comfyanonymous
Support is required to run wan2.1 on the Ascend NPU.
Feature Idea
Support is required to enable the operation of wan2.1 on the Ascend NPU.
Existing Solutions
No response
Other
No response
The basic functionalities of ComfyUI, such as running Stable Diffusion, already work on Ascend. Did you see any error when running Wan?
Yes, wan2.1 encounters an error when using the uni_pc sampler. Switching to the euler sampler does not result in an error, but it generates completely black images.
The basic functionalities of ComfyUI, such as running Stable Diffusion, already work on Ascend. Did you see any error when running Wan?
I use the official sample workflow, 480 model, reported the following error, hope to get your advice, thank you first! 我用的是官方的示例工作流，480的模型，报如下错，希望得到你的指点，先谢为敬！
o see the GUI go to: http://127.0.0.1:8188
127.0.0.1: Client protocols [''] don’t overlap server-known ones ()
got prompt
Using pytorch attention in VAE
Using pytorch attention in VAE
VAE load device: npu:0, offload device: cpu, dtype: torch.bfloat16
Requested to load CLIPVisionModelProjection
loaded completely 60916.70234375 1208.09814453125 True
CLIP/text encoder model load device: npu:0, offload device: cpu, current: cpu, dtype: torch.float16
127.0.0.1: Client protocols [''] don’t overlap server-known ones ()
Requested to load WanTEModel
loaded completely 59657.76875 10835.4765625 True
Requested to load WanVAE
loaded completely 46630.3974609375 242.02829551696777 True
/home/ma-user/work/ComfyUI/comfy/ldm/wan/vae.py:209: UserWarning: current tensor is running as_strided, don't perform inplace operations on the returned value. If you encounter this warning and have precision issues, you can try torch.npu.config.allow_internal_format = False to resolve precision issues. (Triggered internally at build/CMakeFiles/torch_npu.dir/compiler_depend.ts:123.)
feat_cache[idx][:, :, -1, :, :].unsqueeze(2).to(
127.0.0.1: Client protocols [''] don’t overlap server-known ones ()
/home/ma-user/work/ComfyUI/custom_nodes/ComfyUI-GGUF/loader.py:91: UserWarning: The given NumPy array is not writable, and PyTorch does not support non-writable tensors. This means writing to this tensor will result in undefined behavior. You may want to copy the array to protect its data or make it writable before converting it to a tensor. This type of warning will be suppressed for the rest of this program. (Triggered internally at /pytorch/torch/csrc/utils/tensor_numpy.cpp:206.)
torch_tensor = torch.from_numpy(tensor.data) # mmap
model weight dtype torch.float16, manual cast: None
model_type FLOW
Requested to load WAN21
127.0.0.1: Client protocols [''] don’t overlap server-known ones ()
127.0.0.1: Client protocols [''] don’t overlap server-known ones ()
127.0.0.1: Client protocols [''] don’t overlap server-known ones ()
127.0.0.1: Client protocols [''] don’t overlap server-known ones ()
loaded completely 46230.6442546875 31734.013916015625 True
[ComfyUI-Easy-Use] server: v1.2.8 Loaded
[ComfyUI-Easy-Use] web root: /home/ma-user/work/ComfyUI/custom_nodes/ComfyUI-Easy-Use/web_version/v2 Loaded
gguf qtypes: F32 (823), F16 (480)
0%| | 0/20 [00:00<?, ?it/s][W404 21:26:55.710019296 compiler_depend.ts:387] Warning: EZ3002: 2025-04-04-21:26:55.052.111 Optype [Conv3D] of Ops kernel [AIcoreEngine] is unsupported. Reason: [tbe-custom]:op type Conv3D is not found in this op store.[tbe-custom]:op type Conv3D is not found in this op store.[Dynamic shape check]: data type DT_FLOAT of input [x] is not supported. All supported data type and format of tensor input0.x is: Data Type: {DT_FLOAT16,DT_INT8,DT_BFLOAT16}Format:{NDC1HWC0,NDC1HWC0,NDC1HWC0}.
Possible Cause: The operator type is unsupported in the operator information library due to specification mismatch.
Solution: Submit an issue to request for support at https://gitee.com/ascend, or remove this type of operators from your model.
TraceBack (most recent call last):
No supported Ops kernel and engine are found for [Conv3D81], optype [Conv3D].
Assert ((SelectEngine(node_ptr, exclude_engines, is_check_support_success, op_info)) == ge::SUCCESS) failed[FUNC:operator()][FILE:engine_place.cc][LINE:144]
build graph failed, graph id:80, ret:-1[FUNC:BuildModelWithGraphId][FILE:ge_generator.cc][LINE:1608]
[Build][SingleOpModel]call ge interface generator.BuildSingleOpModel failed. ge result = 4294967295[FUNC:ReportCallError][FILE:log_inner.cpp][LINE:161]
[Build][Op]Fail to build op model[FUNC:ReportInnerError][FILE:log_inner.cpp][LINE:145]
build op model failed, result = 500002[FUNC:ReportInnerError][FILE:log_inner.cpp][LINE:145]
(function ExecFunc)
0%| | 0/20 [00:00<?, ?it/s]
!!! Exception during processing !!! The Inner error is reported as above. The process exits for this inner error, and the current working operator name is Conv3D.
Since the operator is called asynchronously, the stacktrace may be inaccurate. If you want to get the accurate stacktrace, pleace set the environment variable ASCEND_LAUNCH_BLOCKING=1.
[ERROR] 2025-04-04-21:26:55 (PID:1310, Device:0, RankID:-1) ERR00100 PTA call acl api failed
Traceback (most recent call last):
File "/home/ma-user/work/ComfyUI/execution.py", line 327, in execute
output_data, output_ui, has_subgraph = get_output_data(obj, input_data_all, execution_block_cb=execution_block_cb, pre_execute_cb=pre_execute_cb)
File "/home/ma-user/work/ComfyUI/execution.py", line 202, in get_output_data
return_values = _map_node_over_list(obj, input_data_all, obj.FUNCTION, allow_interrupt=True, execution_block_cb=execution_block_cb, pre_execute_cb=pre_execute_cb)
File "/home/ma-user/work/ComfyUI/execution.py", line 174, in _map_node_over_list
process_inputs(input_dict, i)
File "/home/ma-user/work/ComfyUI/execution.py", line 163, in process_inputs
results.append(getattr(obj, func)(**inputs))
File "/home/ma-user/work/ComfyUI/nodes.py", line 1543, in sample
return common_ksampler(model, seed, steps, cfg, sampler_name, scheduler, positive, negative, latent_image, denoise=denoise)
File "/home/ma-user/work/ComfyUI/nodes.py", line 1510, in common_ksampler
samples = comfy.sample.sample(model, noise, steps, cfg, sampler_name, scheduler, positive, negative, latent_image,
File "/home/ma-user/work/ComfyUI/comfy/sample.py", line 45, 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 "/home/ma-user/work/ComfyUI/comfy/samplers.py", line 1133, 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 "/home/ma-user/work/ComfyUI/comfy/samplers.py", line 1023, in sample
return cfg_guider.sample(noise, latent_image, sampler, sigmas, denoise_mask, callback, disable_pbar, seed)
File "/home/ma-user/work/ComfyUI/comfy/samplers.py", line 1008, in sample
output = executor.execute(noise, latent_image, sampler, sigmas, denoise_mask, callback, disable_pbar, seed)
File "/home/ma-user/work/ComfyUI/comfy/patcher_extension.py", line 110, in execute
return self.original(*args, **kwargs)
File "/home/ma-user/work/ComfyUI/comfy/samplers.py", line 976, in outer_sample
output = self.inner_sample(noise, latent_image, device, sampler, sigmas, denoise_mask, callback, disable_pbar, seed)
File "/home/ma-user/work/ComfyUI/comfy/samplers.py", line 959, in inner_sample
samples = executor.execute(self, sigmas, extra_args, callback, noise, latent_image, denoise_mask, disable_pbar)
File "/home/ma-user/work/ComfyUI/comfy/patcher_extension.py", line 110, in execute
return self.original(*args, **kwargs)
File "/home/ma-user/work/ComfyUI/comfy/samplers.py", line 738, 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/ma-user/work/ComfyUI/comfy/extra_samplers/uni_pc.py", line 868, in sample_unipc
x = uni_pc.sample(noise, timesteps=timesteps, skip_type="time_uniform", method="multistep", order=order, lower_order_final=True, callback=callback, disable_pbar=disable)
File "/home/ma-user/work/ComfyUI/comfy/extra_samplers/uni_pc.py", line 715, in sample
model_prev_list = [self.model_fn(x, vec_t)]
File "/home/ma-user/work/ComfyUI/comfy/extra_samplers/uni_pc.py", line 410, in model_fn
return self.data_prediction_fn(x, t)
File "/home/ma-user/work/ComfyUI/comfy/extra_samplers/uni_pc.py", line 394, in data_prediction_fn
noise = self.noise_prediction_fn(x, t)
File "/home/ma-user/work/ComfyUI/comfy/extra_samplers/uni_pc.py", line 388, in noise_prediction_fn
return self.model(x, t)
File "/home/ma-user/work/ComfyUI/comfy/extra_samplers/uni_pc.py", line 329, in model_fn
return noise_pred_fn(x, t_continuous)
File "/home/ma-user/work/ComfyUI/comfy/extra_samplers/uni_pc.py", line 297, in noise_pred_fn
output = model(x, t_input, **model_kwargs)
File "/home/ma-user/work/ComfyUI/comfy/extra_samplers/uni_pc.py", line 859, in
I seem to be having the same problem, has it been resolved?
