An error occurs during processing
RuntimeError: "Error(s) in loading state_dict for BSRoformer:
Missing key(s) in state_dict: "mask_estimators.0.to_freqs.0.0.0.weight", "mask_estimators.0.to_freqs.0.0.0.bias", "mask_estimators.0.to_freqs.0.0.2.weight", "mask_estimators.0.to_freqs.0.0.2.bias", "mask_estimators.0.to_freqs.1.0.0.weight", "mask_estimators.0.to_freqs.1.0.0.bias", "mask_estimators.0.to_freqs.1.0.2.weight", "mask_estimators.0.to_freqs.1.0.2.bias",.............
Such lines appear in the log immediately after launching the Inference. Please fix this issue if possible.
Did you edit the code as described in the model card?
I edited it bs_roformer.py but I still haven't figured out how to import it into UVR.
upd: I've read that Music Source Separation Training should be used at the moment, but will future versions allow it to be used in UVR, or will the model require code intervention anyway?
UVR is not currently compatible.
Okay, thank you. I hope to see UVR support in the future
The UVR side needs to support this architectural change.
There is nothing I can do about it.