You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Hi, I would like to use the pretrained model on LibriTTS to adapt the model on two target speakers for which I have about 40 minutes of training data each.
Could you please share how would be the approach for fine tuning it?
Any modules to freeze, decreasing learning rate, if it is actually possible in your opinion with that amount of data etc..
Any info would be useful.
Thanks for your work and have a good day.
The text was updated successfully, but these errors were encountered:
Hi, I would like to use the pretrained model on LibriTTS to adapt the model on two target speakers for which I have about 40 minutes of training data each.
Could you please share how would be the approach for fine tuning it?
Any modules to freeze, decreasing learning rate, if it is actually possible in your opinion with that amount of data etc..
Any info would be useful.
Thanks for your work and have a good day.
The text was updated successfully, but these errors were encountered: