-
统计时长
wav-to-duration scp:data/train/wav.scp ark,t:- 2>/dev/null|awk 'BEGIN{SUM=0}{SUM+=$2}END{print SUM/3600}'
或
awk 'BEGIN{SUM=0}{SUM+=$2}END{print SUM/3600}' data/train/utt2dur
-
生成utt2dur
utils/data/get_utt2dur.sh data/train
-
统计句子数
wc -l data/train/text