Rodrigoandrigo
Rodrigoandrigo
https://devblogs.microsoft.com/directx/in-the-works-opencl-and-opengl-mapping-layers-to-directx/ this may help
Changing torch.float16 to torch.float32 might solve the problem I think
https://github.com/ROCm/ROCm/issues/2536#issuecomment-2120651850
Yes, but it doesn't work with local models with lm-studio and I didn't find this interface in the documentation.