Repository navigation
Tensorboard ? :D #3
Description
Do you have any tensorboard :D, an audio samples sound good except background noise, maybe training longer will solve this problem haha :D, great job :D.
{{ message }}
Do you have any tensorboard :D, an audio samples sound good except background noise, maybe training longer will solve this problem haha :D, great job :D.
Activity
[-]Tensorboard :D[/-][+]Tensorboard ? :D[/+]rishikksh20 commentedon Jul 5, 2020
rishikksh20 commentedon Jul 5, 2020
rishikksh20 commentedon Jul 5, 2020
@dathudeptrai Currently I am using raw pitch and energy with MSE that's why error looks so high but if required I will standardize or normalize in future. Pitch and energy both seem to be static after 10k steps and going towards over-fitting.
dathudeptrai commentedon Jul 6, 2020
@rishikksh20 is the l1 loss use masked ? . I use the same preprocessing as ESPNET but my valid l1 loss is around < 0.3.
rishikksh20 commentedon Jul 6, 2020
@dathudeptrai
I am using l1 loss masked actually l1_loss is combined loss of before and after Postnet l1 loss that's why it's very, although before and after l1 losses are also very high around 0.47 it should be around < 0.3 I don't know why is this happening whereas the quality of generated audio and spectrogram seems fine. And I am not using ESPNet pre-processing actually I am using Nvidia's pre-processing for simplicity and easy integration with
Waveglowand my own implementation ofMelgan. You can look at the code especiallyfastspeech.py, if you find any irregularities in implementation, please let me know.dathudeptrai commentedon Jul 6, 2020
@rishikksh20 is nvidia norm from 0->4 ?
rishikksh20 commentedon Jul 6, 2020
@dathudeptrai I don't think so
FastSpeech2/utils/stft.py
Line 161 in 5bc2b40
they are using spectral normalization:
FastSpeech2/utils/stft.py
Line 153 in 5bc2b40