search

FunAudioLLM / CosyVoice

Multi-lingual large voice generation model, providing inference, training and deployment full-stack ability.
https://funaudiollm.github.io/
Apache License 2.0
5.01k stars 509 forks source link

vioce merge #429

Open CuiRobert opened 2 hours ago

CuiRobert commented 2 hours ago

There is a demonstration of the fusion of male and female timbres in the example. How do you achieve natural timbre fusion?

aluminumbox commented 1 hour ago

well we select these embedding fusion manually

CuiRobert commented 1 hour ago

well we select these embedding fusion manually

thx! Do you use zero-shot or sft for fusion?