search

ZFTurbo / Music-Source-Separation-Training

Repository for training models for music source separation.
MIT License
485 stars 67 forks source link

Using other models? #94

Open AirJCovers34 opened 6 days ago

AirJCovers34 commented 6 days ago

Is it possible to run the inference script using other models? I'm specifically thinking of VR models, such as 5_HP-Karaoke-UVR.pth or UVR-BVE-4B_SN-44100-1.pth.

Similarly, could we consider integrating models like Medley Vox?

ZFTurbo commented 6 days ago

Currently it's not possible

AirJCovers34 commented 5 days ago

Not even by using a rough or less conventional method? 😥

ZFTurbo commented 5 days ago

I think it's possible to add some older models in current code. Pull requests always welcome )