Closed q1446227740 closed 2 months ago
size mismatch for proj_in.weight: copying a param with shape torch.Size([..., ...]) from checkpoint, the shape in current model is torch.Size([..., ...]) Any tensor size mismatch error is caused by the wrong combination of IPAdapter model, image encoder and/or base checkpoint.
All -vit-h models require the SD1.5 image encoder. At the moment only one SDXL model and the vit-G SD1.5 model need the bigger image encoder.
I don't understand what that means
please post the full error from the console
Error occurred when executing IPAdapterAdvanced:
Error(s) in loading state_dict for Resampler: size mismatch for proj_in.weight: copying a param with shape torch.Size([768, 1280]) from checkpoint, the shape in current model is torch.Size([768, 1664]).
File "D:\SD,ai\ComfyUI-aki-v1.3\execution.py", line 151, in recursive_execute output_data, output_ui = get_output_data(obj, input_data_all) File "D:\SD,ai\ComfyUI-aki-v1.3\execution.py", line 81, in get_output_data return_values = map_node_over_list(obj, input_data_all, obj.FUNCTION, allow_interrupt=True) File "D:\SD,ai\ComfyUI-aki-v1.3\execution.py", line 74, in map_node_over_list results.append(getattr(obj, func)(slice_dict(input_data_all, i))) File "D:\SD,ai\ComfyUI-aki-v1.3\custom_nodes\ComfyUI_IPAdapter_plus\IPAdapterPlus.py", line 673, in apply_ipadapter return (ipadapter_execute(model.clone(), ipadapter_model, clip_vision, ipa_args), ) File "D:\SD,ai\ComfyUI-aki-v1.3\custom_nodes\ComfyUI_IPAdapter_plus\IPAdapterPlus.py", line 324, in ipadapter_execute ipa = IPAdapter( File "D:\SD,ai\ComfyUI-aki-v1.3\custom_nodes\ComfyUI_IPAdapter_plus\IPAdapterPlus.py", line 68, in init self.image_proj_model.load_state_dict(ipadapter_model["image_proj"]) File "D:\SD,ai\ComfyUI-aki-v1.3\python\lib\site-packages\torch\nn\modules\module.py", line 2152, in load_state_dict raise RuntimeError('Error(s) in loading state_dict for {}:\n\t{}'.format(
please post the full error from the console
I know you're very busy, can you help me
you are using a wrong combination of models, you probably named them wrong. try to re-download the encoders and rename them as instructed in the documentation
您使用了错误的模型组合,您可能将它们命名错误。尝试重新下载编码器并按照文档中的说明重命名它们
Can I put the 'IP Adapter' model and 'faceid' together
yes
yes
I don't seem to have gotten these names wrong
I have created separate folders circled in red. Let me try it out
the image encoders! not the ipadapter models
Sorry, I'm not sure which one is "the image encoders". Can you mark it on "GitHub"
I don't seem to be making any mistakes with this one
Wtf! Sorry, they seem to be one file!! How could this be? I'm really sorry, thank you very much
What did I do! Maybe my network fluctuated and my download interrupted once. I made a mistake with the file!
Hey @cubiq, I have the same problem but the solution you gave didn't work for me, I don't know why. Mi potete aiutare?
Hello, I'm sorry, I'm a beginner and my English is not very good. I can only rely on translation software to read English, I haven't figured out the problem with "size mismatch for proj_in. weight" and haven't understood what you're saying, Can you explain in detail what 'SD1.5 image encoder' means? I don't understand 'clip vision encoder' I did use the sd1.5 Checkpoint model, I can only use 'VIT-G (medium strength)' at the moment, If you see it, I hope you can help me! Other friends also hope that you can help me, I have been sitting in front of the computer for 3 hours now! thanks!!