SeongYeonPark

Results 3 comments of SeongYeonPark

For this stretch factor, the stretch effect has better performance.

To add on, the above code sometimes does and sometimes doesn't give the warning, depending on the sampled stretch ratio. However, I am not using that code itself, but rather...

I have a question about your WER/CER results in the paper.

Thank you for answering!

Question about the scale of gaussian noise in the mel spectrogram

Also, could you tell me if you had issues with Nan losses (when applying the SR augmentation)?