The most powerful and modular diffusion model GUI, api and backend with a graph/nodes interface. The fastest local inference engine in the world.
Don't enable pytorch attention on AMD if triton isn't available. (#9747)
comfyanonymousbcbd7884e3af
diff --git a/comfy/model_management.py b/comfy/model_management.pyindex 17516b6ed..cb6580f73 100644--- a/comfy/model_management.py+++ b/comfy/model_management.py@@ -22,6 +22,7 @@ from enum import Enumfrom comfy.cli_args import args, PerformanceFeatureimport torchimport sys+import importlibimport platformimport weakrefimport gc@@ -336,12 +337,13 @@ try:logging.info("AMD arch: {}".format(arch))logging.info("ROCm version: {}".format(rocm_version))if args.use_split_cross_attention == False and args.use_quad_cross_attention == False:- if torch_version_numeric >= (2, 7): # works on 2.6 but doesn't actually seem to improve much- if any((a in arch) for a in ["gfx90a", "gfx942", "gfx1100", "gfx1101", "gfx1151"]): # TODO: more arches, TODO: gfx950- ENABLE_PYTORCH_ATTENTION = True-# if torch_version_numeric >= (2, 8):-# if any((a in arch) for a in ["gfx1201"]):-# ENABLE_PYTORCH_ATTENTION = True+ if importlib.util.find_spec('triton') is not None: # AMD efficient attention implementation depends on triton. TODO: better way of detecting if it's compiled in or not.+ if torch_version_numeric >= (2, 7): # works on 2.6 but doesn't actually seem to improve much+ if any((a in arch) for a in ["gfx90a", "gfx942", "gfx1100", "gfx1101", "gfx1151"]): # TODO: more arches, TODO: gfx950+ ENABLE_PYTORCH_ATTENTION = True+# if torch_version_numeric >= (2, 8):+# if any((a in arch) for a in ["gfx1201"]):+# ENABLE_PYTORCH_ATTENTION = Trueif torch_version_numeric >= (2, 7) and rocm_version >= (6, 4):if any((a in arch) for a in ["gfx1201", "gfx942", "gfx950"]): # TODO: more archesSUPPORT_FP8_OPS = True