Open eliranwong opened 6 months ago
I added partial support in my fork, but it's via DirectML and pretty buggy, and last I checked the DirectML Pytorch component hasn't been updated for a year. Keep an eye on this fork: https://github.com/PABannier/bark.cpp they may eventually get AMD support.
Linux users use ROCM
Can Bark support acceleration with AMD GPU with ROCM?