glog to stderr

demos/audio_searching/src/encode.py

@@ -11,11 +11,7 @@
 # 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.
-import os
-
-import librosa
 import numpy as np
-from config import DEFAULT_TABLE
 from logs import LOGGER
 
 from paddlespeech.cli import VectorExecutor

speechx/examples/CMakeLists.txt

@@ -3,3 +3,5 @@ cmake_minimum_required(VERSION 3.14 FATAL_ERROR)
 add_subdirectory(feat)
 add_subdirectory(nnet)
 add_subdirectory(decoder)
+
+add_subdirectory(glog)

speechx/examples/README.md

@@ -1,8 +1,9 @@
 # Examples
 
-* decoder - online decoder to work as offline
+* glog - glog usage
 * feat - mfcc, linear 
 * nnet - ds2 nn
+* decoder - online decoder to work as offline
 
 ## How to run
 

speechx/examples/decoder/offline_decoder_main.cc

@@ -63,6 +63,7 @@ int main(int argc, char* argv[]) {
 
     int32 chunk_size = 35;
     decoder.InitDecoder();
+    LOG(INFO) << "chunk size: " << chunk_size;
 
     for (; !feature_reader.Done(); feature_reader.Next()) {
         string utt = feature_reader.Key();

speechx/examples/decoder/run.sh

@@ -25,6 +25,9 @@ model_dir=../paddle_asr_model
 feat_wspecifier=./feats.ark
 cmvn=./cmvn.ark
 
+
+export GLOG_logtostderr=1
+
 # 3. run feat
 linear_spectrogram_main \
     --wav_rspecifier=scp:$model_dir/wav.scp \
@@ -37,4 +40,4 @@ offline_decoder_main \
     --model_path=$model_dir/avg_1.jit.pdmodel \
     --param_path=$model_dir/avg_1.jit.pdparams \
     --dict_file=$model_dir/vocab.txt \
-    --lm_path=$model_dir/avg_1.jit.klm
+    --lm_path=$model_dir/avg_1.jit.klm

speechx/examples/feat/linear_spectrogram_main.cc

@@ -25,6 +25,8 @@
 #include "kaldi/util/kaldi-io.h"
 #include "kaldi/util/table-types.h"
 
+#include <glog/logging.h>
+
 DEFINE_string(wav_rspecifier, "", "test wav scp path");
 DEFINE_string(feature_wspecifier, "", "output feats wspecifier");
 DEFINE_string(cmvn_write_path, "./cmvn.ark", "write cmvn");
@@ -172,6 +174,9 @@ int main(int argc, char* argv[]) {
     ppspeech::LinearSpectrogramOptions opt;
     opt.frame_opts.frame_length_ms = 20;
     opt.frame_opts.frame_shift_ms = 10;
+    LOG(INFO) << "frame length (ms):" << opt.frame_opts.frame_length_ms;
+    LOG(INFO) << "frame shift (ms):" << opt.frame_opts.frame_shift_ms;
+
     ppspeech::DecibelNormalizerOptions db_norm_opt;
     std::unique_ptr<ppspeech::FeatureExtractorInterface> base_feature_extractor(
         new ppspeech::DecibelNormalizer(db_norm_opt, std::move(data_source)));
@@ -190,6 +195,11 @@ int main(int argc, char* argv[]) {
     int sample_rate = 16000;
     int chunk_sample_size = streaming_chunk * sample_rate;
 
+    LOG(INFO) << "sr:" << sample_rate;
+    LOG(INFO) << "chunk size (s):" << streaming_chunk;
+    LOG(INFO) << "chunk size (sample):" << chunk_sample_size;
+
+
     for (; !wav_reader.Done(); wav_reader.Next()) {
         std::string utt = wav_reader.Key();
         const kaldi::WaveData& wave_data = wav_reader.Value();

speechx/examples/feat/run.sh

@@ -25,6 +25,7 @@ feat_wspecifier=./feats.ark
 cmvn=./cmvn.ark
 
 # 3. run feat
+export GLOG_logtostderr=1
 linear_spectrogram_main \
     --wav_rspecifier=scp:$model_dir/wav.scp \
     --feature_wspecifier=ark,t:$feat_wspecifier \