in compliance with the License.
+// You may obtain a copy of the License at
+//
+//     http://www.apache.org/licenses/LICENSE-2.0
+//
+// Unless required by applicable law or agreed to in writing, software
+// distributed under the License is distributed on an "AS IS" BASIS,
+// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+// See the License for the specific language governing permissions and
+// limitations under the License.
+
 #include "decoder/ctc_tlg_decoder.h"
 namespace ppspeech {
 
 TLGDecoder::TLGDecoder(TLGDecoderOptions opts) {
     fst_.reset(fst::Fst<fst::StdArc>::Read(opts.fst_path));
     CHECK(fst_ != nullptr);
-    word_symbol_table_.reset(fst::SymbolTable::ReadText(opts.word_symbol_table));
+    word_symbol_table_.reset(
+        fst::SymbolTable::ReadText(opts.word_symbol_table));
     decoder_.reset(new kaldi::LatticeFasterOnlineDecoder(*fst_, opts.opts));
     decoder_->InitDecoding();
+    frame_decoded_size_ = 0;
 }
 
 void TLGDecoder::InitDecoder() {
     decoder_->InitDecoding();
+    frame_decoded_size_ = 0;
 }
 
-void TLGDecoder::AdvanceDecode(const std::shared_ptr<kaldi::DecodableInterface>& decodable) {
-    while (1) {
-      AdvanceDecoding(decodable.get());
-      if (decodable->IsLastFrame(num_frame_decoded_)) break;
+void TLGDecoder::AdvanceDecode(
+    const std::shared_ptr<kaldi::DecodableInterface>& decodable) {
+    while (!decodable->IsLastFrame(frame_decoded_size_)) {
+        LOG(INFO) << "num frame decode: " << frame_decoded_size_;
+        AdvanceDecoding(decodable.get());
     }
 }
 
 void TLGDecoder::AdvanceDecoding(kaldi::DecodableInterface* decodable) {
-  // skip blank frame?
-  decoder_->AdvanceDecoding(decodable, 1);
-  num_frame_decoded_++;
+    decoder_->AdvanceDecoding(decodable, 1);
+    frame_decoded_size_++;
 }
 
 void TLGDecoder::Reset() {
-  decoder_->InitDecoding();
-  return;
+    InitDecoder();
+    return;
 }
 
 std::string TLGDecoder::GetFinalBestPath() {
-  decoder_->FinalizeDecoding();
-  kaldi::Lattice lat;
-  kaldi::LatticeWeight weight;
-  std::vector<int> alignment;
-  std::vector<int> words_id;
-  decoder_->GetBestPath(&lat, true);
-  fst::GetLinearSymbolSequence(lat, &alignment, &words_id, &weight);
-  std::string words;
-  for (int32 idx = 0; idx < words_id.size(); ++idx) {
-    std::string word = word_symbol_table_->Find(words_id[idx]);
-    words += word;
-  }
-  return words;
+    decoder_->FinalizeDecoding();
+    kaldi::Lattice lat;
+    kaldi::LatticeWeight weight;
+    std::vector<int> alignment;
+    std::vector<int> words_id;
+    decoder_->GetBestPath(&lat, true);
+    fst::GetLinearSymbolSequence(lat, &alignment, &words_id, &weight);
+    std::string words;
+    for (int32 idx = 0; idx < words_id.size(); ++idx) {
+        std::string word = word_symbol_table_->Find(words_id[idx]);
+        words += word;
+    }
+    return words;
 }
-
 }
\ No newline at end of file
diff --git a/speechx/speechx/decoder/ctc_tlg_decoder.h b/speechx/speechx/decoder/ctc_tlg_decoder.h
index b4cd8c34..361c44af 100644
--- a/speechx/speechx/decoder/ctc_tlg_decoder.h
+++ b/speechx/speechx/decoder/ctc_tlg_decoder.h
@@ -1,21 +1,33 @@
+// Copyright (c) 2022 PaddlePaddle Authors. All Rights Reserved.
+//
+// Licensed under the Apache License, Version 2.0 (the "License");
+// you may not use this file except in compliance with the License.
+// You may obtain a copy of the License at
+//
+//     http://www.apache.org/licenses/LICENSE-2.0
+//
+// Unless required by applicable law or agreed to in writing, software
+// distributed under the License is distributed on an "AS IS" BASIS,
+// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+// See the License for the specific language governing permissions and
+// limitations under the License.
+
 #pragma once
 
-#include "kaldi/decoder/lattice-faster-online-decoder.h"
+#include "base/basic_types.h"
 #include "kaldi/decoder/decodable-itf.h"
+#include "kaldi/decoder/lattice-faster-online-decoder.h"
 #include "util/parse-options.h"
-#include "base/basic_types.h"
 
 namespace ppspeech {
 
 struct TLGDecoderOptions {
-   kaldi::LatticeFasterDecoderConfig opts; 
-   // todo remove later, add into decode resource 
-   std::string word_symbol_table;
-   std::string fst_path;
-   
-   TLGDecoderOptions() 
-       : word_symbol_table(""),
-         fst_path("") {}
+    kaldi::LatticeFasterDecoderConfig opts;
+    // todo remove later, add into decode resource
+    std::string word_symbol_table;
+    std::string fst_path;
+
+    TLGDecoderOptions() : word_symbol_table(""), fst_path("") {}
 };
 
 class TLGDecoder {
@@ -34,14 +46,14 @@ class TLGDecoder {
     void Reset();
 
   private:
-    void AdvanceDecoding(kaldi::DecodableInterface* decodable);    
+    void AdvanceDecoding(kaldi::DecodableInterface* decodable);
 
     std::shared_ptr<kaldi::LatticeFasterOnlineDecoder> decoder_;
-    std::shared_ptr<fst::Fst<fst::StdArc>> fst_;   
+    std::shared_ptr<fst::Fst<fst::StdArc>> fst_;
     std::shared_ptr<fst::SymbolTable> word_symbol_table_;
-    int32 num_frame_decoded_;
-  };
+    // the frame size which have decoded starts from 0.
+    int32 frame_decoded_size_;
+};
 
-    
 
 }  // namespace ppspeech
\ No newline at end of file
diff --git a/speechx/speechx/nnet/decodable.cc b/speechx/speechx/nnet/decodable.cc
index 805c0dca..1bf870d8 100644
--- a/speechx/speechx/nnet/decodable.cc
+++ b/speechx/speechx/nnet/decodable.cc
@@ -22,8 +22,13 @@ using std::vector;
 using kaldi::Vector;
 
 Decodable::Decodable(const std::shared_ptr<NnetInterface>& nnet,
-                     const std::shared_ptr<FrontendInterface>& frontend)
-    : frontend_(frontend), nnet_(nnet), frame_offset_(0), frames_ready_(0) {}
+                     const std::shared_ptr<FrontendInterface>& frontend,
+                     kaldi::BaseFloat acoustic_scale)
+    : frontend_(frontend),
+      nnet_(nnet),
+      frame_offset_(0),
+      frames_ready_(0),
+      acoustic_scale_(acoustic_scale) {}
 
 void Decodable::Acceptlikelihood(const Matrix<BaseFloat>& likelihood) {
     nnet_cache_ = likelihood;
@@ -32,14 +37,14 @@ void Decodable::Acceptlikelihood(const Matrix<BaseFloat>& likelihood) {
 
 // Decodable::Init(DecodableConfig config) {
 //}
-int32 Decodable::NumFramesReady() const {
-    return frames_ready_;
-}
 
+// return the size of frame have computed.
+int32 Decodable::NumFramesReady() const { return frames_ready_; }
+
+// frame idx is from 0 to frame_ready_ -1;
 bool Decodable::IsLastFrame(int32 frame) {
     bool flag = EnsureFrameHaveComputed(frame);
-    //CHECK_LE(frame, frames_ready_);
-    return (flag == false) && (frame == frames_ready_);
+    return frame >= frames_ready_;
 }
 
 int32 Decodable::NumIndices() const { return 0; }
@@ -48,7 +53,8 @@ BaseFloat Decodable::LogLikelihood(int32 frame, int32 index) {
     CHECK_LE(index, nnet_cache_.NumCols());
     CHECK_LE(frame, frames_ready_);
     int32 frame_idx = frame - frame_offset_;
-    return std::log(nnet_cache_(frame_idx, index) + std::numeric_limits<float>::min());
+    return acoustic_scale_ * std::log(nnet_cache_(frame_idx, index - 1) +
+                                      std::numeric_limits<float>::min());
 }
 
 bool Decodable::EnsureFrameHaveComputed(int32 frame) {
@@ -67,20 +73,12 @@ bool Decodable::AdvanceChunk() {
     Vector<BaseFloat> inferences;
     Matrix<BaseFloat> nnet_cache_tmp;
     nnet_->FeedForward(features, frontend_->Dim(), &inferences, &nnet_dim);
-    nnet_cache_tmp.Resize(inferences.Dim() / nnet_dim, nnet_dim);
-    nnet_cache_tmp.CopyRowsFromVec(inferences);
-    // skip blank
-    vector<int> no_blank_record;
-    BaseFloat blank_threshold = 0.98;
-    for (int32 idx = 0; idx < nnet_cache_.NumRows(); ++idx) {
-      if (nnet_cache_(idx, 0) > blank_threshold) {
-             
-      }
-    }
-    
-    
+    nnet_cache_.Resize(inferences.Dim() / nnet_dim, nnet_dim);
+    nnet_cache_.CopyRowsFromVec(inferences);
+
     frame_offset_ = frames_ready_;
     frames_ready_ += nnet_cache_.NumRows();
+    LOG(INFO) << "nnet size: " << nnet_cache_.NumRows();
     return true;
 }
 
@@ -89,7 +87,8 @@ bool Decodable::FrameLogLikelihood(int32 frame, vector<BaseFloat>* likelihood) {
     if (EnsureFrameHaveComputed(frame) == false) return false;
     likelihood->resize(nnet_cache_.NumCols());
     for (int32 idx = 0; idx < nnet_cache_.NumCols(); ++idx) {
-        (*likelihood)[idx] = nnet_cache_(frame - frame_offset_, idx);
+        (*likelihood)[idx] =
+            nnet_cache_(frame - frame_offset_, idx) * acoustic_scale_;
     }
     return true;
 }
diff --git a/speechx/speechx/nnet/decodable.h b/speechx/speechx/nnet/decodable.h
index 5b687f3d..b18ef07c 100644
--- a/speechx/speechx/nnet/decodable.h
+++ b/speechx/speechx/nnet/decodable.h
@@ -14,8 +14,8 @@
 
 #include "base/common.h"
 #include "frontend/audio/frontend_itf.h"
-#include "kaldi/matrix/kaldi-matrix.h"
 #include "kaldi/decoder/decodable-itf.h"
+#include "kaldi/matrix/kaldi-matrix.h"
 #include "nnet/nnet_itf.h"
 
 namespace ppspeech {
@@ -25,7 +25,8 @@ struct DecodableOpts;
 class Decodable : public kaldi::DecodableInterface {
   public:
     explicit Decodable(const std::shared_ptr<NnetInterface>& nnet,
-                       const std::shared_ptr<FrontendInterface>& frontend);
+                       const std::shared_ptr<FrontendInterface>& frontend,
+                       kaldi::BaseFloat acoustic_scale = 1.0);
     // void Init(DecodableOpts config);
     virtual kaldi::BaseFloat LogLikelihood(int32 frame, int32 index);
     virtual bool IsLastFrame(int32 frame);
@@ -38,14 +39,12 @@ class Decodable : public kaldi::DecodableInterface {
     void Reset();
     bool IsInputFinished() const { return frontend_->IsFinished(); }
     bool EnsureFrameHaveComputed(int32 frame);
-    
 
   private:
     bool AdvanceChunk();
     std::shared_ptr<FrontendInterface> frontend_;
     std::shared_ptr<NnetInterface> nnet_;
     kaldi::Matrix<kaldi::BaseFloat> nnet_cache_;
-    // std::vector<std::vector<kaldi::BaseFloat>> nnet_cache_;
     int32 frame_offset_;
     int32 frames_ready_;
     // todo: feature frame mismatch with nnet inference frame
@@ -53,6 +52,7 @@ class Decodable : public kaldi::DecodableInterface {
     // so use subsampled_frame
     int32 current_log_post_subsampled_offset_;
     int32 num_chunk_computed_;
+    kaldi::BaseFloat acoustic_scale_;
 };
 
 }  // namespace ppspeech

From 90d6b6f15d06580587cc4bf08f2a5e26b8bd46d4 Mon Sep 17 00:00:00 2001
From: Yang Zhou <goat.zhou@qq.com>
Date: Mon, 11 Apr 2022 18:49:28 +0800
Subject: [PATCH 7/9] add aishell wfst eg script

---
 speechx/examples/aishell/run.sh               | 42 ++++++++++++++++---
 .../offline_decoder_sliding_chunk_main.cc     |  8 ++--
 .../decoder/offline_wfst_decoder_main.cc      | 28 ++++++++++---
 speechx/speechx/nnet/paddle_nnet.cc           |  6 ---
 4 files changed, 65 insertions(+), 19 deletions(-)

diff --git a/speechx/examples/aishell/run.sh b/speechx/examples/aishell/run.sh
index a21ba086..8a16a865 100755
--- a/speechx/examples/aishell/run.sh
+++ b/speechx/examples/aishell/run.sh
@@ -48,7 +48,7 @@ wer=./aishell_wer
 nj=40
 export GLOG_logtostderr=1
 
-./local/split_data.sh $data $data/$aishell_wav_scp $aishell_wav_scp $nj
+#./local/split_data.sh $data $data/$aishell_wav_scp $aishell_wav_scp $nj
 
 data=$PWD/data
 # 3. gen linear feat
@@ -72,10 +72,42 @@ utils/run.pl JOB=1:$nj $data/split${nj}/JOB/log \
     --param_path=$aishell_online_model/avg_1.jit.pdiparams \
     --model_output_names=softmax_0.tmp_0,tmp_5,concat_0.tmp_0,concat_1.tmp_0 \
     --dict_file=$lm_model_dir/vocab.txt \
-    --lm_path=$lm_model_dir/avg_1.jit.klm \
     --result_wspecifier=ark,t:$data/split${nj}/JOB/result
 
-cat $data/split${nj}/*/result > $label_file
+cat $data/split${nj}/*/result > ${label_file}
+local/compute-wer.py --char=1 --v=1 ${label_file} $text > ${wer}
+
+# 4. decode with lm
+utils/run.pl JOB=1:$nj $data/split${nj}/JOB/log_lm \
+  offline_decoder_sliding_chunk_main \
+    --feature_rspecifier=scp:$data/split${nj}/JOB/feat.scp \
+    --model_path=$aishell_online_model/avg_1.jit.pdmodel \
+    --param_path=$aishell_online_model/avg_1.jit.pdiparams \
+    --model_output_names=softmax_0.tmp_0,tmp_5,concat_0.tmp_0,concat_1.tmp_0 \
+    --dict_file=$lm_model_dir/vocab.txt \
+    --lm_path=$lm_model_dir/avg_1.jit.klm \
+    --result_wspecifier=ark,t:$data/split${nj}/JOB/result_lm
+
+cat $data/split${nj}/*/result_lm > ${label_file}_lm
+local/compute-wer.py --char=1 --v=1 ${label_file}_lm $text > ${wer}_lm
+
+graph_dir=./aishell_graph
+if [ ! -d $ ]; then
+    wget -c https://paddlespeech.bj.bcebos.com/s2t/paddle_asr_online/aishell_graph.zip
+    unzip -d aishell_graph.zip
+fi
+
+# 5. test TLG decoder
+utils/run.pl JOB=1:$nj $data/split${nj}/JOB/log_tlg \
+  offline_wfst_decoder_main \
+    --feature_rspecifier=scp:$data/split${nj}/JOB/feat.scp \
+    --model_path=$aishell_online_model/avg_1.jit.pdmodel \
+    --param_path=$aishell_online_model/avg_1.jit.pdiparams \
+    --word_symbol_table=$graph_dir/words.txt \
+    --model_output_names=softmax_0.tmp_0,tmp_5,concat_0.tmp_0,concat_1.tmp_0 \
+     --graph_path=$graph_dir/TLG.fst --max_active=7500 \
+    --acoustic_scale=1.2 \
+    --result_wspecifier=ark,t:$data/split${nj}/JOB/result_tlg
 
-local/compute-wer.py --char=1 --v=1 $label_file $text > $wer
-tail $wer
+cat $data/split${nj}/*/result_tlg > ${label_file}_tlg
+local/compute-wer.py --char=1 --v=1 ${label_file}_tlg $text > ${wer}_tlg
\ No newline at end of file
diff --git a/speechx/examples/decoder/offline_decoder_sliding_chunk_main.cc b/speechx/examples/decoder/offline_decoder_sliding_chunk_main.cc
index be56342f..1bb9ce22 100644
--- a/speechx/examples/decoder/offline_decoder_sliding_chunk_main.cc
+++ b/speechx/examples/decoder/offline_decoder_sliding_chunk_main.cc
@@ -27,7 +27,7 @@ DEFINE_string(result_wspecifier, "", "test result wspecifier");
 DEFINE_string(model_path, "avg_1.jit.pdmodel", "paddle nnet model");
 DEFINE_string(param_path, "avg_1.jit.pdiparams", "paddle nnet model param");
 DEFINE_string(dict_file, "vocab.txt", "vocabulary of lm");
-DEFINE_string(lm_path, "lm.klm", "language model");
+DEFINE_string(lm_path, "", "language model");
 DEFINE_int32(receptive_field_length,
              7,
              "receptive field of two CNN(kernel=5) downsampling module.");
@@ -45,7 +45,6 @@ using kaldi::BaseFloat;
 using kaldi::Matrix;
 using std::vector;
 
-
 // test ds2 online decoder by feeding speech feature
 int main(int argc, char* argv[]) {
     gflags::ParseCommandLineFlags(&argc, &argv, false);
@@ -63,7 +62,6 @@ int main(int argc, char* argv[]) {
     LOG(INFO) << "dict path: " << dict_file;
     LOG(INFO) << "lm path: " << lm_path;
 
-
     int32 num_done = 0, num_err = 0;
 
     ppspeech::CTCBeamSearchOptions opts;
@@ -139,6 +137,10 @@ int main(int argc, char* argv[]) {
         std::string result;
         result = decoder.GetFinalBestPath();
         KALDI_LOG << " the result of " << utt << " is " << result;
+        if (result.empty()) {
+            // the TokenWriter can not write empty string.
+            result = " ";
+        }
         result_writer.Write(utt, result);
         decodable->Reset();
         decoder.Reset();
diff --git a/speechx/examples/decoder/offline_wfst_decoder_main.cc b/speechx/examples/decoder/offline_wfst_decoder_main.cc
index 90dc8840..2520922a 100644
--- a/speechx/examples/decoder/offline_wfst_decoder_main.cc
+++ b/speechx/examples/decoder/offline_wfst_decoder_main.cc
@@ -22,10 +22,11 @@
 #include "nnet/decodable.h"
 #include "nnet/paddle_nnet.h"
 
-DEFINE_string(feature_respecifier, "", "test feature rspecifier");
+DEFINE_string(feature_rspecifier, "", "test feature rspecifier");
+DEFINE_string(result_wspecifier, "", "test result wspecifier");
 DEFINE_string(model_path, "avg_1.jit.pdmodel", "paddle nnet model");
 DEFINE_string(param_path, "avg_1.jit.pdiparams", "paddle nnet model param");
-DEFINE_string(word_symbol_table, "vocab.txt", "word symbol table");
+DEFINE_string(word_symbol_table, "words.txt", "word symbol table");
 DEFINE_string(graph_path, "TLG", "decoder graph");
 DEFINE_double(acoustic_scale, 1.0, "acoustic scale");
 DEFINE_int32(max_active, 7500, "decoder graph");
@@ -35,22 +36,33 @@ DEFINE_int32(receptive_field_length,
 DEFINE_int32(downsampling_rate,
              4,
              "two CNN(kernel=5) module downsampling rate.");
+DEFINE_string(model_output_names,
+              "save_infer_model/scale_0.tmp_1,save_infer_model/"
+              "scale_1.tmp_1,save_infer_model/scale_2.tmp_1,save_infer_model/"
+              "scale_3.tmp_1",
+              "model output names");
+DEFINE_string(model_cache_names, "5-1-1024,5-1-1024", "model cache names");
 
 using kaldi::BaseFloat;
 using kaldi::Matrix;
 using std::vector;
 
-// test clg decoder by feeding speech feature.
+// test TLG decoder by feeding speech feature.
 int main(int argc, char* argv[]) {
     gflags::ParseCommandLineFlags(&argc, &argv, false);
     google::InitGoogleLogging(argv[0]);
 
     kaldi::SequentialBaseFloatMatrixReader feature_reader(
-        FLAGS_feature_respecifier);
+        FLAGS_feature_rspecifier);
+    kaldi::TokenWriter result_writer(FLAGS_result_wspecifier);
     std::string model_graph = FLAGS_model_path;
     std::string model_params = FLAGS_param_path;
     std::string word_symbol_table = FLAGS_word_symbol_table;
     std::string graph_path = FLAGS_graph_path;
+    LOG(INFO) << "model path: " << model_graph;
+    LOG(INFO) << "model param: " << model_params;
+    LOG(INFO) << "word symbol path: " << word_symbol_table;
+    LOG(INFO) << "graph path: " << graph_path;
 
     int32 num_done = 0, num_err = 0;
 
@@ -65,7 +77,8 @@ int main(int argc, char* argv[]) {
     ppspeech::ModelOptions model_opts;
     model_opts.model_path = model_graph;
     model_opts.params_path = model_params;
-    model_opts.cache_shape = "5-1-1024,5-1-1024";
+    model_opts.cache_shape = FLAGS_model_cache_names;
+    model_opts.output_names = FLAGS_model_output_names;
     std::shared_ptr<ppspeech::PaddleNnet> nnet(
         new ppspeech::PaddleNnet(model_opts));
     std::shared_ptr<ppspeech::DataCache> raw_data(new ppspeech::DataCache());
@@ -127,6 +140,11 @@ int main(int argc, char* argv[]) {
         std::string result;
         result = decoder.GetFinalBestPath();
         KALDI_LOG << " the result of " << utt << " is " << result;
+         if (result.empty()) {
+            // the TokenWriter can not write empty string.
+            result = " ";
+        }
+        result_writer.Write(utt, result);
         decodable->Reset();
         decoder.Reset();
         ++num_done;
diff --git a/speechx/speechx/nnet/paddle_nnet.cc b/speechx/speechx/nnet/paddle_nnet.cc
index c4b91cf6..c065d530 100644
--- a/speechx/speechx/nnet/paddle_nnet.cc
+++ b/speechx/speechx/nnet/paddle_nnet.cc
@@ -94,7 +94,6 @@ PaddleNnet::PaddleNnet(const ModelOptions& opts) : opts_(opts) {
 void PaddleNnet::Reset() { InitCacheEncouts(opts_); }
 
 paddle_infer::Predictor* PaddleNnet::GetPredictor() {
-    LOG(INFO) << "attempt to get a new predictor instance " << std::endl;
     paddle_infer::Predictor* predictor = nullptr;
     std::lock_guard<std::mutex> guard(pool_mutex);
     int pred_id = 0;
@@ -110,7 +109,6 @@ paddle_infer::Predictor* PaddleNnet::GetPredictor() {
     if (predictor) {
         pool_usages[pred_id] = true;
         predictor_to_thread_id[predictor] = pred_id;
-        LOG(INFO) << pred_id << " predictor create success";
     } else {
         LOG(INFO) << "Failed to get predictor from pool !!!";
     }
@@ -119,7 +117,6 @@ paddle_infer::Predictor* PaddleNnet::GetPredictor() {
 }
 
 int PaddleNnet::ReleasePredictor(paddle_infer::Predictor* predictor) {
-    LOG(INFO) << "attempt to releae a predictor";
     std::lock_guard<std::mutex> guard(pool_mutex);
     auto iter = predictor_to_thread_id.find(predictor);
 
@@ -128,10 +125,8 @@ int PaddleNnet::ReleasePredictor(paddle_infer::Predictor* predictor) {
         return 0;
     }
 
-    LOG(INFO) << iter->second << " predictor will be release";
     pool_usages[iter->second] = false;
     predictor_to_thread_id.erase(predictor);
-    LOG(INFO) << "release success";
     return 0;
 }
 
@@ -152,7 +147,6 @@ void PaddleNnet::FeedForward(const Vector<BaseFloat>& features,
     int feat_row = features.Dim() / feature_dim;
     std::vector<std::string> input_names = predictor->GetInputNames();
     std::vector<std::string> output_names = predictor->GetOutputNames();
-    LOG(INFO) << "feat info: rows, cols: " << feat_row << ", " << feature_dim;
 
     std::unique_ptr<paddle_infer::Tensor> input_tensor =
         predictor->GetInputHandle(input_names[0]);

From 1f23c4bd24e70c2e3939f4e631c886459bd3d135 Mon Sep 17 00:00:00 2001
From: Yang Zhou <goat.zhou@qq.com>
Date: Mon, 11 Apr 2022 21:20:14 +0800
Subject: [PATCH 8/9] remove unnecessary log

---
 .../decoder/offline_decoder_sliding_chunk_main.cc    | 10 ++++++----
 .../examples/decoder/offline_wfst_decoder_main.cc    | 12 +++++++-----
 speechx/speechx/nnet/decodable.cc                    |  1 -
 speechx/speechx/nnet/paddle_nnet.cc                  |  1 -
 4 files changed, 13 insertions(+), 11 deletions(-)

diff --git a/speechx/examples/decoder/offline_decoder_sliding_chunk_main.cc b/speechx/examples/decoder/offline_decoder_sliding_chunk_main.cc
index 1bb9ce22..40092de3 100644
--- a/speechx/examples/decoder/offline_decoder_sliding_chunk_main.cc
+++ b/speechx/examples/decoder/offline_decoder_sliding_chunk_main.cc
@@ -136,14 +136,16 @@ int main(int argc, char* argv[]) {
         }
         std::string result;
         result = decoder.GetFinalBestPath();
-        KALDI_LOG << " the result of " << utt << " is " << result;
+        decodable->Reset();
+        decoder.Reset();
         if (result.empty()) {
             // the TokenWriter can not write empty string.
-            result = " ";
+            ++num_err;
+            KALDI_LOG << " the result of " << utt << " is empty";
+            continue;
         }
+        KALDI_LOG << " the result of " << utt << " is " << result;
         result_writer.Write(utt, result);
-        decodable->Reset();
-        decoder.Reset();
         ++num_done;
     }
 
diff --git a/speechx/examples/decoder/offline_wfst_decoder_main.cc b/speechx/examples/decoder/offline_wfst_decoder_main.cc
index 2520922a..06460a45 100644
--- a/speechx/examples/decoder/offline_wfst_decoder_main.cc
+++ b/speechx/examples/decoder/offline_wfst_decoder_main.cc
@@ -139,14 +139,16 @@ int main(int argc, char* argv[]) {
         }
         std::string result;
         result = decoder.GetFinalBestPath();
-        KALDI_LOG << " the result of " << utt << " is " << result;
-         if (result.empty()) {
+        decodable->Reset();
+        decoder.Reset();
+        if (result.empty()) {
             // the TokenWriter can not write empty string.
-            result = " ";
+            ++num_err;
+            KALDI_LOG << " the result of " << utt << " is empty";
+            continue;
         }
+        KALDI_LOG << " the result of " << utt << " is " << result;
         result_writer.Write(utt, result);
-        decodable->Reset();
-        decoder.Reset();
         ++num_done;
     }
 
diff --git a/speechx/speechx/nnet/decodable.cc b/speechx/speechx/nnet/decodable.cc
index 1bf870d8..ce269650 100644
--- a/speechx/speechx/nnet/decodable.cc
+++ b/speechx/speechx/nnet/decodable.cc
@@ -78,7 +78,6 @@ bool Decodable::AdvanceChunk() {
 
     frame_offset_ = frames_ready_;
     frames_ready_ += nnet_cache_.NumRows();
-    LOG(INFO) << "nnet size: " << nnet_cache_.NumRows();
     return true;
 }
 
diff --git a/speechx/speechx/nnet/paddle_nnet.cc b/speechx/speechx/nnet/paddle_nnet.cc
index c065d530..5c4da11a 100644
--- a/speechx/speechx/nnet/paddle_nnet.cc
+++ b/speechx/speechx/nnet/paddle_nnet.cc
@@ -177,7 +177,6 @@ void PaddleNnet::FeedForward(const Vector<BaseFloat>& features,
         LOG(INFO) << "predictor run occurs error";
     }
 
-    LOG(INFO) << "get the model success";
     std::unique_ptr<paddle_infer::Tensor> h_out =
         predictor->GetOutputHandle(output_names[2]);
     assert(h_cache->get_shape() == h_out->shape());

From 3456ae4a516aa192e86bf1391e18f734170c6cf2 Mon Sep 17 00:00:00 2001
From: Yang Zhou <goat.zhou@qq.com>
Date: Tue, 12 Apr 2022 12:07:57 +0800
Subject: [PATCH 9/9] add log & rename LogFrameLikelihood

---
 speechx/speechx/decoder/ctc_beam_search_decoder.cc |  2 +-
 speechx/speechx/kaldi/decoder/decodable-itf.h      |  2 +-
 speechx/speechx/nnet/decodable.cc                  | 11 +++++++++--
 speechx/speechx/nnet/decodable.h                   | 10 +++++++---
 4 files changed, 18 insertions(+), 7 deletions(-)

diff --git a/speechx/speechx/decoder/ctc_beam_search_decoder.cc b/speechx/speechx/decoder/ctc_beam_search_decoder.cc
index 5d7a4f77..b4caa8e7 100644
--- a/speechx/speechx/decoder/ctc_beam_search_decoder.cc
+++ b/speechx/speechx/decoder/ctc_beam_search_decoder.cc
@@ -93,7 +93,7 @@ void CTCBeamSearch::AdvanceDecode(
         vector<vector<BaseFloat>> likelihood;
         vector<BaseFloat> frame_prob;
         bool flag =
-            decodable->FrameLogLikelihood(num_frame_decoded_, &frame_prob);
+            decodable->FrameLikelihood(num_frame_decoded_, &frame_prob);
         if (flag == false) break;
         likelihood.push_back(frame_prob);
         AdvanceDecoding(likelihood);
diff --git a/speechx/speechx/kaldi/decoder/decodable-itf.h b/speechx/speechx/kaldi/decoder/decodable-itf.h
index 19e07498..b8ce9143 100644
--- a/speechx/speechx/kaldi/decoder/decodable-itf.h
+++ b/speechx/speechx/kaldi/decoder/decodable-itf.h
@@ -143,7 +143,7 @@ class DecodableInterface {
     /// this is for compatibility with OpenFst).
     virtual int32 NumIndices() const = 0;
 
-    virtual bool FrameLogLikelihood(
+    virtual bool FrameLikelihood(
         int32 frame, std::vector<kaldi::BaseFloat>* likelihood) = 0;
 
 
diff --git a/speechx/speechx/nnet/decodable.cc b/speechx/speechx/nnet/decodable.cc
index ce269650..d52b249f 100644
--- a/speechx/speechx/nnet/decodable.cc
+++ b/speechx/speechx/nnet/decodable.cc
@@ -49,11 +49,18 @@ bool Decodable::IsLastFrame(int32 frame) {
 
 int32 Decodable::NumIndices() const { return 0; }
 
+// the ilable(TokenId) of wfst(TLG) insert <eps>(id = 0) in front of Nnet prob id.
+int32 Decodable::TokenId2NnetId(int32 token_id) {
+   return token_id - 1;
+}
+
 BaseFloat Decodable::LogLikelihood(int32 frame, int32 index) {
     CHECK_LE(index, nnet_cache_.NumCols());
     CHECK_LE(frame, frames_ready_);
     int32 frame_idx = frame - frame_offset_;
-    return acoustic_scale_ * std::log(nnet_cache_(frame_idx, index - 1) +
+    // the nnet output is prob ranther than log prob
+    // the index - 1, because the ilabel 
+    return acoustic_scale_ * std::log(nnet_cache_(frame_idx, TokenId2NnetId(index)) +
                                       std::numeric_limits<float>::min());
 }
 
@@ -81,7 +88,7 @@ bool Decodable::AdvanceChunk() {
     return true;
 }
 
-bool Decodable::FrameLogLikelihood(int32 frame, vector<BaseFloat>* likelihood) {
+bool Decodable::FrameLikelihood(int32 frame, vector<BaseFloat>* likelihood) {
     std::vector<BaseFloat> result;
     if (EnsureFrameHaveComputed(frame) == false) return false;
     likelihood->resize(nnet_cache_.NumCols());
diff --git a/speechx/speechx/nnet/decodable.h b/speechx/speechx/nnet/decodable.h
index b18ef07c..9555fea7 100644
--- a/speechx/speechx/nnet/decodable.h
+++ b/speechx/speechx/nnet/decodable.h
@@ -31,24 +31,28 @@ class Decodable : public kaldi::DecodableInterface {
     virtual kaldi::BaseFloat LogLikelihood(int32 frame, int32 index);
     virtual bool IsLastFrame(int32 frame);
     virtual int32 NumIndices() const;
-    virtual bool FrameLogLikelihood(int32 frame,
-                                    std::vector<kaldi::BaseFloat>* likelihood);
+    // not logprob
+    virtual bool FrameLikelihood(int32 frame,
+                                 std::vector<kaldi::BaseFloat>* likelihood);
     virtual int32 NumFramesReady() const;
     // for offline test
     void Acceptlikelihood(const kaldi::Matrix<kaldi::BaseFloat>& likelihood);
     void Reset();
     bool IsInputFinished() const { return frontend_->IsFinished(); }
     bool EnsureFrameHaveComputed(int32 frame);
+    int32 TokenId2NnetId(int32 token_id);
 
   private:
     bool AdvanceChunk();
     std::shared_ptr<FrontendInterface> frontend_;
     std::shared_ptr<NnetInterface> nnet_;
     kaldi::Matrix<kaldi::BaseFloat> nnet_cache_;
+    // the frame is nnet prob frame rather than audio feature frame
+    // nnet frame subsample the feature frame
+    // eg: 35 frame features output 8 frame inferences
     int32 frame_offset_;
     int32 frames_ready_;
     // todo: feature frame mismatch with nnet inference frame
-    // eg: 35 frame features output 8 frame inferences
     // so use subsampled_frame
     int32 current_log_post_subsampled_offset_;
     int32 num_chunk_computed_;