Wan2_2-Animate-14B_fp8_scaled_e4m3fn_KJ_v2.safetensors (v1 worked, this one doesnt)
I got a workflow with the v1 (which worked), switched it out with this v2 model. Got this error:
ComfyUI Error Report ## Error Details - Node ID: 56 - Node Type: WanVideoSampler - Exception Type: RuntimeError - Exception Message: self and mat2 must have the same dtype, but got Half and Float8_e4m3fn ## Stack Trace
File "/workspace/madapps/ComfyUI/execution.py", line 496, 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, hidden_inputs=hidden_inputs)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/workspace/madapps/ComfyUI/execution.py", line 315, 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, hidden_inputs=hidden_inputs)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/workspace/madapps/ComfyUI/execution.py", line 289, in _async_map_node_over_list
await process_inputs(input_dict, i)
File "/workspace/madapps/ComfyUI/execution.py", line 277, in process_inputs
result = f(**inputs)
^^^^^^^^^^^
File "/workspace/madapps/ComfyUI/custom_nodes/ComfyUI-WanVideoWrapper/nodes.py", line 4655, in process
raise e
File "/workspace/madapps/ComfyUI/custom_nodes/ComfyUI-WanVideoWrapper/nodes.py", line 4457, in process
noise_pred, self.cache_state = predict_with_cfg(
^^^^^^^^^^^^^^^^^
File "/workspace/madapps/ComfyUI/custom_nodes/ComfyUI-WanVideoWrapper/nodes.py", line 3199, in predict_with_cfg
raise e
File "/workspace/madapps/ComfyUI/custom_nodes/ComfyUI-WanVideoWrapper/nodes.py", line 3086, in predict_with_cfg
noise_pred_cond, cache_state_cond = transformer(
^^^^^^^^^^^^
File "/workspace/madapps/ComfyUI/.venv/lib/python3.12/site-packages/torch/nn/modules/module.py", line 1773, in _wrapped_call_impl
return self._call_impl(*args, **kwargs)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/workspace/madapps/ComfyUI/.venv/lib/python3.12/site-packages/torch/nn/modules/module.py", line 1784, in _call_impl
return forward_call(*args, **kwargs)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/workspace/madapps/ComfyUI/custom_nodes/ComfyUI-WanVideoWrapper/wanvideo/modules/model.py", line 2117, in forward
motion_vec = self.wananimate_face_embedding(wananim_face_pixel_values).to(x[0].dtype)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/workspace/madapps/ComfyUI/custom_nodes/ComfyUI-WanVideoWrapper/wanvideo/modules/model.py", line 1898, in wananimate_face_embedding
motion_vec = self.face_encoder(motion_vec.to(self.face_encoder.dtype))
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/workspace/madapps/ComfyUI/.venv/lib/python3.12/site-packages/torch/nn/modules/module.py", line 1773, in _wrapped_call_impl
return self._call_impl(*args, **kwargs)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/workspace/madapps/ComfyUI/.venv/lib/python3.12/site-packages/torch/nn/modules/module.py", line 1784, in _call_impl
return forward_call(*args, **kwargs)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/workspace/madapps/ComfyUI/custom_nodes/ComfyUI-WanVideoWrapper/wanvideo/modules/wananimate/face_blocks.py", line 61, in forward
x = self.out_proj(x)
^^^^^^^^^^^^^^^^
File "/workspace/madapps/ComfyUI/.venv/lib/python3.12/site-packages/torch/nn/modules/module.py", line 1773, in _wrapped_call_impl
return self._call_impl(*args, **kwargs)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/workspace/madapps/ComfyUI/.venv/lib/python3.12/site-packages/torch/nn/modules/module.py", line 1784, in _call_impl
return forward_call(*args, **kwargs)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/workspace/madapps/ComfyUI/.venv/lib/python3.12/site-packages/torch/nn/modules/linear.py", line 125, in forward
return F.linear(input, self.weight, self.bias)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
System Information - ComfyUI Version: 0.3.60 - Arguments: main.py --listen 0.0.0.0 --port 8188 - OS: posix - Python Version: 3.12.10 (main, Apr 9 2025, 08:55:05) [GCC 11.4.0] - Embedded Python: false - PyTorch Version: 2.8.0+cu128 ## Devices - Name: cuda:0 NVIDIA RTX PRO 6000 Blackwell Server Edition : cudaMallocAsync - Type: cuda - VRAM Total: 101974081536 - VRAM Free: 101388779520 - Torch VRAM Total: 0 - Torch VRAM Free: 0 ## Logs
2025-10-07T09:09:34.103186 -
2025-10-07T09:09:34.114135 - DWPose: Pose 10.72ms on 2 people
2025-10-07T09:09:34.114259 -
2025-10-07T09:09:34.201057 - DWPose: Bbox 9.00ms2025-10-07T09:09:34.201209 -
2025-10-07T09:09:34.211466 - DWPose: Pose 10.01ms on 2 people
2025-10-07T09:09:34.211606 -
2025-10-07T09:09:34.301899 - DWPose: Bbox 8.88ms2025-10-07T09:09:34.302049 -
2025-10-07T09:09:34.312837 - DWPose: Pose 10.57ms on 2 people
2025-10-07T09:09:34.313032 -
2025-10-07T09:09:34.401454 - DWPose: Bbox 8.77ms2025-10-07T09:09:34.401621 -
2025-10-07T09:09:34.412459 - DWPose: Pose 10.61ms on 2 people
2025-10-07T09:09:34.412601 -
2025-10-07T09:09:34.501179 - DWPose: Bbox 8.48ms2025-10-07T09:09:34.501326 -
2025-10-07T09:09:34.511832 - DWPose: Pose 10.30ms on 2 people
2025-10-07T09:09:34.511999 -
2025-10-07T09:09:34.595374 - DWPose: Bbox 72.26ms2025-10-07T09:09:34.595502 -
2025-10-07T09:09:34.605829 - DWPose: Pose 10.10ms on 2 people
2025-10-07T09:09:34.606098 -
2025-10-07T09:09:34.695980 - DWPose: Bbox 77.21ms2025-10-07T09:09:34.696144 -
2025-10-07T09:09:34.708189 - DWPose: Pose 11.75ms on 2 people
2025-10-07T09:09:34.708384 -
2025-10-07T09:09:34.732478 - DWPose: Bbox 9.19ms2025-10-07T09:09:34.732640 -
2025-10-07T09:09:34.798396 - DWPose: Pose 65.49ms on 2 people
2025-10-07T09:09:34.798583 -
2025-10-07T09:09:34.820602 - DWPose: Bbox 9.99ms2025-10-07T09:09:34.820748 -
2025-10-07T09:09:34.898470 - DWPose: Pose 77.48ms on 2 people
2025-10-07T09:09:34.898654 -
2025-10-07T09:09:34.922349 - DWPose: Bbox 9.35ms2025-10-07T09:09:34.922560 -
2025-10-07T09:09:35.004375 - DWPose: Pose 81.57ms on 2 people
2025-10-07T09:09:35.004528 -
2025-10-07T09:09:35.091309 - DWPose: Bbox 9.41ms2025-10-07T09:09:35.091594 -
2025-10-07T09:09:35.102221 - DWPose: Pose 10.39ms on 2 people
2025-10-07T09:09:35.102362 -
2025-10-07T09:09:35.195803 - DWPose: Bbox 79.24ms2025-10-07T09:09:35.195974 -
2025-10-07T09:09:35.206074 - DWPose: Pose 9.85ms on 2 people
2025-10-07T09:09:35.206204 -
2025-10-07T09:09:35.297997 - DWPose: Bbox 74.34ms2025-10-07T09:09:35.298162 -
2025-10-07T09:09:35.310015 - DWPose: Pose 11.60ms on 2 people
2025-10-07T09:09:35.310169 -
2025-10-07T09:09:35.399738 - DWPose: Bbox 79.04ms2025-10-07T09:09:35.399934 -
2025-10-07T09:09:35.410281 - DWPose: Pose 10.12ms on 2 people
2025-10-07T09:09:35.410409 -
2025-10-07T09:09:35.506080 - DWPose: Bbox 9.23ms2025-10-07T09:09:35.506272 -
2025-10-07T09:09:35.591765 - DWPose: Pose 85.24ms on 2 people
2025-10-07T09:09:35.592014 -
2025-10-07T09:09:35.611660 - DWPose: Bbox 9.16ms2025-10-07T09:09:35.611847 -
2025-10-07T09:09:35.623900 - DWPose: Pose 11.80ms on 2 people
2025-10-07T09:09:35.691547 -
2025-10-07T09:09:35.710091 - DWPose: Bbox 8.45ms2025-10-07T09:09:35.710236 -
2025-10-07T09:09:35.793734 - DWPose: Pose 83.25ms on 2 people
2025-10-07T09:09:35.793908 -
2025-10-07T09:09:35.812874 - DWPose: Bbox 8.57ms2025-10-07T09:09:35.813015 -
2025-10-07T09:09:35.894480 - DWPose: Pose 81.24ms on 2 people
2025-10-07T09:09:35.894650 -
2025-10-07T09:09:35.912713 - DWPose: Bbox 8.39ms2025-10-07T09:09:35.912898 -
2025-10-07T09:09:35.993966 - DWPose: Pose 80.85ms on 2 people
2025-10-07T09:09:35.994115 -
2025-10-07T09:09:36.017171 - DWPose: Bbox 10.38ms2025-10-07T09:09:36.017361 -
2025-10-07T09:09:36.027615 - DWPose: Pose 10.05ms on 2 people
2025-10-07T09:09:36.027747 -
2025-10-07T09:09:36.108938 - DWPose: Bbox 8.45ms2025-10-07T09:09:36.109106 -
2025-10-07T09:09:36.120200 - DWPose: Pose 10.88ms on 2 people
2025-10-07T09:09:36.120339 -
2025-10-07T09:09:36.209476 - DWPose: Bbox 8.38ms2025-10-07T09:09:36.209632 -
2025-10-07T09:09:36.219637 - DWPose: Pose 9.80ms on 2 people
2025-10-07T09:09:36.219792 -
2025-10-07T09:09:36.311552 - DWPose: Bbox 8.42ms2025-10-07T09:09:36.311732 -
2025-10-07T09:09:36.321938 - DWPose: Pose 9.97ms on 2 people
2025-10-07T09:09:36.322083 -
2025-10-07T09:09:36.413276 - DWPose: Bbox 10.31ms2025-10-07T09:09:36.413457 -
2025-10-07T09:09:36.496034 - DWPose: Pose 82.29ms on 2 people
2025-10-07T09:09:36.496311 -
2025-10-07T09:09:36.591334 - DWPose: Bbox 16.89ms2025-10-07T09:09:36.591634 -
2025-10-07T09:09:36.603668 - DWPose: Pose 11.78ms on 2 people
2025-10-07T09:09:36.603841 -
2025-10-07T09:09:36.692174 - DWPose: Bbox 76.98ms2025-10-07T09:09:36.692339 -
2025-10-07T09:09:36.703807 - DWPose: Pose 11.26ms on 2 people
2025-10-07T09:09:36.704114 -
2025-10-07T09:09:36.798432 - DWPose: Bbox 79.95ms2025-10-07T09:09:36.798630 -
2025-10-07T09:09:36.811096 - DWPose: Pose 12.17ms on 2 people
2025-10-07T09:09:36.811268 -
2025-10-07T09:09:36.902806 - DWPose: Bbox 8.95ms2025-10-07T09:09:36.902978 -
2025-10-07T09:09:36.913117 - DWPose: Pose 9.91ms on 2 people
2025-10-07T09:09:36.913286 -
2025-10-07T09:09:37.010850 - DWPose: Bbox 9.53ms2025-10-07T09:09:37.011016 -
2025-10-07T09:09:37.095547 - DWPose: Pose 84.33ms on 2 people
2025-10-07T09:09:37.095725 -
2025-10-07T09:09:37.115229 - DWPose: Bbox 8.99ms2025-10-07T09:09:37.115378 -
2025-10-07T09:09:37.195227 - DWPose: Pose 79.63ms on 2 people
2025-10-07T09:09:37.195409 -
2025-10-07T09:09:37.214448 - DWPose: Bbox 8.35ms2025-10-07T09:09:37.214595 -
2025-10-07T09:09:37.226146 - DWPose: Pose 11.35ms on 2 people
2025-10-07T09:09:37.226303 -
2025-10-07T09:09:37.309688 - DWPose: Bbox 8.37ms2025-10-07T09:09:37.309846 -
2025-10-07T09:09:37.393985 - DWPose: Pose 83.79ms on 2 people
2025-10-07T09:09:37.394208 -
2025-10-07T09:09:37.415316 - DWPose: Bbox 8.81ms2025-10-07T09:09:37.415446 -
2025-10-07T09:09:37.495279 - DWPose: Pose 79.57ms on 2 people
2025-10-07T09:09:37.495478 -
2025-10-07T09:09:37.514784 - DWPose: Bbox 8.80ms2025-10-07T09:09:37.514988 -
2025-10-07T09:09:37.593603 - DWPose: Pose 78.37ms on 2 people
2025-10-07T09:09:37.593750 -
2025-10-07T09:09:37.614854 - DWPose: Bbox 8.77ms2025-10-07T09:09:37.615031 -
2025-10-07T09:09:37.625382 - DWPose: Pose 10.16ms on 2 people
2025-10-07T09:09:37.625550 -
2025-10-07T09:09:37.710208 - DWPose: Bbox 8.94ms2025-10-07T09:09:37.710358 -
2025-10-07T09:09:37.721008 - DWPose: Pose 10.44ms on 2 people
2025-10-07T09:09:37.721149 -
2025-10-07T09:09:37.809276 - DWPose: Bbox 8.83ms2025-10-07T09:09:37.809415 -
2025-10-07T09:09:37.820295 - DWPose: Pose 10.63ms on 2 people
2025-10-07T09:09:37.820474 -
2025-10-07T09:09:37.909438 - DWPose: Bbox 8.79ms2025-10-07T09:09:37.909593 -
2025-10-07T09:09:37.920317 - DWPose: Pose 10.45ms on 2 people
2025-10-07T09:09:37.920464 -
2025-10-07T09:09:38.009794 - DWPose: Bbox 8.70ms2025-10-07T09:09:38.009936 -
2025-10-07T09:09:38.091925 - DWPose: Pose 81.71ms on 2 people
2025-10-07T09:09:38.092097 -
2025-10-07T09:09:38.112666 - DWPose: Bbox 8.80ms2025-10-07T09:09:38.112811 -
2025-10-07T09:09:38.196336 - DWPose: Pose 83.27ms on 2 people
2025-10-07T09:09:38.196489 -
2025-10-07T09:09:38.215824 - DWPose: Bbox 8.83ms2025-10-07T09:09:38.216118 -
2025-10-07T09:09:38.299323 - DWPose: Pose 82.94ms on 2 people
2025-10-07T09:09:38.299478 -
2025-10-07T09:09:38.393182 - DWPose: Bbox 78.43ms2025-10-07T09:09:38.393337 -
2025-10-07T09:09:38.404565 - DWPose: Pose 11.00ms on 2 people
2025-10-07T09:09:38.404705 -
2025-10-07T09:09:38.502584 - DWPose: Bbox 8.73ms2025-10-07T09:09:38.502756 -
2025-10-07T09:09:38.513534 - DWPose: Pose 10.54ms on 2 people
2025-10-07T09:09:38.513698 -
2025-10-07T09:09:38.613297 - DWPose: Bbox 8.71ms2025-10-07T09:09:38.613440 -
2025-10-07T09:09:38.693291 - DWPose: Pose 79.59ms on 2 people
2025-10-07T09:09:38.693468 -
2025-10-07T09:09:38.792118 - DWPose: Bbox 86.72ms2025-10-07T09:09:38.792282 -
2025-10-07T09:09:38.804827 - DWPose: Pose 12.36ms on 2 people
2025-10-07T09:09:38.804990 -
2025-10-07T09:09:38.897629 - DWPose: Bbox 80.67ms2025-10-07T09:09:38.897778 -
2025-10-07T09:09:38.908711 - DWPose: Pose 10.72ms on 2 people
2025-10-07T09:09:38.908872 -
2025-10-07T09:09:38.997417 - DWPose: Bbox 76.74ms2025-10-07T09:09:38.997572 -
2025-10-07T09:09:39.008243 - DWPose: Pose 10.47ms on 2 people
2025-10-07T09:09:39.008365 -
2025-10-07T09:09:39.102306 - DWPose: Bbox 8.78ms2025-10-07T09:09:39.102511 -
2025-10-07T09:09:39.112699 - DWPose: Pose 10.00ms on 2 people
2025-10-07T09:09:39.112873 -
2025-10-07T09:09:39.212071 - DWPose: Bbox 8.45ms2025-10-07T09:09:39.212220 -
2025-10-07T09:09:39.222369 - DWPose: Pose 9.95ms on 2 people
2025-10-07T09:09:39.222519 -
2025-10-07T09:09:39.320020 - DWPose: Bbox 11.06ms2025-10-07T09:09:39.320192 -
2025-10-07T09:09:39.392110 - DWPose: Pose 71.67ms on 2 people
2025-10-07T09:09:39.392287 -
2025-10-07T09:09:39.423197 - DWPose: Bbox 17.02ms2025-10-07T09:09:39.423369 -
2025-10-07T09:09:39.497458 - DWPose: Pose 73.85ms on 2 people
2025-10-07T09:09:39.497599 -
2025-10-07T09:09:39.594997 - DWPose: Bbox 84.46ms2025-10-07T09:09:39.595169 -
2025-10-07T09:09:39.607028 - DWPose: Pose 11.66ms on 2 people
2025-10-07T09:09:39.607163 -
2025-10-07T09:09:39.704447 - DWPose: Bbox 8.34ms2025-10-07T09:09:39.704674 -
2025-10-07T09:09:39.794994 - DWPose: Pose 90.14ms on 2 people
2025-10-07T09:09:39.795133 -
2025-10-07T09:09:39.815620 - DWPose: Bbox 8.67ms2025-10-07T09:09:39.815793 -
2025-10-07T09:09:39.895079 - DWPose: Pose 79.07ms on 2 people
2025-10-07T09:09:39.895228 -
2025-10-07T09:09:39.917480 - DWPose: Bbox 9.22ms2025-10-07T09:09:39.917607 -
2025-10-07T09:09:39.995519 - DWPose: Pose 77.69ms on 2 people
2025-10-07T09:09:39.995666 -
2025-10-07T09:09:40.014718 - DWPose: Bbox 8.52ms2025-10-07T09:09:40.014848 -
2025-10-07T09:09:40.094846 - DWPose: Pose 79.77ms on 2 people
2025-10-07T09:09:40.094991 -
2025-10-07T09:09:40.114388 - DWPose: Bbox 10.04ms2025-10-07T09:09:40.114512 -
2025-10-07T09:09:40.124451 - DWPose: Pose 9.77ms on 2 people
2025-10-07T09:09:40.124582 -
2025-10-07T09:09:40.209223 - DWPose: Bbox 8.62ms2025-10-07T09:09:40.209402 -
2025-10-07T09:09:40.219084 - DWPose: Pose 9.51ms on 2 people
2025-10-07T09:09:40.219197 -
2025-10-07T09:09:40.313076 - DWPose: Bbox 9.71ms2025-10-07T09:09:40.313225 -
2025-10-07T09:09:40.323383 - DWPose: Pose 9.96ms on 2 people
2025-10-07T09:09:40.323522 -
2025-10-07T09:09:40.412155 - DWPose: Bbox 9.12ms2025-10-07T09:09:40.412295 -
2025-10-07T09:09:40.423118 - DWPose: Pose 10.58ms on 2 people
2025-10-07T09:09:40.423671 -
2025-10-07T09:09:40.510498 - DWPose: Bbox 8.93ms2025-10-07T09:09:40.510924 -
2025-10-07T09:09:40.596143 - DWPose: Pose 84.96ms on 2 people
2025-10-07T09:09:40.596353 -
2025-10-07T09:09:40.616961 - DWPose: Bbox 10.18ms2025-10-07T09:09:40.617104 -
2025-10-07T09:09:40.696517 - DWPose: Pose 79.07ms on 2 people
2025-10-07T09:09:40.696673 -
2025-10-07T09:09:40.717561 - DWPose: Bbox 10.33ms2025-10-07T09:09:40.717732 -
2025-10-07T09:09:40.794708 - DWPose: Pose 76.75ms on 2 people
2025-10-07T09:09:40.794886 -
2025-10-07T09:09:40.815545 - DWPose: Bbox 9.78ms2025-10-07T09:09:40.815690 -
2025-10-07T09:09:40.895682 - DWPose: Pose 79.74ms on 2 people
2025-10-07T09:09:40.895865 -
2025-10-07T09:09:42.296786 - tensor.shape:2025-10-07T09:09:42.296963 - 2025-10-07T09:09:42.297103 - torch.Size([136, 1920, 1080, 3])2025-10-07T09:09:42.297198 -
2025-10-07T09:10:15.727427 - Requested to load CLIPVisionModelProjection
2025-10-07T09:10:15.733981 - loaded completely 93611.64956359864 1208.09814453125 True
2025-10-07T09:10:15.907984 - Clip embeds shape: torch.Size([1, 257, 1280]), dtype: torch.float32
2025-10-07T09:10:15.908391 - Combined clip embeds shape: torch.Size([1, 257, 1280])
2025-10-07T09:10:24.171541 - pose_latents2025-10-07T09:10:24.171768 - 2025-10-07T09:10:24.171892 - torch.Size([1, 16, 34, 160, 90])2025-10-07T09:10:24.172011 -
2025-10-07T09:10:43.991902 - CUDA Compute Capability: 12.0
2025-10-07T09:10:43.992386 - Detected model in_channels: 36
2025-10-07T09:10:43.992582 - Model cross attention type: i2v, num_heads: 40, num_layers: 40
2025-10-07T09:10:43.992837 - Model variant detected: 14B
2025-10-07T09:10:44.165745 - model_type FLOW
2025-10-07T09:10:44.238254 - Loading LoRA: WanAnimate_relight_lora_fp16 with strength: 1.0
2025-10-07T09:10:47.866520 - Loading LoRA: low_noise_model with strength: 1.0
2025-10-07T09:10:50.868204 - Loading LoRA: Jiggle/BounceLowWan2_2 with strength: 0.5
2025-10-07T09:10:53.214777 - Loading LoRA: Wan21_PusaV1_LoRA_14B_rank512_bf16 with strength: 1.0
2025-10-07T09:10:56.636927 - Loading LoRA: Wan2.2-Fun-A14B-InP-LOW-HPS2.1_resized_dynamic_avg_rank_15_bf16 with strength: 0.5
2025-10-07T09:10:57.242169 - Using accelerate to load and assign model weights to device...
2025-10-07T09:10:58.398977 -
Loading transformer parameters to cuda:0: 91%|█████████▏| 1304/1427 [00:01<00:00, 1787.92it/s]2025-10-07T09:11:03.118979 -
Loading transformer parameters to cuda:0: 100%|██████████| 1427/1427 [00:05<00:00, 242.92it/s] 2025-10-07T09:11:03.119190 -
2025-10-07T09:11:03.119894 - Using 486 LoRA weight patches for WanVideo model
2025-10-07T09:11:21.802477 - ------- Scheduler info -------
2025-10-07T09:11:21.824197 - Total timesteps: tensor([999, 957, 882, 714], device='cuda:0')
2025-10-07T09:11:21.827683 - Using timesteps: tensor([999, 957, 882, 714], device='cuda:0')
2025-10-07T09:11:21.863597 - Using sigmas: tensor([1.0000, 0.9574, 0.8823, 0.7142, 0.0000])
2025-10-07T09:11:21.863806 - ------------------------------
2025-10-07T09:11:21.864145 - sigmas: tensor([1.0000, 0.9574, 0.8823, 0.7142, 0.0000])
2025-10-07T09:11:23.552997 - WanAnimate: Face input torch.Size([1, 3, 136, 512, 512]) padded to shape torch.Size([1, 3, 154, 512, 512])
2025-10-07T09:11:23.554928 - WanAnimate: Ref masks torch.Size([4, 34, 160, 90]) padded to shape torch.Size([4, 34, 160, 90])
2025-10-07T09:11:23.709850 - WanAnimate: BG images torch.Size([3, 136, 1280, 720]) padded to shape torch.Size([3, 136, 1280, 720])
2025-10-07T09:11:23.710366 - Sampling 133 frames in 2 windows, at 720x1280 with 4 steps
2025-10-07T09:11:29.092201 -
Frames 0-77: 0%| | 0/4 [00:00<?, ?it/s]2025-10-07T09:11:29.809224 - Error during model prediction: self and mat2 must have the same dtype, but got Half and Float8_e4m3fn
2025-10-07T09:11:30.555142 - Error during sampling: self and mat2 must have the same dtype, but got Half and Float8_e4m3fn
2025-10-07T09:11:31.412154 - !!! Exception during processing !!! self and mat2 must have the same dtype, but got Half and Float8_e4m3fn
2025-10-07T09:11:31.443579 - Traceback (most recent call last):
File "/workspace/madapps/ComfyUI/execution.py", line 496, 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, hidden_inputs=hidden_inputs)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/workspace/madapps/ComfyUI/execution.py", line 315, 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, hidden_inputs=hidden_inputs)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/workspace/madapps/ComfyUI/execution.py", line 289, in _async_map_node_over_list
await process_inputs(input_dict, i)
File "/workspace/madapps/ComfyUI/execution.py", line 277, in process_inputs
result = f(**inputs)
^^^^^^^^^^^
File "/workspace/madapps/ComfyUI/custom_nodes/ComfyUI-WanVideoWrapper/nodes.py", line 4655, in process
raise e
File "/workspace/madapps/ComfyUI/custom_nodes/ComfyUI-WanVideoWrapper/nodes.py", line 4457, in process
noise_pred, self.cache_state = predict_with_cfg(
^^^^^^^^^^^^^^^^^
File "/workspace/madapps/ComfyUI/custom_nodes/ComfyUI-WanVideoWrapper/nodes.py", line 3199, in predict_with_cfg
raise e
File "/workspace/madapps/ComfyUI/custom_nodes/ComfyUI-WanVideoWrapper/nodes.py", line 3086, in predict_with_cfg
noise_pred_cond, cache_state_cond = transformer(
^^^^^^^^^^^^
File "/workspace/madapps/ComfyUI/.venv/lib/python3.12/site-packages/torch/nn/modules/module.py", line 1773, in _wrapped_call_impl
return self._call_impl(*args, **kwargs)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/workspace/madapps/ComfyUI/.venv/lib/python3.12/site-packages/torch/nn/modules/module.py", line 1784, in _call_impl
return forward_call(*args, **kwargs)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/workspace/madapps/ComfyUI/custom_nodes/ComfyUI-WanVideoWrapper/wanvideo/modules/model.py", line 2117, in forward
motion_vec = self.wananimate_face_embedding(wananim_face_pixel_values).to(x[0].dtype)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/workspace/madapps/ComfyUI/custom_nodes/ComfyUI-WanVideoWrapper/wanvideo/modules/model.py", line 1898, in wananimate_face_embedding
motion_vec = self.face_encoder(motion_vec.to(self.face_encoder.dtype))
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/workspace/madapps/ComfyUI/.venv/lib/python3.12/site-packages/torch/nn/modules/module.py", line 1773, in _wrapped_call_impl
return self._call_impl(*args, **kwargs)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/workspace/madapps/ComfyUI/.venv/lib/python3.12/site-packages/torch/nn/modules/module.py", line 1784, in _call_impl
return forward_call(*args, **kwargs)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/workspace/madapps/ComfyUI/custom_nodes/ComfyUI-WanVideoWrapper/wanvideo/modules/wananimate/face_blocks.py", line 61, in forward
x = self.out_proj(x)
^^^^^^^^^^^^^^^^
File "/workspace/madapps/ComfyUI/.venv/lib/python3.12/site-packages/torch/nn/modules/module.py", line 1773, in _wrapped_call_impl
return self._call_impl(*args, **kwargs)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/workspace/madapps/ComfyUI/.venv/lib/python3.12/site-packages/torch/nn/modules/module.py", line 1784, in _call_impl
return forward_call(*args, **kwargs)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/workspace/madapps/ComfyUI/.venv/lib/python3.12/site-packages/torch/nn/modules/linear.py", line 125, in forward
return F.linear(input, self.weight, self.bias)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
RuntimeError: self and mat2 must have the same dtype, but got Half and Float8_e4m3fn
2025-10-07T09:11:31.457739 - Prompt executed in 256.57 seconds
2025-10-07T09:11:32.153837 -
Frames 0-77: 0%|
This should be fixed now.
Note that the "v2" is mostly quantization bugfix for native workflows, and the benefit in the wrapper is only memory wise, quality may even be slightly worse as v1 doesn't have some layers in fp8 at all.
我和他一样的问题,使用V1版本的工作流和V1版本的animate正常,下载的先前的V2版本和现在的V2版本如果只使用姿态控制可以使用,加上面部和替换就出现和他一样的错误,但是如果把merge_loars开关打开,就可以使用了,不知道我这是不是属于个例还是咋地,其他博主做的视频看的运行的好好的,我必须把合并lora开关打开才行,但是用官方采样器的工作流没有此问题。但是我喜欢用KJ的流,因为运行的比官方的快