sesame

/

csm-1b

like 2.16k

Follow

Sesame 1.4k

try out this repo https://github.com/isaiahbjork/csm-voice-cloning

Just pointing out, fine tuning is different from cloning.

check this out: https://github.com/davidbrowne17/csm-streaming

Use this one for finetuning and training from scratch: https://github.com/knottwill/sesame-finetune

It finetunes by modifying the original weights, as opposed to the LoRA approach in the previous responses, which is optimal for domain shifts like new languages.