Skip to content
New issue

Have a question about this project? # for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “#”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? # to your account

Error(s) in loading state_dict for DataParallel: #27

Open
cosmolu opened this issue Oct 15, 2018 · 11 comments
Open

Error(s) in loading state_dict for DataParallel: #27

cosmolu opened this issue Oct 15, 2018 · 11 comments

Comments

@cosmolu
Copy link

cosmolu commented Oct 15, 2018

When I download the pre_trained model and resume it. there is an error.
model.load_state_dict(checkpoint['state_dict'])
It seems that the name are not matched.(e.g. "module.features.0.weight" v.s. "features.module.0.weight")
How could I solve it if I wish to use the pre_trained model on Cifar10?
Thank you !

Traceback (most recent call last):
File "test_0.py", line 130, in
model = load_model()
File "test_0.py", line 104, in load_model
model.load_state_dict(checkpoint['state_dict'])
File "/home/cosmo/anaconda3/envs/tf8/lib/python3.6/site-packages/torch/nn/modules/module.py", line 719, in load_state_dict
self.class.name, "\n\t".join(error_msgs)))
RuntimeError: Error(s) in loading state_dict for DataParallel:
Missing key(s) in state_dict: "module.features.0.weight", "module.features.0.bias", "module.features.3.weight", "module.features.3.bias", "module.features.6.weight", "module.features.6.bias", "module.features.8.weight", "module.features.8.bias", "module.features.10.weight", "module.features.10.bias", "module.classifier.weight", "module.classifier.bias".
Unexpected key(s) in state_dict: "features.module.0.weight", "features.module.0.bias", "features.module.3.weight", "features.module.3.bias", "features.module.6.weight", "features.module.6.bias", "features.module.8.weight", "features.module.8.bias", "features.module.10.weight", "features.module.10.bias", "classifier.weight", "classifier.bias".

@imirzadeh
Copy link

Hi,

The problem is the module is load with dataparallel activated and you are trying to load it without data parallel. That's why there's an extra module at the beginning of each key!

Refer to this link for more information:
https://discuss.pytorch.org/t/missing-keys-unexpected-keys-in-state-dict-when-loading-self-trained-model/22379

@goncalomordido
Copy link

You can also manually updated the dic. Like this:

        state_dict =checkpoint['state_dict']
        from collections import OrderedDict
        new_state_dict = OrderedDict()

        for k, v in state_dict.items():
            if 'module' not in k:
                k = 'module.'+k
            else:
                k = k.replace('features.module.', 'module.features.')
            new_state_dict[k]=v

        model.load_state_dict(new_state_dict)

@xjock
Copy link

xjock commented Apr 17, 2019

The mutiple GPUs usage in pytorch is a little difficult.
In TF, you just set the os.environ["CUDA_VISIBLE_DEVICES"]='0,1,2,3'

@raghav1810
Copy link

Am getting an error similar to this one
This is what I am running:
python cifar.py -a preresnet --depth 110 --epochs 3 --schedule 81 122 --gamma 0.1 --wd 1e-4 --checkpoint checkpoints/cifar10/preresnet-110 --resume 'checkpoint.pth.tar'
('checkpoint.pth.tar' is from the onedrive folder)

RuntimeError: Error(s) in loading state_dict for DataParallel:
Missing key(s) in state_dict: "module.bn.weight", "module.bn.bias", "module.bn.running_mean", "module.bn.running_var".
Unexpected key(s) in state_dict: "module.bn1.weight", "module.bn1.bias", "module.bn1.running_mean", "module.bn1.running_var", "module.layer1.0.conv3.weight", "module.layer1.0.bn3.weight", "module.layer1.0.bn3.bias", "module.layer1.0.bn3.running_mean", "module.layer1.0.bn3.running_var", "module.layer1.0.downsample.0.weight", "module.layer1.0.downsample.1.weight", "module.layer1.0.downsample.1.bias", "module.layer1.0.downsample.1.running_mean", "module.layer1.0.downsample.1.running_var", "module.layer1.1.conv3.weight", "module.layer1.1.bn3.weight", "module.layer1.1.bn3.bias", "module.layer1.1.bn3.running_mean", "module.layer1.1.bn3.running_var", "module.layer1.2.conv3.weight", "module.layer1.2.bn3.weight", "module.layer1.2.bn3.bias", "module.layer1.2.bn3.running_mean", "module.layer1.2.bn3.running_var", "module.layer1.3.conv3.weight", "module.layer1.3.bn3.weight", "module.layer1.3.bn3.bias", "module.layer1.3.bn3.running_mean", "module.layer1.3.bn3.running_var", "module.layer1.4.conv3.weight", "module.layer1.4.bn3.weight", "module.layer1.4.bn3.bias", "module.layer1.4.bn3.running_mean", "module.layer1.4.bn3.running_var", "module.layer1.5.conv3.weight", "module.layer1.5.bn3.weight", "module.layer1.5.bn3.bias", "module.layer1.5.bn3.running_mean", "module.layer1.5.bn3.running_var", "module.layer1.6.conv3.weight", "module.layer1.6.bn3.weight", "module.layer1.6.bn3.bias", "module.layer1.6.bn3.running_mean", "module.layer1.6.bn3.running_var", "module.layer1.7.conv3.weight", "module.layer1.7.bn3.weight", "module.layer1.7.bn3.bias", "module.layer1.7.bn3.running_mean", "module.layer1.7.bn3.running_var", "module.layer1.8.conv3.weight", "module.layer1.8.bn3.weight", "module.layer1.8.bn3.bias", "module.layer1.8.bn3.running_mean", "module.layer1.8.bn3.running_var", "module.layer1.9.conv3.weight", "module.layer1.9.bn3.weight", "module.layer1.9.bn3.bias", "module.layer1.9.bn3.running_mean", "module.layer1.9.bn3.running_var", "module.layer1.10.conv3.weight", "module.layer1.10.bn3.weight", "module.layer1.10.bn3.bias", "module.layer1.10.bn3.running_mean", "module.layer1.10.bn3.running_var", "module.layer1.11.conv3.weight", "module.layer1.11.bn3.weight", "module.layer1.11.bn3.bias", "module.layer1.11.bn3.running_mean", "module.layer1.11.bn3.running_var", "module.layer1.12.conv3.weight", "module.layer1.12.bn3.weight", "module.layer1.12.bn3.bias", "module.layer1.12.bn3.running_mean", "module.layer1.12.bn3.running_var", "module.layer1.13.conv3.weight", "module.layer1.13.bn3.weight", "module.layer1.13.bn3.bias", "module.layer1.13.bn3.running_mean", "module.layer1.13.bn3.running_var", "module.layer1.14.conv3.weight", "module.layer1.14.bn3.weight", "module.layer1.14.bn3.bias", "module.layer1.14.bn3.running_mean", "module.layer1.14.bn3.running_var", "module.layer1.15.conv3.weight", "module.layer1.15.bn3.weight", "module.layer1.15.bn3.bias", "module.layer1.15.bn3.running_mean", "module.layer1.15.bn3.running_var", "module.layer1.16.conv3.weight", "module.layer1.16.bn3.weight", "module.layer1.16.bn3.bias", "module.layer1.16.bn3.running_mean", "module.layer1.16.bn3.running_var", "module.layer1.17.conv3.weight", "module.layer1.17.bn3.weight", "module.layer1.17.bn3.bias", "module.layer1.17.bn3.running_mean", "module.layer1.17.bn3.running_var", "module.layer2.0.conv3.weight", "module.layer2.0.bn3.weight", "module.layer2.0.bn3.bias", "module.layer2.0.bn3.running_mean", "module.layer2.0.bn3.running_var", "module.layer2.0.downsample.1.weight", "module.layer2.0.downsample.1.bias", "module.layer2.0.downsample.1.running_mean", "module.layer2.0.downsample.1.running_var", "module.layer2.1.conv3.weight", "module.layer2.1.bn3.weight", "module.layer2.1.bn3.bias", "module.layer2.1.bn3.running_mean", "module.layer2.1.bn3.running_var", "module.layer2.2.conv3.weight", "module.layer2.2.bn3.weight", "module.layer2.2.bn3.bias", "module.layer2.2.bn3.running_mean", "module.layer2.2.bn3.running_var", "module.layer2.3.conv3.weight", "module.layer2.3.bn3.weight", "module.layer2.3.bn3.bias", "module.layer2.3.bn3.running_mean", "module.layer2.3.bn3.running_var", "module.layer2.4.conv3.weight", "module.layer2.4.bn3.weight", "module.layer2.4.bn3.bias", "module.layer2.4.bn3.running_mean", "module.layer2.4.bn3.running_var", "module.layer2.5.conv3.weight", "module.layer2.5.bn3.weight", "module.layer2.5.bn3.bias", "module.layer2.5.bn3.running_mean", "module.layer2.5.bn3.running_var", "module.layer2.6.conv3.weight", "module.layer2.6.bn3.weight", "module.layer2.6.bn3.bias", "module.layer2.6.bn3.running_mean", "module.layer2.6.bn3.running_var", "module.layer2.7.conv3.weight", "module.layer2.7.bn3.weight", "module.layer2.7.bn3.bias", "module.layer2.7.bn3.running_mean", "module.layer2.7.bn3.running_var", "module.layer2.8.conv3.weight", "module.layer2.8.bn3.weight", "module.layer2.8.bn3.bias", "module.layer2.8.bn3.running_mean", "module.layer2.8.bn3.running_var", "module.layer2.9.conv3.weight", "module.layer2.9.bn3.weight", "module.layer2.9.bn3.bias", "module.layer2.9.bn3.running_mean", "module.layer2.9.bn3.running_var", "module.layer2.10.conv3.weight", "module.layer2.10.bn3.weight", "module.layer2.10.bn3.bias", "module.layer2.10.bn3.running_mean", "module.layer2.10.bn3.running_var", "module.layer2.11.conv3.weight", "module.layer2.11.bn3.weight", "module.layer2.11.bn3.bias", "module.layer2.11.bn3.running_mean", "module.layer2.11.bn3.running_var", "module.layer2.12.conv3.weight", "module.layer2.12.bn3.weight", "module.layer2.12.bn3.bias", "module.layer2.12.bn3.running_mean", "module.layer2.12.bn3.running_var", "module.layer2.13.conv3.weight", "module.layer2.13.bn3.weight", "module.layer2.13.bn3.bias", "module.layer2.13.bn3.running_mean", "module.layer2.13.bn3.running_var", "module.layer2.14.conv3.weight", "module.layer2.14.bn3.weight", "module.layer2.14.bn3.bias", "module.layer2.14.bn3.running_mean", "module.layer2.14.bn3.running_var", "module.layer2.15.conv3.weight", "module.layer2.15.bn3.weight", "module.layer2.15.bn3.bias", "module.layer2.15.bn3.running_mean", "module.layer2.15.bn3.running_var", "module.layer2.16.conv3.weight", "module.layer2.16.bn3.weight", "module.layer2.16.bn3.bias", "module.layer2.16.bn3.running_mean", "module.layer2.16.bn3.running_var", "module.layer2.17.conv3.weight", "module.layer2.17.bn3.weight", "module.layer2.17.bn3.bias", "module.layer2.17.bn3.running_mean", "module.layer2.17.bn3.running_var", "module.layer3.0.conv3.weight", "module.layer3.0.bn3.weight", "module.layer3.0.bn3.bias", "module.layer3.0.bn3.running_mean", "module.layer3.0.bn3.running_var", "module.layer3.0.downsample.1.weight", "module.layer3.0.downsample.1.bias", "module.layer3.0.downsample.1.running_mean", "module.layer3.0.downsample.1.running_var", "module.layer3.1.conv3.weight", "module.layer3.1.bn3.weight", "module.layer3.1.bn3.bias", "module.layer3.1.bn3.running_mean", "module.layer3.1.bn3.running_var", "module.layer3.2.conv3.weight", "module.layer3.2.bn3.weight", "module.layer3.2.bn3.bias", "module.layer3.2.bn3.running_mean", "module.layer3.2.bn3.running_var", "module.layer3.3.conv3.weight", "module.layer3.3.bn3.weight", "module.layer3.3.bn3.bias", "module.layer3.3.bn3.running_mean", "module.layer3.3.bn3.running_var", "module.layer3.4.conv3.weight", "module.layer3.4.bn3.weight", "module.layer3.4.bn3.bias", "module.layer3.4.bn3.running_mean", "module.layer3.4.bn3.running_var", "module.layer3.5.conv3.weight", "module.layer3.5.bn3.weight", "module.layer3.5.bn3.bias", "module.layer3.5.bn3.running_mean", "module.layer3.5.bn3.running_var", "module.layer3.6.conv3.weight", "module.layer3.6.bn3.weight", "module.layer3.6.bn3.bias", "module.layer3.6.bn3.running_mean", "module.layer3.6.bn3.running_var", "module.layer3.7.conv3.weight", "module.layer3.7.bn3.weight", "module.layer3.7.bn3.bias", "module.layer3.7.bn3.running_mean", "module.layer3.7.bn3.running_var", "module.layer3.8.conv3.weight", "module.layer3.8.bn3.weight", "module.layer3.8.bn3.bias", "module.layer3.8.bn3.running_mean", "module.layer3.8.bn3.running_var", "module.layer3.9.conv3.weight", "module.layer3.9.bn3.weight", "module.layer3.9.bn3.bias", "module.layer3.9.bn3.running_mean", "module.layer3.9.bn3.running_var", "module.layer3.10.conv3.weight", "module.layer3.10.bn3.weight", "module.layer3.10.bn3.bias", "module.layer3.10.bn3.running_mean", "module.layer3.10.bn3.running_var", "module.layer3.11.conv3.weight", "module.layer3.11.bn3.weight", "module.layer3.11.bn3.bias", "module.layer3.11.bn3.running_mean", "module.layer3.11.bn3.running_var", "module.layer3.12.conv3.weight", "module.layer3.12.bn3.weight", "module.layer3.12.bn3.bias", "module.layer3.12.bn3.running_mean", "module.layer3.12.bn3.running_var", "module.layer3.13.conv3.weight", "module.layer3.13.bn3.weight", "module.layer3.13.bn3.bias", "module.layer3.13.bn3.running_mean", "module.layer3.13.bn3.running_var", "module.layer3.14.conv3.weight", "module.layer3.14.bn3.weight", "module.layer3.14.bn3.bias", "module.layer3.14.bn3.running_mean", "module.layer3.14.bn3.running_var", "module.layer3.15.conv3.weight", "module.layer3.15.bn3.weight", "module.layer3.15.bn3.bias", "module.layer3.15.bn3.running_mean", "module.layer3.15.bn3.running_var", "module.layer3.16.conv3.weight", "module.layer3.16.bn3.weight", "module.layer3.16.bn3.bias", "module.layer3.16.bn3.running_mean", "module.layer3.16.bn3.running_var", "module.layer3.17.conv3.weight", "module.layer3.17.bn3.weight", "module.layer3.17.bn3.bias", "module.layer3.17.bn3.running_mean", "module.layer3.17.bn3.running_var".
size mismatch for module.layer1.0.conv1.weight: copying a param with shape torch.Size([16, 16, 1, 1]) from checkpoint, the shape in current model is torch.Size([16, 16, 3, 3]).
size mismatch for module.layer1.1.conv1.weight: copying a param with shape torch.Size([16, 64, 1, 1]) from checkpoint, the shape in current model is torch.Size([16, 16, 3, 3]).
size mismatch for module.layer1.2.conv1.weight: copying a param with shape torch.Size([16, 64, 1, 1]) from checkpoint, the shape in current model is torch.Size([16, 16, 3, 3]).
size mismatch for module.layer1.3.conv1.weight: copying a param with shape torch.Size([16, 64, 1, 1]) from checkpoint, the shape in current model is torch.Size([16, 16, 3, 3]).
size mismatch for module.layer1.4.conv1.weight: copying a param with shape torch.Size([16, 64, 1, 1]) from checkpoint, the shape in current model is torch.Size([16, 16, 3, 3]).
size mismatch for module.layer1.5.conv1.weight: copying a param with shape torch.Size([16, 64, 1, 1]) from checkpoint, the shape in current model is torch.Size([16, 16, 3, 3]).
size mismatch for module.layer1.6.conv1.weight: copying a param with shape torch.Size([16, 64, 1, 1]) from checkpoint, the shape in current model is torch.Size([16, 16, 3, 3]).
size mismatch for module.layer1.7.conv1.weight: copying a param with shape torch.Size([16, 64, 1, 1]) from checkpoint, the shape in current model is torch.Size([16, 16, 3, 3]).
size mismatch for module.layer1.8.conv1.weight: copying a param with shape torch.Size([16, 64, 1, 1]) from checkpoint, the shape in current model is torch.Size([16, 16, 3, 3]).
size mismatch for module.layer1.9.conv1.weight: copying a param with shape torch.Size([16, 64, 1, 1]) from checkpoint, the shape in current model is torch.Size([16, 16, 3, 3]).
size mismatch for module.layer1.10.conv1.weight: copying a param with shape torch.Size([16, 64, 1, 1]) from checkpoint, the shape in current model is torch.Size([16, 16, 3, 3]).
size mismatch for module.layer1.11.conv1.weight: copying a param with shape torch.Size([16, 64, 1, 1]) from checkpoint, the shape in current model is torch.Size([16, 16, 3, 3]).
size mismatch for module.layer1.12.conv1.weight: copying a param with shape torch.Size([16, 64, 1, 1]) from checkpoint, the shape in current model is torch.Size([16, 16, 3, 3]).
size mismatch for module.layer1.13.conv1.weight: copying a param with shape torch.Size([16, 64, 1, 1]) from checkpoint, the shape in current model is torch.Size([16, 16, 3, 3]).
size mismatch for module.layer1.14.conv1.weight: copying a param with shape torch.Size([16, 64, 1, 1]) from checkpoint, the shape in current model is torch.Size([16, 16, 3, 3]).
size mismatch for module.layer1.15.conv1.weight: copying a param with shape torch.Size([16, 64, 1, 1]) from checkpoint, the shape in current model is torch.Size([16, 16, 3, 3]).
size mismatch for module.layer1.16.conv1.weight: copying a param with shape torch.Size([16, 64, 1, 1]) from checkpoint, the shape in current model is torch.Size([16, 16, 3, 3]).
size mismatch for module.layer1.17.conv1.weight: copying a param with shape torch.Size([16, 64, 1, 1]) from checkpoint, the shape in current model is torch.Size([16, 16, 3, 3]).
size mismatch for module.layer2.0.bn1.weight: copying a param with shape torch.Size([32]) from checkpoint, the shape in current model is torch.Size([16]).
size mismatch for module.layer2.0.bn1.bias: copying a param with shape torch.Size([32]) from checkpoint, the shape in current model is torch.Size([16]).
size mismatch for module.layer2.0.bn1.running_mean: copying a param with shape torch.Size([32]) from checkpoint, the shape in current model is torch.Size([16]).
size mismatch for module.layer2.0.bn1.running_var: copying a param with shape torch.Size([32]) from checkpoint, the shape in current model is torch.Size([16]).
size mismatch for module.layer2.0.conv1.weight: copying a param with shape torch.Size([32, 64, 1, 1]) from checkpoint, the shape in current model is torch.Size([32, 16, 3, 3]).
size mismatch for module.layer2.0.downsample.0.weight: copying a param with shape torch.Size([128, 64, 1, 1]) from checkpoint, the shape in current model is torch.Size([32, 16, 1, 1]).
size mismatch for module.layer2.1.conv1.weight: copying a param with shape torch.Size([32, 128, 1, 1]) from checkpoint, the shape in current model is torch.Size([32, 32, 3, 3]).
size mismatch for module.layer2.2.conv1.weight: copying a param with shape torch.Size([32, 128, 1, 1]) from checkpoint, the shape in current model is torch.Size([32, 32, 3, 3]).
size mismatch for module.layer2.3.conv1.weight: copying a param with shape torch.Size([32, 128, 1, 1]) from checkpoint, the shape in current model is torch.Size([32, 32, 3, 3]).
size mismatch for module.layer2.4.conv1.weight: copying a param with shape torch.Size([32, 128, 1, 1]) from checkpoint, the shape in current model is torch.Size([32, 32, 3, 3]).
size mismatch for module.layer2.5.conv1.weight: copying a param with shape torch.Size([32, 128, 1, 1]) from checkpoint, the shape in current model is torch.Size([32, 32, 3, 3]).
size mismatch for module.layer2.6.conv1.weight: copying a param with shape torch.Size([32, 128, 1, 1]) from checkpoint, the shape in current model is torch.Size([32, 32, 3, 3]).
size mismatch for module.layer2.7.conv1.weight: copying a param with shape torch.Size([32, 128, 1, 1]) from checkpoint, the shape in current model is torch.Size([32, 32, 3, 3]).
size mismatch for module.layer2.8.conv1.weight: copying a param with shape torch.Size([32, 128, 1, 1]) from checkpoint, the shape in current model is torch.Size([32, 32, 3, 3]).
size mismatch for module.layer2.9.conv1.weight: copying a param with shape torch.Size([32, 128, 1, 1]) from checkpoint, the shape in current model is torch.Size([32, 32, 3, 3]).
size mismatch for module.layer2.10.conv1.weight: copying a param with shape torch.Size([32, 128, 1, 1]) from checkpoint, the shape in current model is torch.Size([32, 32, 3, 3]).
size mismatch for module.layer2.11.conv1.weight: copying a param with shape torch.Size([32, 128, 1, 1]) from checkpoint, the shape in current model is torch.Size([32, 32, 3, 3]).
size mismatch for module.layer2.12.conv1.weight: copying a param with shape torch.Size([32, 128, 1, 1]) from checkpoint, the shape in current model is torch.Size([32, 32, 3, 3]).
size mismatch for module.layer2.13.conv1.weight: copying a param with shape torch.Size([32, 128, 1, 1]) from checkpoint, the shape in current model is torch.Size([32, 32, 3, 3]).
size mismatch for module.layer2.14.conv1.weight: copying a param with shape torch.Size([32, 128, 1, 1]) from checkpoint, the shape in current model is torch.Size([32, 32, 3, 3]).
size mismatch for module.layer2.15.conv1.weight: copying a param with shape torch.Size([32, 128, 1, 1]) from checkpoint, the shape in current model is torch.Size([32, 32, 3, 3]).
size mismatch for module.layer2.16.conv1.weight: copying a param with shape torch.Size([32, 128, 1, 1]) from checkpoint, the shape in current model is torch.Size([32, 32, 3, 3]).
size mismatch for module.layer2.17.conv1.weight: copying a param with shape torch.Size([32, 128, 1, 1]) from checkpoint, the shape in current model is torch.Size([32, 32, 3, 3]).
size mismatch for module.layer3.0.bn1.weight: copying a param with shape torch.Size([64]) from checkpoint, the shape in current model is torch.Size([32]).
size mismatch for module.layer3.0.bn1.bias: copying a param with shape torch.Size([64]) from checkpoint, the shape in current model is torch.Size([32]).
size mismatch for module.layer3.0.bn1.running_mean: copying a param with shape torch.Size([64]) from checkpoint, the shape in current model is torch.Size([32]).
size mismatch for module.layer3.0.bn1.running_var: copying a param with shape torch.Size([64]) from checkpoint, the shape in current model is torch.Size([32]).
size mismatch for module.layer3.0.conv1.weight: copying a param with shape torch.Size([64, 128, 1, 1]) from checkpoint, the shape in current model is torch.Size([64, 32, 3, 3]).
size mismatch for module.layer3.0.downsample.0.weight: copying a param with shape torch.Size([256, 128, 1, 1]) from checkpoint, the shape in current model is torch.Size([64, 32, 1, 1]).
size mismatch for module.layer3.1.conv1.weight: copying a param with shape torch.Size([64, 256, 1, 1]) from checkpoint, the shape in current model is torch.Size([64, 64, 3, 3]).
size mismatch for module.layer3.2.conv1.weight: copying a param with shape torch.Size([64, 256, 1, 1]) from checkpoint, the shape in current model is torch.Size([64, 64, 3, 3]).
size mismatch for module.layer3.3.conv1.weight: copying a param with shape torch.Size([64, 256, 1, 1]) from checkpoint, the shape in current model is torch.Size([64, 64, 3, 3]).
size mismatch for module.layer3.4.conv1.weight: copying a param with shape torch.Size([64, 256, 1, 1]) from checkpoint, the shape in current model is torch.Size([64, 64, 3, 3]).
size mismatch for module.layer3.5.conv1.weight: copying a param with shape torch.Size([64, 256, 1, 1]) from checkpoint, the shape in current model is torch.Size([64, 64, 3, 3]).
size mismatch for module.layer3.6.conv1.weight: copying a param with shape torch.Size([64, 256, 1, 1]) from checkpoint, the shape in current model is torch.Size([64, 64, 3, 3]).
size mismatch for module.layer3.7.conv1.weight: copying a param with shape torch.Size([64, 256, 1, 1]) from checkpoint, the shape in current model is torch.Size([64, 64, 3, 3]).
size mismatch for module.layer3.8.conv1.weight: copying a param with shape torch.Size([64, 256, 1, 1]) from checkpoint, the shape in current model is torch.Size([64, 64, 3, 3]).
size mismatch for module.layer3.9.conv1.weight: copying a param with shape torch.Size([64, 256, 1, 1]) from checkpoint, the shape in current model is torch.Size([64, 64, 3, 3]).
size mismatch for module.layer3.10.conv1.weight: copying a param with shape torch.Size([64, 256, 1, 1]) from checkpoint, the shape in current model is torch.Size([64, 64, 3, 3]).
size mismatch for module.layer3.11.conv1.weight: copying a param with shape torch.Size([64, 256, 1, 1]) from checkpoint, the shape in current model is torch.Size([64, 64, 3, 3]).
size mismatch for module.layer3.12.conv1.weight: copying a param with shape torch.Size([64, 256, 1, 1]) from checkpoint, the shape in current model is torch.Size([64, 64, 3, 3]).
size mismatch for module.layer3.13.conv1.weight: copying a param with shape torch.Size([64, 256, 1, 1]) from checkpoint, the shape in current model is torch.Size([64, 64, 3, 3]).
size mismatch for module.layer3.14.conv1.weight: copying a param with shape torch.Size([64, 256, 1, 1]) from checkpoint, the shape in current model is torch.Size([64, 64, 3, 3]).
size mismatch for module.layer3.15.conv1.weight: copying a param with shape torch.Size([64, 256, 1, 1]) from checkpoint, the shape in current model is torch.Size([64, 64, 3, 3]).
size mismatch for module.layer3.16.conv1.weight: copying a param with shape torch.Size([64, 256, 1, 1]) from checkpoint, the shape in current model is torch.Size([64, 64, 3, 3]).
size mismatch for module.layer3.17.conv1.weight: copying a param with shape torch.Size([64, 256, 1, 1]) from checkpoint, the shape in current model is torch.Size([64, 64, 3, 3]).
size mismatch for module.fc.weight: copying a param with shape torch.Size([100, 256]) from checkpoint, the shape in current model is torch.Size([10, 64]).
size mismatch for module.fc.bias: copying a param with shape torch.Size([100]) from checkpoint, the shape in current model is torch.Size([10]).

@jiteshm17
Copy link

You can also manually updated the dic. Like this:

        state_dict =checkpoint['state_dict']
        from collections import OrderedDict
        new_state_dict = OrderedDict()

        for k, v in state_dict.items():
            if 'module' not in k:
                k = 'module.'+k
            else:
                k = k.replace('features.module.', 'module.features.')
            new_state_dict[k]=v

        model.load_state_dict(new_state_dict)

I suppose this issue can be closed as the referenced post mentions the cause of the error and offers a solution

@yustiks
Copy link

yustiks commented Jan 28, 2021

You can also manually updated the dic. Like this:

        state_dict =checkpoint['state_dict']
        from collections import OrderedDict
        new_state_dict = OrderedDict()

        for k, v in state_dict.items():
            if 'module' not in k:
                k = 'module.'+k
            else:
                k = k.replace('features.module.', 'module.features.')
            new_state_dict[k]=v

        model.load_state_dict(new_state_dict)

life saver!

@ersamo
Copy link

ersamo commented Sep 27, 2021

Am getting an error similar to this one
This is what I am running:
python cifar.py -a preresnet --depth 110 --epochs 3 --schedule 81 122 --gamma 0.1 --wd 1e-4 --checkpoint checkpoints/cifar10/preresnet-110 --resume 'checkpoint.pth.tar'
('checkpoint.pth.tar' is from the onedrive folder)

RuntimeError: Error(s) in loading state_dict for DataParallel:
Missing key(s) in state_dict: "module.bn.weight", "module.bn.bias", "module.bn.running_mean", "module.bn.running_var".
Unexpected key(s) in state_dict: "module.bn1.weight", "module.bn1.bias", "module.bn1.running_mean", "module.bn1.running_var", "module.layer1.0.conv3.weight", "module.layer1.0.bn3.weight", "module.layer1.0.bn3.bias", "module.layer1.0.bn3.running_mean", "module.layer1.0.bn3.running_var", "module.layer1.0.downsample.0.weight", "module.layer1.0.downsample.1.weight", "module.layer1.0.downsample.1.bias", "module.layer1.0.downsample.1.running_mean", "module.layer1.0.downsample.1.running_var", "module.layer1.1.conv3.weight", "module.layer1.1.bn3.weight", "module.layer1.1.bn3.bias", "module.layer1.1.bn3.running_mean", "module.layer1.1.bn3.running_var", "module.layer1.2.conv3.weight", "module.layer1.2.bn3.weight", "module.layer1.2.bn3.bias", "module.layer1.2.bn3.running_mean", "module.layer1.2.bn3.running_var", "module.layer1.3.conv3.weight", "module.layer1.3.bn3.weight", "module.layer1.3.bn3.bias", "module.layer1.3.bn3.running_mean", "module.layer1.3.bn3.running_var", "module.layer1.4.conv3.weight", "module.layer1.4.bn3.weight", "module.layer1.4.bn3.bias", "module.layer1.4.bn3.running_mean", "module.layer1.4.bn3.running_var", "module.layer1.5.conv3.weight", "module.layer1.5.bn3.weight", "module.layer1.5.bn3.bias", "module.layer1.5.bn3.running_mean", "module.layer1.5.bn3.running_var", "module.layer1.6.conv3.weight", "module.layer1.6.bn3.weight", "module.layer1.6.bn3.bias", "module.layer1.6.bn3.running_mean", "module.layer1.6.bn3.running_var", "module.layer1.7.conv3.weight", "module.layer1.7.bn3.weight", "module.layer1.7.bn3.bias", "module.layer1.7.bn3.running_mean", "module.layer1.7.bn3.running_var", "module.layer1.8.conv3.weight", "module.layer1.8.bn3.weight", "module.layer1.8.bn3.bias", "module.layer1.8.bn3.running_mean", "module.layer1.8.bn3.running_var", "module.layer1.9.conv3.weight", "module.layer1.9.bn3.weight", "module.layer1.9.bn3.bias", "module.layer1.9.bn3.running_mean", "module.layer1.9.bn3.running_var", "module.layer1.10.conv3.weight", "module.layer1.10.bn3.weight", "module.layer1.10.bn3.bias", "module.layer1.10.bn3.running_mean", "module.layer1.10.bn3.running_var", "module.layer1.11.conv3.weight", "module.layer1.11.bn3.weight", "module.layer1.11.bn3.bias", "module.layer1.11.bn3.running_mean", "module.layer1.11.bn3.running_var", "module.layer1.12.conv3.weight", "module.layer1.12.bn3.weight", "module.layer1.12.bn3.bias", "module.layer1.12.bn3.running_mean", "module.layer1.12.bn3.running_var", "module.layer1.13.conv3.weight", "module.layer1.13.bn3.weight", "module.layer1.13.bn3.bias", "module.layer1.13.bn3.running_mean", "module.layer1.13.bn3.running_var", "module.layer1.14.conv3.weight", "module.layer1.14.bn3.weight", "module.layer1.14.bn3.bias", "module.layer1.14.bn3.running_mean", "module.layer1.14.bn3.running_var", "module.layer1.15.conv3.weight", "module.layer1.15.bn3.weight", "module.layer1.15.bn3.bias", "module.layer1.15.bn3.running_mean", "module.layer1.15.bn3.running_var", "module.layer1.16.conv3.weight", "module.layer1.16.bn3.weight", "module.layer1.16.bn3.bias", "module.layer1.16.bn3.running_mean", "module.layer1.16.bn3.running_var", "module.layer1.17.conv3.weight", "module.layer1.17.bn3.weight", "module.layer1.17.bn3.bias", "module.layer1.17.bn3.running_mean", "module.layer1.17.bn3.running_var", "module.layer2.0.conv3.weight", "module.layer2.0.bn3.weight", "module.layer2.0.bn3.bias", "module.layer2.0.bn3.running_mean", "module.layer2.0.bn3.running_var", "module.layer2.0.downsample.1.weight", "module.layer2.0.downsample.1.bias", "module.layer2.0.downsample.1.running_mean", "module.layer2.0.downsample.1.running_var", "module.layer2.1.conv3.weight", "module.layer2.1.bn3.weight", "module.layer2.1.bn3.bias", "module.layer2.1.bn3.running_mean", "module.layer2.1.bn3.running_var", "module.layer2.2.conv3.weight", "module.layer2.2.bn3.weight", "module.layer2.2.bn3.bias", "module.layer2.2.bn3.running_mean", "module.layer2.2.bn3.running_var", "module.layer2.3.conv3.weight", "module.layer2.3.bn3.weight", "module.layer2.3.bn3.bias", "module.layer2.3.bn3.running_mean", "module.layer2.3.bn3.running_var", "module.layer2.4.conv3.weight", "module.layer2.4.bn3.weight", "module.layer2.4.bn3.bias", "module.layer2.4.bn3.running_mean", "module.layer2.4.bn3.running_var", "module.layer2.5.conv3.weight", "module.layer2.5.bn3.weight", "module.layer2.5.bn3.bias", "module.layer2.5.bn3.running_mean", "module.layer2.5.bn3.running_var", "module.layer2.6.conv3.weight", "module.layer2.6.bn3.weight", "module.layer2.6.bn3.bias", "module.layer2.6.bn3.running_mean", "module.layer2.6.bn3.running_var", "module.layer2.7.conv3.weight", "module.layer2.7.bn3.weight", "module.layer2.7.bn3.bias", "module.layer2.7.bn3.running_mean", "module.layer2.7.bn3.running_var", "module.layer2.8.conv3.weight", "module.layer2.8.bn3.weight", "module.layer2.8.bn3.bias", "module.layer2.8.bn3.running_mean", "module.layer2.8.bn3.running_var", "module.layer2.9.conv3.weight", "module.layer2.9.bn3.weight", "module.layer2.9.bn3.bias", "module.layer2.9.bn3.running_mean", "module.layer2.9.bn3.running_var", "module.layer2.10.conv3.weight", "module.layer2.10.bn3.weight", "module.layer2.10.bn3.bias", "module.layer2.10.bn3.running_mean", "module.layer2.10.bn3.running_var", "module.layer2.11.conv3.weight", "module.layer2.11.bn3.weight", "module.layer2.11.bn3.bias", "module.layer2.11.bn3.running_mean", "module.layer2.11.bn3.running_var", "module.layer2.12.conv3.weight", "module.layer2.12.bn3.weight", "module.layer2.12.bn3.bias", "module.layer2.12.bn3.running_mean", "module.layer2.12.bn3.running_var", "module.layer2.13.conv3.weight", "module.layer2.13.bn3.weight", "module.layer2.13.bn3.bias", "module.layer2.13.bn3.running_mean", "module.layer2.13.bn3.running_var", "module.layer2.14.conv3.weight", "module.layer2.14.bn3.weight", "module.layer2.14.bn3.bias", "module.layer2.14.bn3.running_mean", "module.layer2.14.bn3.running_var", "module.layer2.15.conv3.weight", "module.layer2.15.bn3.weight", "module.layer2.15.bn3.bias", "module.layer2.15.bn3.running_mean", "module.layer2.15.bn3.running_var", "module.layer2.16.conv3.weight", "module.layer2.16.bn3.weight", "module.layer2.16.bn3.bias", "module.layer2.16.bn3.running_mean", "module.layer2.16.bn3.running_var", "module.layer2.17.conv3.weight", "module.layer2.17.bn3.weight", "module.layer2.17.bn3.bias", "module.layer2.17.bn3.running_mean", "module.layer2.17.bn3.running_var", "module.layer3.0.conv3.weight", "module.layer3.0.bn3.weight", "module.layer3.0.bn3.bias", "module.layer3.0.bn3.running_mean", "module.layer3.0.bn3.running_var", "module.layer3.0.downsample.1.weight", "module.layer3.0.downsample.1.bias", "module.layer3.0.downsample.1.running_mean", "module.layer3.0.downsample.1.running_var", "module.layer3.1.conv3.weight", "module.layer3.1.bn3.weight", "module.layer3.1.bn3.bias", "module.layer3.1.bn3.running_mean", "module.layer3.1.bn3.running_var", "module.layer3.2.conv3.weight", "module.layer3.2.bn3.weight", "module.layer3.2.bn3.bias", "module.layer3.2.bn3.running_mean", "module.layer3.2.bn3.running_var", "module.layer3.3.conv3.weight", "module.layer3.3.bn3.weight", "module.layer3.3.bn3.bias", "module.layer3.3.bn3.running_mean", "module.layer3.3.bn3.running_var", "module.layer3.4.conv3.weight", "module.layer3.4.bn3.weight", "module.layer3.4.bn3.bias", "module.layer3.4.bn3.running_mean", "module.layer3.4.bn3.running_var", "module.layer3.5.conv3.weight", "module.layer3.5.bn3.weight", "module.layer3.5.bn3.bias", "module.layer3.5.bn3.running_mean", "module.layer3.5.bn3.running_var", "module.layer3.6.conv3.weight", "module.layer3.6.bn3.weight", "module.layer3.6.bn3.bias", "module.layer3.6.bn3.running_mean", "module.layer3.6.bn3.running_var", "module.layer3.7.conv3.weight", "module.layer3.7.bn3.weight", "module.layer3.7.bn3.bias", "module.layer3.7.bn3.running_mean", "module.layer3.7.bn3.running_var", "module.layer3.8.conv3.weight", "module.layer3.8.bn3.weight", "module.layer3.8.bn3.bias", "module.layer3.8.bn3.running_mean", "module.layer3.8.bn3.running_var", "module.layer3.9.conv3.weight", "module.layer3.9.bn3.weight", "module.layer3.9.bn3.bias", "module.layer3.9.bn3.running_mean", "module.layer3.9.bn3.running_var", "module.layer3.10.conv3.weight", "module.layer3.10.bn3.weight", "module.layer3.10.bn3.bias", "module.layer3.10.bn3.running_mean", "module.layer3.10.bn3.running_var", "module.layer3.11.conv3.weight", "module.layer3.11.bn3.weight", "module.layer3.11.bn3.bias", "module.layer3.11.bn3.running_mean", "module.layer3.11.bn3.running_var", "module.layer3.12.conv3.weight", "module.layer3.12.bn3.weight", "module.layer3.12.bn3.bias", "module.layer3.12.bn3.running_mean", "module.layer3.12.bn3.running_var", "module.layer3.13.conv3.weight", "module.layer3.13.bn3.weight", "module.layer3.13.bn3.bias", "module.layer3.13.bn3.running_mean", "module.layer3.13.bn3.running_var", "module.layer3.14.conv3.weight", "module.layer3.14.bn3.weight", "module.layer3.14.bn3.bias", "module.layer3.14.bn3.running_mean", "module.layer3.14.bn3.running_var", "module.layer3.15.conv3.weight", "module.layer3.15.bn3.weight", "module.layer3.15.bn3.bias", "module.layer3.15.bn3.running_mean", "module.layer3.15.bn3.running_var", "module.layer3.16.conv3.weight", "module.layer3.16.bn3.weight", "module.layer3.16.bn3.bias", "module.layer3.16.bn3.running_mean", "module.layer3.16.bn3.running_var", "module.layer3.17.conv3.weight", "module.layer3.17.bn3.weight", "module.layer3.17.bn3.bias", "module.layer3.17.bn3.running_mean", "module.layer3.17.bn3.running_var".
size mismatch for module.layer1.0.conv1.weight: copying a param with shape torch.Size([16, 16, 1, 1]) from checkpoint, the shape in current model is torch.Size([16, 16, 3, 3]).
size mismatch for module.layer1.1.conv1.weight: copying a param with shape torch.Size([16, 64, 1, 1]) from checkpoint, the shape in current model is torch.Size([16, 16, 3, 3]).
size mismatch for module.layer1.2.conv1.weight: copying a param with shape torch.Size([16, 64, 1, 1]) from checkpoint, the shape in current model is torch.Size([16, 16, 3, 3]).
size mismatch for module.layer1.3.conv1.weight: copying a param with shape torch.Size([16, 64, 1, 1]) from checkpoint, the shape in current model is torch.Size([16, 16, 3, 3]).
size mismatch for module.layer1.4.conv1.weight: copying a param with shape torch.Size([16, 64, 1, 1]) from checkpoint, the shape in current model is torch.Size([16, 16, 3, 3]).
size mismatch for module.layer1.5.conv1.weight: copying a param with shape torch.Size([16, 64, 1, 1]) from checkpoint, the shape in current model is torch.Size([16, 16, 3, 3]).
size mismatch for module.layer1.6.conv1.weight: copying a param with shape torch.Size([16, 64, 1, 1]) from checkpoint, the shape in current model is torch.Size([16, 16, 3, 3]).
size mismatch for module.layer1.7.conv1.weight: copying a param with shape torch.Size([16, 64, 1, 1]) from checkpoint, the shape in current model is torch.Size([16, 16, 3, 3]).
size mismatch for module.layer1.8.conv1.weight: copying a param with shape torch.Size([16, 64, 1, 1]) from checkpoint, the shape in current model is torch.Size([16, 16, 3, 3]).
size mismatch for module.layer1.9.conv1.weight: copying a param with shape torch.Size([16, 64, 1, 1]) from checkpoint, the shape in current model is torch.Size([16, 16, 3, 3]).
size mismatch for module.layer1.10.conv1.weight: copying a param with shape torch.Size([16, 64, 1, 1]) from checkpoint, the shape in current model is torch.Size([16, 16, 3, 3]).
size mismatch for module.layer1.11.conv1.weight: copying a param with shape torch.Size([16, 64, 1, 1]) from checkpoint, the shape in current model is torch.Size([16, 16, 3, 3]).
size mismatch for module.layer1.12.conv1.weight: copying a param with shape torch.Size([16, 64, 1, 1]) from checkpoint, the shape in current model is torch.Size([16, 16, 3, 3]).
size mismatch for module.layer1.13.conv1.weight: copying a param with shape torch.Size([16, 64, 1, 1]) from checkpoint, the shape in current model is torch.Size([16, 16, 3, 3]).
size mismatch for module.layer1.14.conv1.weight: copying a param with shape torch.Size([16, 64, 1, 1]) from checkpoint, the shape in current model is torch.Size([16, 16, 3, 3]).
size mismatch for module.layer1.15.conv1.weight: copying a param with shape torch.Size([16, 64, 1, 1]) from checkpoint, the shape in current model is torch.Size([16, 16, 3, 3]).
size mismatch for module.layer1.16.conv1.weight: copying a param with shape torch.Size([16, 64, 1, 1]) from checkpoint, the shape in current model is torch.Size([16, 16, 3, 3]).
size mismatch for module.layer1.17.conv1.weight: copying a param with shape torch.Size([16, 64, 1, 1]) from checkpoint, the shape in current model is torch.Size([16, 16, 3, 3]).
size mismatch for module.layer2.0.bn1.weight: copying a param with shape torch.Size([32]) from checkpoint, the shape in current model is torch.Size([16]).
size mismatch for module.layer2.0.bn1.bias: copying a param with shape torch.Size([32]) from checkpoint, the shape in current model is torch.Size([16]).
size mismatch for module.layer2.0.bn1.running_mean: copying a param with shape torch.Size([32]) from checkpoint, the shape in current model is torch.Size([16]).
size mismatch for module.layer2.0.bn1.running_var: copying a param with shape torch.Size([32]) from checkpoint, the shape in current model is torch.Size([16]).
size mismatch for module.layer2.0.conv1.weight: copying a param with shape torch.Size([32, 64, 1, 1]) from checkpoint, the shape in current model is torch.Size([32, 16, 3, 3]).
size mismatch for module.layer2.0.downsample.0.weight: copying a param with shape torch.Size([128, 64, 1, 1]) from checkpoint, the shape in current model is torch.Size([32, 16, 1, 1]).
size mismatch for module.layer2.1.conv1.weight: copying a param with shape torch.Size([32, 128, 1, 1]) from checkpoint, the shape in current model is torch.Size([32, 32, 3, 3]).
size mismatch for module.layer2.2.conv1.weight: copying a param with shape torch.Size([32, 128, 1, 1]) from checkpoint, the shape in current model is torch.Size([32, 32, 3, 3]).
size mismatch for module.layer2.3.conv1.weight: copying a param with shape torch.Size([32, 128, 1, 1]) from checkpoint, the shape in current model is torch.Size([32, 32, 3, 3]).
size mismatch for module.layer2.4.conv1.weight: copying a param with shape torch.Size([32, 128, 1, 1]) from checkpoint, the shape in current model is torch.Size([32, 32, 3, 3]).
size mismatch for module.layer2.5.conv1.weight: copying a param with shape torch.Size([32, 128, 1, 1]) from checkpoint, the shape in current model is torch.Size([32, 32, 3, 3]).
size mismatch for module.layer2.6.conv1.weight: copying a param with shape torch.Size([32, 128, 1, 1]) from checkpoint, the shape in current model is torch.Size([32, 32, 3, 3]).
size mismatch for module.layer2.7.conv1.weight: copying a param with shape torch.Size([32, 128, 1, 1]) from checkpoint, the shape in current model is torch.Size([32, 32, 3, 3]).
size mismatch for module.layer2.8.conv1.weight: copying a param with shape torch.Size([32, 128, 1, 1]) from checkpoint, the shape in current model is torch.Size([32, 32, 3, 3]).
size mismatch for module.layer2.9.conv1.weight: copying a param with shape torch.Size([32, 128, 1, 1]) from checkpoint, the shape in current model is torch.Size([32, 32, 3, 3]).
size mismatch for module.layer2.10.conv1.weight: copying a param with shape torch.Size([32, 128, 1, 1]) from checkpoint, the shape in current model is torch.Size([32, 32, 3, 3]).
size mismatch for module.layer2.11.conv1.weight: copying a param with shape torch.Size([32, 128, 1, 1]) from checkpoint, the shape in current model is torch.Size([32, 32, 3, 3]).
size mismatch for module.layer2.12.conv1.weight: copying a param with shape torch.Size([32, 128, 1, 1]) from checkpoint, the shape in current model is torch.Size([32, 32, 3, 3]).
size mismatch for module.layer2.13.conv1.weight: copying a param with shape torch.Size([32, 128, 1, 1]) from checkpoint, the shape in current model is torch.Size([32, 32, 3, 3]).
size mismatch for module.layer2.14.conv1.weight: copying a param with shape torch.Size([32, 128, 1, 1]) from checkpoint, the shape in current model is torch.Size([32, 32, 3, 3]).
size mismatch for module.layer2.15.conv1.weight: copying a param with shape torch.Size([32, 128, 1, 1]) from checkpoint, the shape in current model is torch.Size([32, 32, 3, 3]).
size mismatch for module.layer2.16.conv1.weight: copying a param with shape torch.Size([32, 128, 1, 1]) from checkpoint, the shape in current model is torch.Size([32, 32, 3, 3]).
size mismatch for module.layer2.17.conv1.weight: copying a param with shape torch.Size([32, 128, 1, 1]) from checkpoint, the shape in current model is torch.Size([32, 32, 3, 3]).
size mismatch for module.layer3.0.bn1.weight: copying a param with shape torch.Size([64]) from checkpoint, the shape in current model is torch.Size([32]).
size mismatch for module.layer3.0.bn1.bias: copying a param with shape torch.Size([64]) from checkpoint, the shape in current model is torch.Size([32]).
size mismatch for module.layer3.0.bn1.running_mean: copying a param with shape torch.Size([64]) from checkpoint, the shape in current model is torch.Size([32]).
size mismatch for module.layer3.0.bn1.running_var: copying a param with shape torch.Size([64]) from checkpoint, the shape in current model is torch.Size([32]).
size mismatch for module.layer3.0.conv1.weight: copying a param with shape torch.Size([64, 128, 1, 1]) from checkpoint, the shape in current model is torch.Size([64, 32, 3, 3]).
size mismatch for module.layer3.0.downsample.0.weight: copying a param with shape torch.Size([256, 128, 1, 1]) from checkpoint, the shape in current model is torch.Size([64, 32, 1, 1]).
size mismatch for module.layer3.1.conv1.weight: copying a param with shape torch.Size([64, 256, 1, 1]) from checkpoint, the shape in current model is torch.Size([64, 64, 3, 3]).
size mismatch for module.layer3.2.conv1.weight: copying a param with shape torch.Size([64, 256, 1, 1]) from checkpoint, the shape in current model is torch.Size([64, 64, 3, 3]).
size mismatch for module.layer3.3.conv1.weight: copying a param with shape torch.Size([64, 256, 1, 1]) from checkpoint, the shape in current model is torch.Size([64, 64, 3, 3]).
size mismatch for module.layer3.4.conv1.weight: copying a param with shape torch.Size([64, 256, 1, 1]) from checkpoint, the shape in current model is torch.Size([64, 64, 3, 3]).
size mismatch for module.layer3.5.conv1.weight: copying a param with shape torch.Size([64, 256, 1, 1]) from checkpoint, the shape in current model is torch.Size([64, 64, 3, 3]).
size mismatch for module.layer3.6.conv1.weight: copying a param with shape torch.Size([64, 256, 1, 1]) from checkpoint, the shape in current model is torch.Size([64, 64, 3, 3]).
size mismatch for module.layer3.7.conv1.weight: copying a param with shape torch.Size([64, 256, 1, 1]) from checkpoint, the shape in current model is torch.Size([64, 64, 3, 3]).
size mismatch for module.layer3.8.conv1.weight: copying a param with shape torch.Size([64, 256, 1, 1]) from checkpoint, the shape in current model is torch.Size([64, 64, 3, 3]).
size mismatch for module.layer3.9.conv1.weight: copying a param with shape torch.Size([64, 256, 1, 1]) from checkpoint, the shape in current model is torch.Size([64, 64, 3, 3]).
size mismatch for module.layer3.10.conv1.weight: copying a param with shape torch.Size([64, 256, 1, 1]) from checkpoint, the shape in current model is torch.Size([64, 64, 3, 3]).
size mismatch for module.layer3.11.conv1.weight: copying a param with shape torch.Size([64, 256, 1, 1]) from checkpoint, the shape in current model is torch.Size([64, 64, 3, 3]).
size mismatch for module.layer3.12.conv1.weight: copying a param with shape torch.Size([64, 256, 1, 1]) from checkpoint, the shape in current model is torch.Size([64, 64, 3, 3]).
size mismatch for module.layer3.13.conv1.weight: copying a param with shape torch.Size([64, 256, 1, 1]) from checkpoint, the shape in current model is torch.Size([64, 64, 3, 3]).
size mismatch for module.layer3.14.conv1.weight: copying a param with shape torch.Size([64, 256, 1, 1]) from checkpoint, the shape in current model is torch.Size([64, 64, 3, 3]).
size mismatch for module.layer3.15.conv1.weight: copying a param with shape torch.Size([64, 256, 1, 1]) from checkpoint, the shape in current model is torch.Size([64, 64, 3, 3]).
size mismatch for module.layer3.16.conv1.weight: copying a param with shape torch.Size([64, 256, 1, 1]) from checkpoint, the shape in current model is torch.Size([64, 64, 3, 3]).
size mismatch for module.layer3.17.conv1.weight: copying a param with shape torch.Size([64, 256, 1, 1]) from checkpoint, the shape in current model is torch.Size([64, 64, 3, 3]).
size mismatch for module.fc.weight: copying a param with shape torch.Size([100, 256]) from checkpoint, the shape in current model is torch.Size([10, 64]).
size mismatch for module.fc.bias: copying a param with shape torch.Size([100]) from checkpoint, the shape in current model is torch.Size([10]).

did you solve it please ?

@Kin-Zhang
Copy link

Kin-Zhang commented Feb 11, 2022

        state_dict =checkpoint['state_dict']
        from collections import OrderedDict
        new_state_dict = OrderedDict()

        for k, v in state_dict.items():
            if 'module' not in k:
                k = 'module.'+k
            else:
                k = k.replace('features.module.', 'module.features.')
            new_state_dict[k]=v

        model.load_state_dict(new_state_dict)

This is the solution!!!! Thanks!!!!!

@bilalghanem
Copy link

change:
model.load_state_dict(torch.load(path + '/pytorch_model.pt'))
to
model.load_state_dict(torch.load(path + '/pytorch_model.pt'), strict=False)

@mgrachten
Copy link

change: model.load_state_dict(torch.load(path + '/pytorch_model.pt')) to model.load_state_dict(torch.load(path + '/pytorch_model.pt'), strict=False)

Although it will make the RuntimeError go away, don't do this unless you know what you are doing. It will leave any parameters it can't find in the checkpoint with random values. That's not what you want if the issue is caused by a mix-up of parameter names, as was the case for the issue reporter.

@Chenny0808
Copy link

Use model.module.state_dict() instead of model.state_dict() in DP mode

# for free to join this conversation on GitHub. Already have an account? # to comment
Labels
None yet
Projects
None yet
Development

No branches or pull requests