Got to run a model in ComfyUI
Hi @GaboxR67 I only want to share that i got to run voc_fv6 in ComfyUI. I just convert it from .ckpt to .safetensors
And some question that I have:
1- Do you made a fine tuning or train from scratch? (Can you share info, i'm interested)
2- Can you explain to me why there are "experimental" (i think they are test before"release"), "instrumental", "karaoke", "vocals" and what's the differents between each one.
every model i release are finetunes
experimental models are some weights that i plan to finetune more on the future
Which datasets did you use to train the karaoke model? How was the data annotation? Was it just the lead singer + harmony or the lead singer + harmony + accompaniment?
Inst+Backing Vocals = Instrumental.flac
Lead vocal = Vocals.flac