search

niliqiang / kaldi_lre

My language recognition system based on Kaldi, using CommonVoice dataset.
1 stars 1 forks source link

运行过程中报错：utils/split_scp.pl: Refusing to split data... #2

Open niliqiang opened 4 years ago

niliqiang commented 4 years ago

完整报错log: utils/split_scp.pl: Refusing to split data because number of speakers 5 is less than the number of output .scp files 20

niliqiang commented 4 years ago

为了便于每个jobs处理数据，kaldi会将数据按照jobs的数量或者jobs*num_processes进行切分，要注意的是，语种数（说话人数）要大于数据切分数，否则会报上述错误。