Fix for running via DirectML (#6542)

* Fix for running via DirectML

Fix DirectML empty image generation issue with Flux1. add CPU fallback for unsupported path. Verified the model works on AMD GPUs

* fix formating

* update casual mask calculation
This commit is contained in:
HishamC
2025-02-11 14:11:32 -08:00
committed by GitHub
parent af4b7c91be
commit b124256817
3 changed files with 13 additions and 2 deletions

View File

@@ -22,7 +22,7 @@ def attention(q: Tensor, k: Tensor, v: Tensor, pe: Tensor, mask=None) -> Tensor:
def rope(pos: Tensor, dim: int, theta: int) -> Tensor:
assert dim % 2 == 0
if comfy.model_management.is_device_mps(pos.device) or comfy.model_management.is_intel_xpu():
if comfy.model_management.is_device_mps(pos.device) or comfy.model_management.is_intel_xpu() or comfy.model_management.is_directml_enabled():
device = torch.device("cpu")
else:
device = pos.device