update repo

egs/aishell/conformer/conf/train_asr_conformer.yaml

@@ -83,6 +83,8 @@ specaug_conf:
     num_time_mask: 2
 
 dataset_conf:
+    data_names: speech,text
+    data_types: sound,text
     shuffle: True
     shuffle_conf:
         shuffle_size: 2048
@@ -93,4 +95,4 @@ dataset_conf:
     num_workers: 8
 
 log_interval: 50
-normalize: None
+normalize: None

egs/aishell2/conformer/run.sh

@@ -103,8 +103,6 @@ if [ ${stage} -le 2 ] && [ ${stop_stage} -ge 2 ]; then
     utils/text2token.py -s 1 -n 1 --space "" ${feats_dir}/data/${train_set}/text | cut -f 2- -d" " | tr " " "\n" \
         | sort | uniq | grep -a -v -e '^\s*$' | awk '{print $0}' >> ${token_list}
     echo "<unk>" >> ${token_list}
-    mkdir -p ${feats_dir}/asr_stats_fbank_zh_char/${train_set}
-    mkdir -p ${feats_dir}/asr_stats_fbank_zh_char/${valid_set}
  fi
 
 # LM Training Stage

egs/aishell2/data2vec_pretrain/run.sh

@@ -82,8 +82,6 @@ if [ ${stage} -le 2 ] && [ ${stop_stage} -ge 2 ]; then
     utils/text2token.py -s 1 -n 1 --space "" ${feats_dir}/data/${train_set}/text | cut -f 2- -d" " | tr " " "\n" \
         | sort | uniq | grep -a -v -e '^\s*$' | awk '{print $0}' >> ${token_list}
     echo "<unk>" >> ${token_list}
-    mkdir -p ${feats_dir}/asr_stats_fbank_zh_char/${train_set}
-    mkdir -p ${feats_dir}/asr_stats_fbank_zh_char/${valid_set}
  fi
 
 # Training Stage

egs/alimeeting/sa-asr/conf/train_sa_asr_conformer.yaml

@@ -43,7 +43,6 @@ spk_encoder_conf:
   pooling_type: statistic
   num_nodes_resnet1: 256
   num_nodes_last_layer: 256
-  batchnorm_momentum: 0.5
 
 # decoder related
 decoder: sa_decoder