so-vits-svc-fork
so-vits-svc-fork copied to clipboard
Optimizations for the project
trafficstars
These are some ideas I have for the project
- Bat file that automatically opens svc GUI
- If file does cuda out of memory it loops the process but cuts file into multiple pieces so it doesn't fail this time
- Bulk infer for multiple audio files back to back
- Easy way to change default training directories
- Easy way to change default parameters like autoplay and +1 octave
I don't think these are worth implementing at least except 2 and 3, 5 is already implemented
Im a noob can you tell me how to do 5? 2 & 3 would be nice only if they can be done easily due to longer audio, a lot of my audio runs out of memory because its too many words, even if its only 2 minutes long.
Each Issue has been opened (and some closed), so please refer to it as appropriate.