Take note of the first argument we pass toStableDiffusionOnnxPipeline.from_pretrained()."./stable_diffusion_onnx". That's a file path to the Onnx-ified model we just created. Andproviderneeds to be"DmlExecutionProvider"in order to actually instruct Stable Diffusion to use DirectML, instead ...
bhackmentioned this issueMay 11, 2024 Cuda on AMDpytorch/pytorch#126004 Closed Sign up for freeto join this conversation on GitHub. Already have an account?Sign in to comment Assignees No one assigned Labels None yet Projects None yet