search

sungwon23 / BSRNN

79 stars 13 forks source link

reproductioin #5

Closed Janeliao-C closed 10 months ago

Janeliao-C commented 1 year ago

Hello，thank you for your sharing and generous！ But I got a problem when I reproduct your result with totally the same code. I use VoiceBank+Demand dataset , sampling rate 16kHz. And I split the 28 speakers in the trainset into train and test for 26 and 2 spkears ,respectively. I set batch size as 18, and train epoch 100 in total. When I test the model in VoiceBank+Demand testset, BUT I got PESQ at 1.97, STOI 0.92, it's weird and just like the model doesn't has been trained. Are you use the same testset in train and test? Or can you share your opinion about this question? Thanks a lot.

Janeliao-C commented 1 year ago

Sorry to bother you agian. I can't get a result totally the same as yours. PESQ 3.07, CSIG 4.48, CBAK 3.68, CVOL 3.85, SSNR 9.55, STOI 0.95 is my best result. All of them had an error less than 1.5%, But I still wonder how do you look at this issue, is this a normal error?