Skip to content

The most powerful and modular diffusion model GUI, api and backend with a graph/nodes interface. The fastest local inference engine in the world.

Enable triton comfy kitchen via cli-arg (#12730)

Silverb138133ffa43
diff --git a/comfy/cli_args.py b/comfy/cli_args.py
index cef1a5e6b..d2fde8b67 100644
--- a/comfy/cli_args.py
+++ b/comfy/cli_args.py
@@ -91,6 +91,7 @@ parser.add_argument("--directml", type=int, nargs="?", metavar="DIRECTML_DEVICE"
parser.add_argument("--oneapi-device-selector", type=str, default=None, metavar="SELECTOR_STRING", help="Sets the oneAPI device(s) this instance will use.")
parser.add_argument("--supports-fp8-compute", action="store_true", help="ComfyUI will act like if the device supports fp8 compute.")
+parser.add_argument("--enable-triton-backend", action="store_true", help="ComfyUI will enable the use of Triton backend in comfy-kitchen. Is disabled at launch by default.")
class LatentPreviewMethod(enum.Enum):
NoPreviews = "none"
diff --git a/comfy/quant_ops.py b/comfy/quant_ops.py
index 42ee08fb2..b90bcfd25 100644
--- a/comfy/quant_ops.py
+++ b/comfy/quant_ops.py
@@ -1,6 +1,8 @@
import torch
import logging
+from comfy.cli_args import args
+
try:
import comfy_kitchen as ck
from comfy_kitchen.tensor import (
@@ -21,7 +23,15 @@ try:
ck.registry.disable("cuda")
logging.warning("WARNING: You need pytorch with cu130 or higher to use optimized CUDA operations.")
- ck.registry.disable("triton")
+ if args.enable_triton_backend:
+ try:
+ import triton
+ logging.info("Found triton %s. Enabling comfy-kitchen triton backend.", triton.__version__)
+ except ImportError as e:
+ logging.error(f"Failed to import triton, Error: {e}, the comfy-kitchen triton backend will not be available.")
+ ck.registry.disable("triton")
+ else:
+ ck.registry.disable("triton")
for k, v in ck.list_backends().items():
logging.info(f"Found comfy_kitchen backend {k}: {v}")
except ImportError as e: