train_metric smaller than val_metric

Question

train_metric smaller than val_metric

Opened this issue 4 years ago · 4 comments

Hallo,

I have run this code on same dataset (DEMAND+VoiceBank). But during the training the train_metric(pesq) is much smaller than val_metric, namely, train_metirc is about 2.7 and val_metric is about 3.2. Do you have the similar situation?

Answer 1 · 2020-11-27T08:12:35.000Z

I run this code with the latest version of pypesq as metric, it returns nan for every step. Do you suffer this problem? How did you fix?

Answer 2 · 2020-11-27T22:20:56.000Z

@sailorbj Yeah I also have that problem. Using another pesq function can solve it.

Answer 3 · 2020-11-28T01:34:14.000Z

Thanks for you reply. Which pesq function works well?

Answer 4 · 2020-11-28T08:13:48.000Z

@sailorbj It was given by my tutor, sorry for can't share it. I am not sure if he is willing to.