problem training about multilingual_text_to_speech

[Tomiinek]

Hello 🙂

So do I understand it correctly that you are training with the CSS10 on a Russian ₊ French subset? No other data?

1. Hm, no it isn't. There is a pre-proccessing at the begining of training (calculating per-channel spectrogram characteristics because of normalization), but it usually takes just a few minutes.
2. If I remember correctly, I run my models on the whole CSS10 + CV until convergence and it took something like 5-6 days on two GTX1080. However, it produced intelligible speech pretty soon.

Could you share the config file?

[stale]

This issue has been automatically marked as stale because it has not had recent activity. It will be closed if no further activity occurs. Thank you for your contributions.