update repo

egs/aishell/conformer/run.sh

@@ -85,7 +85,7 @@ fi
 
 if [ ${stage} -le 1 ] && [ ${stop_stage} -ge 1 ]; then
     echo "stage 1: Feature and CMVN Generation"
-    utils/compute_cmvn.sh --cmd "$train_cmd" --nj $nj --feats_dim ${feats_dim} ${feats_dir}/data/${train_set}
+    utils/compute_cmvn.sh ${feats_dir}/data/${train_set} --cmd "$train_cmd" --nj $nj --feats_dim ${feats_dim} --config $asr_config --scale 1.0
 fi
 
 token_list=${feats_dir}/data/${lang}_token_list/char/tokens.txt

egs/aishell/data2vec_paraformer_finetune/run.sh

@@ -88,7 +88,7 @@ fi
 
 if [ ${stage} -le 1 ] && [ ${stop_stage} -ge 1 ]; then
     echo "stage 1: Feature and CMVN Generation"
-    utils/compute_cmvn.sh --cmd "$train_cmd" --nj $nj --feats_dim ${feats_dim} ${feats_dir}/data/${train_set}
+    utils/compute_cmvn.sh ${feats_dir}/data/${train_set} --cmd "$train_cmd" --nj $nj --feats_dim ${feats_dim} --config $asr_config --scale 1.0
 fi
 
 token_list=${feats_dir}/data/${lang}_token_list/char/tokens.txt

egs/aishell/data2vec_transformer_finetune/run.sh

@@ -88,7 +88,7 @@ fi
 
 if [ ${stage} -le 1 ] && [ ${stop_stage} -ge 1 ]; then
     echo "stage 1: Feature and CMVN Generation"
-    utils/compute_cmvn.sh --cmd "$train_cmd" --nj $nj --feats_dim ${feats_dim} ${feats_dir}/data/${train_set}
+    utils/compute_cmvn.sh ${feats_dir}/data/${train_set} --cmd "$train_cmd" --nj $nj --feats_dim ${feats_dim} --config $asr_config --scale 1.0
 fi
 
 token_list=${feats_dir}/data/${lang}_token_list/char/tokens.txt

egs/aishell/paraformer/run.sh

@@ -85,7 +85,7 @@ fi
 
 if [ ${stage} -le 1 ] && [ ${stop_stage} -ge 1 ]; then
     echo "stage 1: Feature and CMVN Generation"
-    utils/compute_cmvn.sh --cmd "$train_cmd" --nj $nj --feats_dim ${feats_dim} ${feats_dir}/data/${train_set}
+    utils/compute_cmvn.sh ${feats_dir}/data/${train_set} --cmd "$train_cmd" --nj $nj --feats_dim ${feats_dim} --config $asr_config --scale 1.0
 fi
 
 token_list=${feats_dir}/data/${lang}_token_list/char/tokens.txt

egs/aishell/paraformerbert/run.sh

@@ -89,7 +89,7 @@ fi
 
 if [ ${stage} -le 1 ] && [ ${stop_stage} -ge 1 ]; then
     echo "stage 1: Feature and CMVN Generation"
-    utils/compute_cmvn.sh --cmd "$train_cmd" --nj $nj --feats_dim ${feats_dim} ${feats_dir}/data/${train_set}
+    utils/compute_cmvn.sh ${feats_dir}/data/${train_set} --cmd "$train_cmd" --nj $nj --feats_dim ${feats_dim} --config $asr_config --scale 1.0
 fi
 
 token_list=${feats_dir}/data/${lang}_token_list/char/tokens.txt

egs/aishell/transformer/run.sh

@@ -85,7 +85,7 @@ fi
 
 if [ ${stage} -le 1 ] && [ ${stop_stage} -ge 1 ]; then
     echo "stage 1: Feature and CMVN Generation"
-    utils/compute_cmvn.sh --cmd "$train_cmd" --nj $nj --feats_dim ${feats_dim} ${feats_dir}/data/${train_set}
+    utils/compute_cmvn.sh ${feats_dir}/data/${train_set} --cmd "$train_cmd" --nj $nj --feats_dim ${feats_dim} --config $asr_config --scale 1.0
 fi
 
 token_list=${feats_dir}/data/${lang}_token_list/char/tokens.txt

egs/aishell/transformer/utils/compute_cmvn.py

@@ -5,6 +5,7 @@ import os
 import numpy as np
 import torchaudio
 import torchaudio.compliance.kaldi as kaldi
+import yaml
 
 
 def get_parser():
@@ -25,6 +26,11 @@ def get_parser():
         type=str,
         help="the path of wav scps",
     )
+    parser.add_argument(
+        "--config",
+        type=str,
+        help="the config file for computing cmvn",
+    )
     parser.add_argument(
         "--idx",
         default=1,
@@ -82,11 +88,27 @@ def main():
     #         mean_stats += np.sum(mat, axis=0)
     #         var_stats += np.sum(np.square(mat), axis=0)
     #         total_frames += mat.shape[0]
+
+    with open(args.config) as f:
+        configs = yaml.safe_load(f)
+        frontend_configs = configs.get("frontend_conf", {})
+        num_mel_bins = frontend_configs.get("n_mels", 80)
+        frame_length = frontend_configs.get("frame_length", 25)
+        frame_shift = frontend_configs.get("frame_shift", 10)
+        window_type = frontend_configs.get("window", "hamming")
+        resample_rate = frontend_configs.get("fs", 16000)
+        assert num_mel_bins == args.dim
+
     with open(wav_scp_file) as f:
         lines = f.readlines()
         for line in lines:
             _, wav_file = line.strip().split()
-            fbank = compute_fbank(wav_file, num_mel_bins=args.dim)
+            fbank = compute_fbank(wav_file,
+                                  num_mel_bins=args.dim,
+                                  frame_length=frame_length,
+                                  frame_shift=frame_shift,
+                                  resample_rate=resample_rate,
+                                  window_type=window_type)
             mean_stats += np.sum(fbank, axis=0)
             var_stats += np.sum(np.square(fbank), axis=0)
             total_frames += fbank.shape[0]

egs/aishell/transformer/utils/compute_cmvn.sh

@@ -2,15 +2,19 @@
 
 . ./path.sh || exit 1;
 # Begin configuration section.
+fbankdir=$1
 nj=32
 cmd=./utils/run.pl
 feats_dim=80
+config=
+scale=1.0
 
 echo "$0 $@"
 
 . utils/parse_options.sh || exit 1;
 
-fbankdir=$1
+# shellcheck disable=SC2046
+head -n $(awk -v lines="$(wc -l < ${fbankdir}/wav.scp)" -v scale="$scale" 'BEGIN { printf "%.0f\n", lines*scale }') ${fbankdir}/wav.scp > ${fbankdir}/wav.scp.scale
 
 split_dir=${fbankdir}/cmvn/split_${nj};
 mkdir -p $split_dir
@@ -18,14 +22,15 @@ split_scps=""
 for n in $(seq $nj); do
     split_scps="$split_scps $split_dir/wav.$n.scp"
 done
-utils/split_scp.pl ${fbankdir}/wav.scp $split_scps || exit 1;
+utils/split_scp.pl ${fbankdir}/wav.scp.scale $split_scps || exit 1;
 
 logdir=${fbankdir}/cmvn/log
 $cmd JOB=1:$nj $logdir/cmvn.JOB.log \
     python utils/compute_cmvn.py \
       --dim ${feats_dim} \
       --wav_path $split_dir \
-      --idx JOB
+      --config $config \
+      --idx JOB \
 
 python utils/combine_cmvn_file.py --dim ${feats_dim} --cmvn_dir $split_dir --nj $nj --output_dir ${fbankdir}/cmvn