Getting started
jbbqqf opened this issue · 8 comments
I'm really interested in generating music from text. However, I have poor knowledge of the pytorch framework and ML in general.
I appreciate the README a lot. However it's quite implicit, especially for a newbe like me. Is there a way I can get started ?
How to get the ./hubert/hubert_base_ls960.pt and ./hubert/hubert_base_ls960_L9_km500.bin files ? Where to download audio files ?
I'm also looking to get this running but haven't had the time to mess around with it yet. As far as getting the audio files I did find this, which I think it's what google used for generating the examples shown in the project page, hope that helps
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.
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 that, thanks a lot!
hi. May I ask if you have successfully created a demo and could you share your ideas? I really feel like if I could share your code. thx
I'm really interested in generating music from text. However, I have poor knowledge of the pytorch framework and ML in general.
what is "/path/to/audio/files", Thank a lots!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.
I downloaded as below
and then, my problem was "KeyError: 'hubert_pretraining'"
when I downloaded HuBERT Large (and edited code to checkpoint_path = './hubert/hubert_large_ll60k.pt',), my problem was "AttributeError: 'NoneType' object has no attribute 'task'"