第二次绘制图形报错

[328king]

当我第一次绘制图形时，一切正常，但第二次绘制时，就会报错。 Erroroccurred whenexecuting iclight diffusers sampler: Given groups=1,weight of size [320,8,3,3], expected input[2, 4, 64, 64] to have 8 channels, but got 4channels instead File“/root/ComfyUI/execution.py", line 151,in recursive executeoutput data, output ui = get output data(obj,input data all)File"/root/Comfyul/execution.py”line 8l, in get output datareturn values = map node over list(obi, input data all, obj.FUNCTION,allow interrupt=TrueFile “/root/ComfyuI/execution.py",line 74,in map node over listresults,append(getattr(obj, func)(slice_dict(input data all, i)))File "/root/ComfyuI/custom_nodes/ComfyUI-IC-Light-Wrapper/nodes. py"line 304, in processimages = pipe(File "/root/miniconda3/lib/python3.10/site packages/torch/utils/_contextlib.py", line 115, indecorate contextreturn fumc(*args, *kwargs)File"/root/miniconda3/lib/python3.10/sitepackages/diffusers/pipelines/stable_diffusion/pipeline_stable_diffusion. py", line 971, in __callnoise_pred = self.unet(File "/root/miniconda3/lib/python3. 10/site packages/torch/nn/modules/module.py", line 1501, in _call_implreturn forward call(args, kwargs)File "/root/miniconda3/lib/python3.10/site packages/accelerate/hooks.py", line 164, in new_forwardoutput = module. old forward(*args, :**kwargs)File "/root/mini conda3/lib/python3. 10/site-packages/diffusers/models/unets/unet_2d condition.py", line 1169,in forward

[HydraBucket]

Make sure your sizes are either 512, 768 or 1024. This will resolve.

[jiangdi0924]

same issue , Second generation error。 only reload comfyUI can fix it

[328king]

Make sure your sizes are either 512, 768 or 1024. This will resolve.

My image is 512 * 512， The first time it can be completed smoothly, but the second time it will report an error.