iamliulong

when i change the 'args' of the ckeckpoint the author offered from the model i trained, still checkpoint parameter shape is different from the model. when run lyric2beats = TransformerModel.from_pretrained(f'checkpoints/{lyric2beat_prefix}',...

in https://huggingface.co/facebook/hubert-base-ls960, there is no .pt file, 

> > in https://huggingface.co/facebook/hubert-base-ls960, there is no .pt file,  > > You can install it in: https://github.com/pytorch/fairseq/tree/master/examples/hubert > > that can be found in the model card page. got...

> https://github.com/mozilla/TTS/tree/master/TTS/speaker_encoder thanks a lot!

> Hi! Awesome that you are training models by yourself. > > I'm also curious whether or not the training works on short audio. > > Please share the results...