Unable to find VC webui

Open sekkit opened this issue 1 year ago • 1 comments

Currently for v2, only VC to TTS webui is available, but where is the Voice Conversion only feature?

Aug 12 '24 03:08 sekkit

Same question, I tried to modify the script in GPT-SOVITS-VC based on the new pretrained model but failed. The output only contains noise without any pronunciation.

ge = vq_model.ref_enc(spec[:,:704])

This line is the new execution in the new version. It adopts spec[:,:704] instead of spec. I copied that but failed to reconstrcuct the audio.

Aug 14 '24 02:08 WindyyamCY