You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
I assume I have all the correct dependencies since I had errors before and resolved them by installing cuda 9.0 with gcc-5 and g++-5, but now I can not seem to solve this issue.
When I run the evaluate.py script I get the following error:
Traceback (most recent call last):
File "evaluate.py", line 234, in <module>
main()
File "evaluate.py", line 37, in main
model = get_model(num_classes=args.num_classes)
AttributeError: 'Namespace' object has no attribute 'num_classes'
If I change the num_classes to anything but 7 I get the following err (this example is without putting in a parameter):
Traceback (most recent call last):
File "evaluate.py", line 234, in <module>
main()
File "evaluate.py", line 41, in main
model.load_state_dict(saved_state_dict, strict=False)
File "/home/lukak/anaconda3/envs/CompositionalHumanParsing/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 Stream:
size mismatch for decoder.layer_dsn.3.weight: copying a param of torch.Size([20, 512, 1, 1]) from checkpoint, where the shape is torch.Size([7, 512, 1, 1]) in current model.
size mismatch for decoder.layer_dsn.3.bias: copying a param of torch.Size([20]) from checkpoint, where the shape is torch.Size([7]) in current model.
When I change the num_classes to 7 I get the following err with long list of missing keys and unexpected keys:
Traceback (most recent call last):
File "evaluate.py", line 234, in <module>
main()
File "evaluate.py", line 41, in main
model.load_state_dict(saved_state_dict)
File "/home/lukak/anaconda3/envs/CompositionalHumanParsing/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 Stream:
Missing key(s) in state_dict: "decoder.layer5.b0.0.weight", "decoder.layer5.b0.1.weight", "decoder.layer5.b0.1.bias", ...
.
.
.
Unexpected key(s) in state_dict: "decoder.layer6.alpha", "decoder.layer6.conv0.0.weight", "decoder.layer6.conv0.1.weight", ...
.
.
.
I have searched the internet for answers but found nothing that helped and as I am not an experienced software engineer any help would be appreciated.
The text was updated successfully, but these errors were encountered:
ge97qir
changed the title
AttributeError: 'Namespace' object has no attribute 'num_classes' / Missing and Unexpected keys
AttributeError: 'Namespace' object has no attribute 'num_classes' / Missing and Unexpected keys in state_dict
Jan 14, 2021
I also don't understand why there are 7 nodes in the final layer of the model, since there are 20 classes to identify.
Is the model for download correct?
I assume I have all the correct dependencies since I had errors before and resolved them by installing cuda 9.0 with gcc-5 and g++-5, but now I can not seem to solve this issue.
When I run the evaluate.py script I get the following error:
If I change the num_classes to anything but 7 I get the following err (this example is without putting in a parameter):
When I change the num_classes to 7 I get the following err with long list of missing keys and unexpected keys:
I have searched the internet for answers but found nothing that helped and as I am not an experienced software engineer any help would be appreciated.
The text was updated successfully, but these errors were encountered: