Working on this issue for another project~ 50XX series have a newer architecture (SM12.0) that isn't supported in current PyTorch CUDA builds. You can maybe wait 1-3 months for a new stable build to be released or you can try your luck with getting the nightly build to work:
pip install --pre --upgrade --no-user torch torchaudio --index-url https://download.pytorch.org/whl/nightly/cu128 --target "C:\your_torch_torchaudio__lib_dir"

Good luck~! :3

Working on this issue for another project~ 50XX series have a newer architecture (SM12.0) that isn't supported in current PyTorch CUDA builds. You can maybe wait 1-3 months for a new stable build to be released or you can try your luck with getting the nightly build to work: pip install --pre --upgrade --no-user torch torchaudio --index-url https://download.pytorch.org/whl/nightly/cu128 --target "C:\your_torch_torchaudio__lib_dir"

Any luck with this now that the stable builds are released? I've tried upgrading both torch and xformers, but now it needs torch-directml which hasn't been updated for a while. I've only had success by forcing RVC to run on CPU.