VoiceCraft
VoiceCraft copied to clipboard
jasonppy
Where are the 330/830TTSEnhanced .pth models?
I just wanted to download them.
Op zo 21 apr. 2024 20:57 schreef Puyuan Peng @.***>:
what is the error that you encounterred?
— Reply to this email directly, view it on GitHub https://github.com/jasonppy/VoiceCraft/issues/107#issuecomment-2068166063, or unsubscribe https://github.com/notifications/unsubscribe-auth/ASB2UQKK35MNFC26SE3HQW3Y6QD3LAVCNFSM6AAAAABGRSYKY6VHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZDANRYGE3DMMBWGM . You are receiving this because you authored the thread.Message ID: @.***>
They are not on Huggingface (https://huggingface.co/pyp1/VoiceCraft/tree/main) or their own repos there except as safetensor.
Also, what is the difference between the 330MTTSEnhanced and the HalfLibri330MTTSEnhanced?
PS. Perhaps you just haven't released them yet? Though the space seems to be using them?
Yes, I can access it, but I see only safetensor files for 330M/830MTTSEnhanced models.
ah sorry about that, a while ago I updated the files to be safetensors, they should serve the same purpose right?
Sure, but isn't the code in the repo still expecting .pth files, eg. in gradio.py:
voicecraft_name = f"{voicecraft_model_name}.pth"
model = voicecraft.VoiceCraft.from_pretrained(f"pyp1/VoiceCraft_{voicecraft_name.replace('.pth', '')}")
phn2num = model.args.phn2num
config = model.args
model.to(device)
well, .pth got replaced with ''
model = voicecraft.VoiceCraft.from_pretrained(f"pyp1/VoiceCraft_{voicecraft_name.replace('.pth', '')}")
You can find the .pth files here https://huggingface.co/pyp1/VoiceCraft/tree/main