Comfy-Kitchen errors?

[Cubey42]

Custom Node Testing

• I have tried disabling custom nodes and the issue persists (see how to disable custom nodes if you need help)

Your question

Found comfy_kitchen backend triton: {'available': True, 'disabled': True, 'unavailable_reason': None, 'capabilities': ['apply_rope', 'apply_rope1', 'dequantize_nvfp4', 'dequantize_per_tensor_fp8', 'quantize_nvfp4', 'quantize_per_tensor_fp8']}
Found comfy_kitchen backend eager: {'available': True, 'disabled': False, 'unavailable_reason': None, 'capabilities': ['apply_rope', 'apply_rope1', 'dequantize_nvfp4', 'dequantize_per_tensor_fp8', 'quantize_nvfp4', 'quantize_per_tensor_fp8', 'scaled_mm_nvfp4']}
Found comfy_kitchen backend cuda: {'available': False, 'disabled': False, 'unavailable_reason': "cannot import name '_C' from partially initialized module 'comfy_kitchen.backends.cuda' (most likely due to a circular import) (I:\\ComfyUI\\venv\\Lib\\site-packages\\comfy_kitchen\\backends\\cuda\\__init__.py)", 'capabilities': []}

latest pull gave me these errors, made it so I couldn't use wan models. rolling back temporarily resolved the issue, but I'm not sure how to fix it.

Logs

Other

No response

[huanghetv]

Yes，I'm also

[jovan2009]

Did you install comfy-kitchen? https://github.com/Comfy-Org/comfy-kitchen.git
It's a new package requirement, it's not clear from your post that you have it installed or not. I did install it and I still see some similar errors but I am able to run my Wan workflows.

[contentis]

This is not necerssrialy an error. You have multiple compute backends, where "CUDA" is limited to NVIDIA (sm80+) GPUs. In this case it'll fallback to triton, and if that is not available to the default torch eager behaviour.
If you have an RTX GPU and still see the issue, could you repot the

• python version
• torch + cuda version
• driver version
• operating system
• GPU

[Xyc2016]

Observed the same issue in downstream CI (comfy-cli, Python 3.9).

CI example:
https://github.com/Comfy-Org/comfy-cli/actions/runs/20749717414/job/59575644118?pr=346

[contentis]

Observed the same issue in downstream CI (comfy-cli, Python 3.9).

CI example: https://github.com/Comfy-Org/comfy-cli/actions/runs/20749717414/job/59575644118?pr=346

If I'm not mistaken, 3.9 has been depricated as supported version in TOT ComfyUI. Also the CI run show a different failure.

[jovan2009]

This is not necerssrialy an error. You have multiple compute backends, where "CUDA" is limited to NVIDIA (sm80+) GPUs. In this case it'll fallback to triton, and if that is not available to the default torch eager behaviour. If you have an RTX GPU and still see the issue, could you repot the

• python version
• torch + cuda version
• driver version
• operating system
• GPU

OS: Windows 11
NVIDIA GeForce RTX 5060 Ti
Driver version: 591.74
This is the beginning of my log, what I see when start the ComfyUI server, I guess rest of the data is there:

Adding extra search path checkpoints E:\ComfyUI\models\checkpoints
Adding extra search path clip E:\ComfyUI\models\clip
Adding extra search path clip_vision E:\ComfyUI\models\clip_vision
Adding extra search path configs E:\ComfyUI\models\configs
Adding extra search path controlnet E:\ComfyUI\models\controlnet
Adding extra search path diffusion_models E:\ComfyUI\models\diffusion_models
Adding extra search path diffusion_models E:\ComfyUI\models\unet
Adding extra search path embeddings E:\ComfyUI\models\embeddings
Adding extra search path loras E:\ComfyUI\models\loras
Adding extra search path upscale_models E:\ComfyUI\models\upscale_models
Adding extra search path vae E:\ComfyUI\models\vae
Checkpoint files will always be loaded safely.
Total VRAM 15977 MB, total RAM 65465 MB
pytorch version: 2.11.0.dev20260106+cu130
xformers version: 0.0.34+14f0a4fc.d20260102
Enabled fp16 accumulation.
Set vram state to: NORMAL_VRAM
Device: cuda:0 NVIDIA GeForce RTX 5060 Ti : cudaMallocAsync
Using async weight offloading with 2 streams
Enabled pinned memory 29459.0
Found comfy_kitchen backend eager: {'available': True, 'disabled': False, 'unavailable_reason': None, 'capabilities': ['apply_rope', 'apply_rope1', 'dequantize_nvfp4', 'dequantize_per_tensor_fp8', 'quantize_nvfp4', 'quantize_per_tensor_fp8', 'scaled_mm_nvfp4']}
Found comfy_kitchen backend triton: {'available': True, 'disabled': True, 'unavailable_reason': None, 'capabilities': ['apply_rope', 'apply_rope1', 'dequantize_nvfp4', 'dequantize_per_tensor_fp8', 'quantize_nvfp4', 'quantize_per_tensor_fp8']}
Found comfy_kitchen backend cuda: {'available': False, 'disabled': False, 'unavailable_reason': "cannot import name '_C' from partially initialized module 'comfy_kitchen.backends.cuda' (most likely due to a circular import) (F:\\Python314\\install\\Lib\\site-packages\\comfy_kitchen\\backends\\cuda\\__init__.py)", 'capabilities': []}
Using sage attention
Python version: 3.14.2 (main, Jan  5 2026, 17:36:44) [MSC v.1944 64 bit (AMD64)]
ComfyUI version: 0.7.0
Initializing frontend: Comfy-Org/ComfyUI_frontend@latest, requesting version details from GitHub...
[Prompt Server] web root: C:\ComfyUI\web_custom_versions\Comfy-Org_ComfyUI_frontend\1.37.3
Total VRAM 15977 MB, total RAM 65465 MB
pytorch version: 2.11.0.dev20260106+cu130
xformers version: 0.0.34+14f0a4fc.d20260102
Enabled fp16 accumulation.
Set vram state to: NORMAL_VRAM
Device: cuda:0 NVIDIA GeForce RTX 5060 Ti : cudaMallocAsync
Using async weight offloading with 2 streams
Enabled pinned memory 29459.0
WanVideoWrapper WARNING: FantasyPortrait nodes not available: No module named 'onnxruntime'
ColorMod: Ignoring node 'CV2TonemapDurand' due to cv2 edition/version

Import times for custom nodes:
   0.0 seconds: C:\ComfyUI\custom_nodes\websocket_image_save.py
   0.0 seconds: C:\ComfyUI\custom_nodes\IAMCCS-nodes
   0.0 seconds: C:\ComfyUI\custom_nodes\ComfyUI_ColorMod
   0.0 seconds: C:\ComfyUI\custom_nodes\ComfyUI-VideoHelperSuite
   0.1 seconds: C:\ComfyUI\custom_nodes\ComfyUI-KJNodes
   1.9 seconds: C:\ComfyUI\custom_nodes\ComfyUI-WanVideoWrapper

Context impl SQLiteImpl.
Will assume non-transactional DDL.
No target revision found.
Starting server

To see the GUI go to: http://127.0.0.1:8188

I can run my usual core ComfyUI WAN 2.2 workflows (it's all I tried so far), beside the comfy_kitchen console errors I didn't notice any other problems so far.

[Gabled-Waters]

I'm seeing the same issue, but on MacOS:

Found comfy_kitchen backend eager: {'available': True, 'disabled': False, 'unavailable_reason': None, 'capabilities': ['apply_rope', 'apply_rope1', 'dequantize_nvfp4', 'dequantize_per_tensor_fp8', 'quantize_nvfp4', 'quantize_per_tensor_fp8', 'scaled_mm_nvfp4']}
Found comfy_kitchen backend triton: {'available': False, 'disabled': True, 'unavailable_reason': "ImportError: No module named 'triton'", 'capabilities': []}
Found comfy_kitchen backend cuda: {'available': False, 'disabled': False, 'unavailable_reason': "cannot import name '_C' from partially initialized module 'comfy_kitchen.backends.cuda' (most likely due to a circular import) (/Users/.../ComfyUI/ltx2/venv/lib/python3.12/site-packages/comfy_kitchen/backends/cuda/init.py)", 'capabilities': []}

I made sure that I installed kitchen with the cpu settings mentioned there. I guess that's quite right, no CUDA on a Mac, but this circular import makes me wonder. back on an older version for the time being. ltx2 isn't working on MPS anyway, with the checkpoint in fp8 :-(

[Cubey42]

This is not necerssrialy an error. You have multiple compute backends, where "CUDA" is limited to NVIDIA (sm80+) GPUs. In this case it'll fallback to triton, and if that is not available to the default torch eager behaviour. If you have an RTX GPU and still see the issue, could you repot the

• python version
• torch + cuda version
• driver version
• operating system
• GPU

Python version: 3.12.10
torch==2.9.1+cu130
Driver: 591.74
Windows 11
4090

[Kosinkadink]

These aren't errors, just logs. If you try to run anything everything is fine. We may want to reformat these messages since they are causing confusion.

[aswordok]

[START] Security scan
[ComfyUI-Manager] python -m pip not available. Falling back to uv.
[ComfyUI-Manager] Using uv as Python module for pip operations.
[DONE] Security scan

ComfyUI-Manager: installing dependencies done.

** ComfyUI startup time: 2026-01-07 11:35:39.585
** Platform: Windows
** Python version: 3.13.8 (main, Oct 7 2025, 15:31:04) [MSC v.1944 64 bit (AMD64)]
** Python executable: D:+AI\uv\313.venv\Scripts\python.exe
** ComfyUI Path: D:+AI\uv\313\ComfyUI
** ComfyUI Base Folder Path: D:+AI\uv\313\ComfyUI
** User directory: D:+AI\uv\313\ComfyUI\user
** ComfyUI-Manager config path: D:+AI\uv\313\ComfyUI\user\default\ComfyUI-Manager\config.ini
** Log path: D:+AI\uv\313\ComfyUI\user\comfyui.log
[SAM3] ComfyUI-SAM3 prestartup script running...
[SAM3] Script directory: D:+AI\uv\313\ComfyUI\custom_nodes\ComfyUI-SAM3
[SAM3] ComfyUI root: D:+AI\uv\313\ComfyUI
[SAM3] Copying image assets to D:+AI\uv\313\ComfyUI\input...
[SAM3] Image assets: 0 copied, 3 skipped
[SAM3] Copying workflows to D:+AI\uv\313\ComfyUI\user\default\workflows...
[SAM3] Workflows: 0 copied, 5 skipped
[SAM3] Prestartup script completed

Prestartup times for custom nodes:
0.0 seconds: D:+AI\uv\313\ComfyUI\custom_nodes\rgthree-comfy
0.0 seconds: D:+AI\uv\313\ComfyUI\custom_nodes\ComfyUI-Easy-Use
0.0 seconds: D:+AI\uv\313\ComfyUI\custom_nodes\ComfyUI-SAM3
0.8 seconds: D:+AI\uv\313\ComfyUI\custom_nodes\comfyui-manager

D:+AI\uv\313.venv\Lib\site-packages\torch\cuda_init_.py:63: FutureWarning: The pynvml package is deprecated. Please install nvidia-ml-py instead. If you did not install pynvml directly, please report this to the maintainers of the package that installed pynvml for you.
import pynvml # type: ignore[import]
Checkpoint files will always be loaded safely.
Total VRAM 16380 MB, total RAM 65253 MB
pytorch version: 2.9.1+cu130
xformers version: 0.0.33+00a7a5f0.d20251020
Set vram state to: NORMAL_VRAM
Device: cuda:0 NVIDIA GeForce RTX 4060 Ti : cudaMallocAsync
Using async weight offloading with 2 streams
Enabled pinned memory 29364.0
working around nvidia conv3d memory bug.
Found comfy_kitchen backend eager: {'available': True, 'disabled': False, 'unavailable_reason': None, 'capabilities': ['apply_rope', 'apply_rope1', 'dequantize_nvfp4', 'dequantize_per_tensor_fp8', 'quantize_nvfp4', 'quantize_per_tensor_fp8', 'scaled_mm_nvfp4']}
Found comfy_kitchen backend cuda: {'available': False, 'disabled': False, 'unavailable_reason': ### 'DLL load failed while importing _C: 找不到指定的模块。', 'capabilities': []}
Found comfy_kitchen backend triton: {'available': True, 'disabled': True, 'unavailable_reason': None, 'capabilities': ['apply_rope', 'apply_rope1', 'dequantize_nvfp4', 'dequantize_per_tensor_fp8', 'quantize_nvfp4', 'quantize_per_tensor_fp8']}
Using sage attention
Python version: 3.13.8 (main, Oct 7 2025, 15:31:04) [MSC v.1944 64 bit (AMD64)]
ComfyUI version: 0.7.0
ComfyUI frontend version: 1.35.9
[Prompt Server] web root: D:+AI\uv\313.venv\Lib\site-packages\comfyui_frontend_package\static
Total VRAM 16380 MB, total RAM 65253 MB
pytorch version: 2.9.1+cu130
xformers version: 0.0.33+00a7a5f0.d20251020
Set vram state to: NORMAL_VRAM
Device: cuda:0 NVIDIA GeForce RTX 4060 Ti : cudaMallocAsync
Using async weight offloading with 2 streams
Enabled pinned memory 29364.0
Web extensions folder found at D:+AI\uv\313\ComfyUI\web\extensions\ComfyLiterals

[ai201414]

I just upgraded the Comfy-Kitchen wheel to version 0.2.3, but when I tried to launch COMFYUI, I received this message:
"Found comfy_kitchen backend cuda: {'available': True, 'disabled': False, 'unavailable_reason': None, 'capabilities': ['apply_rope', 'apply_rope1', 'dequantize_nvfp4', 'dequantize_per_tensor_fp8', 'quantize_nvfp4', 'quantize_per_tensor_fp8', 'scaled_mm_nvfp4']}"
Found comfy_kitchen backend triton: {'available': True, 'disabled': True, 'unavailable_reason': None, 'capabilities': ['apply_rope', 'apply_rope1', 'dequantize_nvfp4', 'dequantize_per_tensor_fp8', 'quantize_nvfp4', 'quantize_per_tensor_fp8']}
Found comfy_kitchen backend eager: {'available': True, 'disabled': False, 'unavailable_reason': None, 'capabilities': ['apply_rope', 'apply_rope1', 'dequantize_nvfp4', 'dequantize_per_tensor_fp8', 'quantize_nvfp4', 'quantize_per_tensor_fp8', 'scaled_mm_nvfp4']},

[praveenmaniyan]

Getting errors in ComfyUI for Nunchaku Z-Image. Getting a Cuda Illegal memory access issue. Details are here nunchaku-tech/nunchaku#865

[praveenmaniyan]

With the latest commit getting the following error:

loaded partially; 2419.36 MB usable, 2340.83 MB loaded, 263.30 MB offloaded, 78.53 MB buffer reserved, lowvram patches: 0
0%| | 0/9 [00:00<?, ?it/s]
!!! Exception during processing !!! CUDA kernel launch failed: an illegal memory access was encountered
Traceback (most recent call last):
File "/mnt/oldnvme/ComfyUI/execution.py", line 518, in execute
output_data, output_ui, has_subgraph, has_pending_tasks = await get_output_data(prompt_id, unique_id, obj, input_data_all, execution_block_cb=execution_block_cb, pre_execute_cb=pre_execute_cb, v3_data=v3_data)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/mnt/oldnvme/ComfyUI/execution.py", line 329, in get_output_data
return_values = await _async_map_node_over_list(prompt_id, unique_id, obj, input_data_all, obj.FUNCTION, allow_interrupt=True, execution_block_cb=execution_block_cb, pre_execute_cb=pre_execute_cb, v3_data=v3_data)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/mnt/oldnvme/ComfyUI/execution.py", line 303, in _async_map_node_over_list
await process_inputs(input_dict, i)
File "/mnt/oldnvme/ComfyUI/execution.py", line 291, in process_inputs
result = f(**inputs)
^^^^^^^^^^^
File "/mnt/oldnvme/ComfyUI/nodes.py", line 1542, in sample
return common_ksampler(model, seed, steps, cfg, sampler_name, scheduler, positive, negative, latent_image, denoise=denoise)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/mnt/oldnvme/ComfyUI/nodes.py", line 1509, in common_ksampler
samples = comfy.sample.sample(model, noise, steps, cfg, sampler_name, scheduler, positive, negative, latent_image,
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/mnt/oldnvme/ComfyUI/comfy/sample.py", line 60, in sample
samples = sampler.sample(noise, positive, negative, cfg=cfg, latent_image=latent_image, start_step=start_step, last_step=last_step, force_full_denoise=force_full_denoise, denoise_mask=noise_mask, sigmas=sigmas, callback=callback, disable_pbar=disable_pbar, seed=seed)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/mnt/oldnvme/ComfyUI/comfy/samplers.py", line 1178, in sample
return sample(self.model, noise, positive, negative, cfg, self.device, sampler, sigmas, self.model_options, latent_image=latent_image, denoise_mask=denoise_mask, callback=callback, disable_pbar=disable_pbar, seed=seed)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/mnt/oldnvme/ComfyUI/comfy/samplers.py", line 1068, in sample
return cfg_guider.sample(noise, latent_image, sampler, sigmas, denoise_mask, callback, disable_pbar, seed)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/mnt/oldnvme/ComfyUI/comfy/samplers.py", line 1050, in sample
output = executor.execute(noise, latent_image, sampler, sigmas, denoise_mask, callback, disable_pbar, seed, latent_shapes=latent_shapes)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/mnt/oldnvme/ComfyUI/comfy/patcher_extension.py", line 112, in execute
return self.original(*args, **kwargs)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/mnt/oldnvme/ComfyUI/comfy/samplers.py", line 994, in outer_sample
output = self.inner_sample(noise, latent_image, device, sampler, sigmas, denoise_mask, callback, disable_pbar, seed, latent_shapes=latent_shapes)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/mnt/oldnvme/ComfyUI/comfy/samplers.py", line 980, in inner_sample
samples = executor.execute(self, sigmas, extra_args, callback, noise, latent_image, denoise_mask, disable_pbar)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/mnt/oldnvme/ComfyUI/comfy/patcher_extension.py", line 112, in execute
return self.original(*args, **kwargs)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/mnt/oldnvme/ComfyUI/comfy/samplers.py", line 752, in sample
samples = self.sampler_function(model_k, noise, sigmas, extra_args=extra_args, callback=k_callback, disable=disable_pbar, **self.extra_options)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/home/pm/anaconda3/envs/comfygpu/lib/python3.12/site-packages/torch/utils/_contextlib.py", line 120, in decorate_context
return func(*args, **kwargs)
^^^^^^^^^^^^^^^^^^^^^
File "/mnt/oldnvme/ComfyUI/comfy/k_diffusion/sampling.py", line 1432, in sample_res_multistep
return res_multistep(model, x, sigmas, extra_args=extra_args, callback=callback, disable=disable, s_noise=s_noise, noise_sampler=noise_sampler, eta=0., cfg_pp=False)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/home/pm/anaconda3/envs/comfygpu/lib/python3.12/site-packages/torch/utils/_contextlib.py", line 120, in decorate_context
return func(*args, **kwargs)
^^^^^^^^^^^^^^^^^^^^^
File "/mnt/oldnvme/ComfyUI/comfy/k_diffusion/sampling.py", line 1390, in res_multistep
denoised = model(x, sigmas[i] * s_in, **extra_args)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/mnt/oldnvme/ComfyUI/comfy/samplers.py", line 401, in call
out = self.inner_model(x, sigma, model_options=model_options, seed=seed)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/mnt/oldnvme/ComfyUI/comfy/samplers.py", line 953, in call
return self.outer_predict_noise(*args, **kwargs)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/mnt/oldnvme/ComfyUI/comfy/samplers.py", line 960, in outer_predict_noise
).execute(x, timestep, model_options, seed)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/mnt/oldnvme/ComfyUI/comfy/patcher_extension.py", line 112, in execute
return self.original(*args, **kwargs)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/mnt/oldnvme/ComfyUI/comfy/samplers.py", line 963, in predict_noise
return sampling_function(self.inner_model, x, timestep, self.conds.get("negative", None), self.conds.get("positive", None), self.cfg, model_options=model_options, seed=seed)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/mnt/oldnvme/ComfyUI/comfy/samplers.py", line 381, in sampling_function
out = calc_cond_batch(model, conds, x, timestep, model_options)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/mnt/oldnvme/ComfyUI/comfy/samplers.py", line 206, in calc_cond_batch
return _calc_cond_batch_outer(model, conds, x_in, timestep, model_options)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/mnt/oldnvme/ComfyUI/comfy/samplers.py", line 214, in _calc_cond_batch_outer
return executor.execute(model, conds, x_in, timestep, model_options)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/mnt/oldnvme/ComfyUI/comfy/patcher_extension.py", line 112, in execute
return self.original(*args, **kwargs)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/mnt/oldnvme/ComfyUI/comfy/samplers.py", line 326, in calc_cond_batch
output = model.apply_model(input_x, timestep, **c).chunk(batch_chunks)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/mnt/oldnvme/ComfyUI/comfy/model_base.py", line 163, in apply_model
return comfy.patcher_extension.WrapperExecutor.new_class_executor(
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/mnt/oldnvme/ComfyUI/comfy/patcher_extension.py", line 112, in execute
return self.original(*args, **kwargs)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/mnt/oldnvme/ComfyUI/comfy/model_base.py", line 205, in _apply_model
model_output = self.diffusion_model(xc, t, context=context, control=control, transformer_options=transformer_options, **extra_conds)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/home/pm/anaconda3/envs/comfygpu/lib/python3.12/site-packages/torch/nn/modules/module.py", line 1775, in _wrapped_call_impl
return self._call_impl(*args, **kwargs)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/home/pm/anaconda3/envs/comfygpu/lib/python3.12/site-packages/torch/nn/modules/module.py", line 1786, in _call_impl
return forward_call(*args, **kwargs)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/mnt/oldnvme/ComfyUI/comfy/ldm/lumina/model.py", line 600, in forward
return comfy.patcher_extension.WrapperExecutor.new_class_executor(
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/mnt/oldnvme/ComfyUI/comfy/patcher_extension.py", line 112, in execute
return self.original(*args, **kwargs)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/mnt/oldnvme/ComfyUI/comfy/ldm/lumina/model.py", line 635, in _forward
img, mask, img_size, cap_size, freqs_cis = self.patchify_and_embed(x, cap_feats, cap_mask, adaln_input, num_tokens, transformer_options=transformer_options)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/mnt/oldnvme/ComfyUI/comfy/ldm/lumina/model.py", line 581, in patchify_and_embed
cap_feats = layer(cap_feats, cap_mask, freqs_cis[:, :cap_pos_ids.shape[1]], transformer_options=transformer_options)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/home/pm/anaconda3/envs/comfygpu/lib/python3.12/site-packages/torch/nn/modules/module.py", line 1775, in _wrapped_call_impl
return self._call_impl(*args, **kwargs)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/home/pm/anaconda3/envs/comfygpu/lib/python3.12/site-packages/torch/nn/modules/module.py", line 1786, in _call_impl
return forward_call(*args, **kwargs)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/mnt/oldnvme/ComfyUI/comfy/ldm/lumina/model.py", line 295, in forward
clamp_fp16(self.attention(
^^^^^^^^^^^^^^^
File "/home/pm/anaconda3/envs/comfygpu/lib/python3.12/site-packages/torch/nn/modules/module.py", line 1775, in _wrapped_call_impl
return self._call_impl(*args, **kwargs)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/home/pm/anaconda3/envs/comfygpu/lib/python3.12/site-packages/torch/nn/modules/module.py", line 1786, in _call_impl
return forward_call(*args, **kwargs)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/mnt/oldnvme/ComfyUI/custom_nodes/nunchaku_nodes/models/zimage.py", line 78, in forward
return super().forward(x, x_mask, freqs_cis, transformer_options)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/mnt/oldnvme/ComfyUI/comfy/ldm/lumina/model.py", line 114, in forward
xq, xk = apply_rope(xq, xk, freqs_cis)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/home/pm/anaconda3/envs/comfygpu/lib/python3.12/site-packages/comfy_kitchen/init.py", line 179, in apply_rope
return torch.ops.comfy_kitchen.apply_rope(xq, xk, freqs_cis)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/home/pm/anaconda3/envs/comfygpu/lib/python3.12/site-packages/torch/_ops.py", line 1255, in call
return self._op(*args, **kwargs)
^^^^^^^^^^^^^^^^^^^^^^^^^
File "/home/pm/anaconda3/envs/comfygpu/lib/python3.12/site-packages/torch/_library/custom_ops.py", line 343, in backend_impl
result = self._backend_fns[device_type](*args, **kwargs)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/home/pm/anaconda3/envs/comfygpu/lib/python3.12/site-packages/torch/_compile.py", line 53, in inner
return disable_fn(*args, **kwargs)
^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/home/pm/anaconda3/envs/comfygpu/lib/python3.12/site-packages/torch/_dynamo/eval_frame.py", line 1044, in _fn
return fn(*args, **kwargs)
^^^^^^^^^^^^^^^^^^^
File "/home/pm/anaconda3/envs/comfygpu/lib/python3.12/site-packages/torch/_library/custom_ops.py", line 376, in wrapped_fn
return fn(*args, **kwargs)
^^^^^^^^^^^^^^^^^^^
File "/home/pm/anaconda3/envs/comfygpu/lib/python3.12/site-packages/comfy_kitchen/backends/eager/rope.py", line 33, in _op_apply_rope
return impl(**kwargs)
^^^^^^^^^^^^^^
File "/home/pm/anaconda3/envs/comfygpu/lib/python3.12/site-packages/comfy_kitchen/backends/cuda/init.py", line 430, in apply_rope
_C.apply_rope(
RuntimeError: CUDA kernel launch failed: an illegal memory access was encountered

Prompt executed in 8.62 seconds
Aborted (core dumped)