tugstugi/pytorch-dc-tts

transfer learning ljspeech

fivevolt opened this issue · 2 comments

First of all thank you for the repository, the result is really awesome.
Is it possible to transfer learning from pretrained models (ljspeech) into my recording data?

If it is english it is possible, but requires some code. You have to modify train-text2mel.py and train-ssrn.py to include such code:

text2mel.load_state_dict(torch.load('pretrained_ljspeech_text2mel.pth').state_dict())

and

ssrn.load_state_dict(torch.load('pretrained_ljspeech_ssrn.pth').state_dict())

Where would this piece of code be added in the files?

@tugstugi