remove streaming opt, test=doc

speechx/examples/custom_asr/run.sh

@@ -71,7 +71,6 @@ if [ ${stage} -le 4 ] && [ ${stop_stage} -ge 4 ]; then
   recognizer_test_main \
     --wav_rspecifier=scp:$wav_scp \
     --cmvn_file=$cmvn \
-    --streaming_chunk=30 \
     --use_fbank=true \
     --model_path=$model_dir/avg_10.jit.pdmodel \
     --param_path=$model_dir/avg_10.jit.pdiparams \

speechx/examples/ds2_ol/aishell/run.sh

@@ -78,7 +78,6 @@ if [ ${stage} -le 1 ] && [ ${stop_stage} -ge 1 ]; then
         --wav_rspecifier=scp:$data/split${nj}/JOB/${aishell_wav_scp} \
         --feature_wspecifier=ark,scp:$data/split${nj}/JOB/feat.ark,$data/split${nj}/JOB/feat.scp \
         --cmvn_file=$cmvn \
-        --streaming_chunk=0.36
     echo "feature make have finished!!!"
 fi
 
@@ -155,7 +154,6 @@ if [ ${stage} -le 5 ] && [ ${stop_stage} -ge 5 ]; then
         --wav_rspecifier=scp:$data/split${nj}/JOB/${aishell_wav_scp} \
         --cmvn_file=$cmvn \
         --model_path=$model_dir/avg_1.jit.pdmodel \
-        --streaming_chunk=30 \
         --param_path=$model_dir/avg_1.jit.pdiparams \
         --word_symbol_table=$wfst/words.txt \
         --model_output_names=softmax_0.tmp_0,tmp_5,concat_0.tmp_0,concat_1.tmp_0 \

speechx/examples/ds2_ol/aishell/run_fbank.sh

@@ -152,7 +152,6 @@ if [ ${stage} -le 5 ] && [ ${stop_stage} -ge 5 ]; then
         --wav_rspecifier=scp:$data/split${nj}/JOB/${aishell_wav_scp} \
         --cmvn_file=$cmvn \
         --model_path=$model_dir/avg_5.jit.pdmodel \
-        --streaming_chunk=30 \
         --use_fbank=true \
         --param_path=$model_dir/avg_5.jit.pdiparams \
         --word_symbol_table=$wfst/words.txt \

speechx/examples/ds2_ol/websocket/websocket_client.sh

@@ -32,4 +32,4 @@ export GLOG_logtostderr=1
 
 # websocket client
 websocket_client_main \
-    --wav_rspecifier=scp:$data/$aishell_wav_scp --streaming_chunk=0.36
+    --wav_rspecifier=scp:$data/$aishell_wav_scp

speechx/examples/ds2_ol/websocket/websocket_server.sh

@@ -62,7 +62,6 @@ fi
 websocket_server_main \
     --cmvn_file=$cmvn \
     --model_path=$model_dir/avg_1.jit.pdmodel \
-    --streaming_chunk=0.1 \
     --param_path=$model_dir/avg_1.jit.pdiparams \
     --word_symbol_table=$wfst/words.txt \
     --model_output_names=softmax_0.tmp_0,tmp_5,concat_0.tmp_0,concat_1.tmp_0 \

speechx/speechx/decoder/param.h

@@ -25,7 +25,6 @@ DEFINE_bool(use_fbank, false, "False for fbank; or linear feature");
 // feature, or fbank");
 DEFINE_int32(num_bins, 161, "num bins of mel");
 DEFINE_string(cmvn_file, "", "read cmvn");
-DEFINE_double(streaming_chunk, 0.1, "streaming feature chunk size");
 // feature sliding window
 DEFINE_int32(receptive_field_length,
              7,

speechx/speechx/decoder/recognizer_main.cc

@@ -33,7 +33,7 @@ int main(int argc, char* argv[]) {
     kaldi::TokenWriter result_writer(FLAGS_result_wspecifier);
 
     int sample_rate = FLAGS_sample_rate;
-    float streaming_chunk = FLAGS_streaming_chunk;
+    float streaming_chunk = 0.1;
     int chunk_sample_size = streaming_chunk * sample_rate;
     LOG(INFO) << "sr: " << sample_rate;
     LOG(INFO) << "chunk size (s): " << streaming_chunk;
@@ -96,4 +96,4 @@ int main(int argc, char* argv[]) {
     KALDI_LOG << " cost:" << elapsed << " s";
     KALDI_LOG << "total wav duration is: " << tot_wav_duration << " s";
     KALDI_LOG << "the RTF is: " << elapsed / tot_wav_duration;
-}
+}