Open artony35 opened 1 year ago
Hi @artony35, I would suggest you print both the structure of your pre-trained weights and the structure of the network to see if there is anything different. I guess it may be due to the "module.". Hi @tangy5, do you have any suggestions on how to be compatible when users load with their own pre-trained weights? Thanks in advance!
I am having some issues using a private pretrained encoder weights. When following the swin-unetr pretrain repo to train a pretrained encoder weights, the model can't be loaded for downstream segmentation training. Here is the error:
KeyError: 'module.patch_embed.proj.weight'
Is there anything that might be missing from the pretrain repo? Would it be possible to fix this issue without retraining? Thank you!!