How to run it?
#1
by
supercharge19
- opened
Original model can be used with a wav file. How long should the wav file be? should it contain only single person's voice? is it for voice cloning?
How is voice file uploaded to llama cpp (cli or server)? can't just give it the file location can I? or do I need to encode it with xcodec2 first? if so how? and how will it decode it (it is a gguf model and when I tried full model it also installed facebook's wordvector (for audio, whatever its name is) and encoder and decoder for voice (xcodec2) in huggingface models directory. But a single gguf model can work? How, please guide.