FileNotFoundError: [Errno 2] No such file or directory: 'outputs/models/non_adapted_cnn.pt'

[tchinchuthomas]

Hi,

I tried to reproduce the results according to the steps mentioned and I am getting this error. Could you please suggest if there is anything missing?

Thank you.

[dr-costas]

Hi,

As the error says, the code cannot find the file non_adapted_cnn.pt.

I now see that the file is non_adapted_cnn.pytorch. That is, the file name is non_adapted_cnn, but the ending is pt and not pytorch.

Please confirm that you have the file called non_adapted_cnn at the directory outputs/models. Then, check the extension of the file. If the extension is pytorch, please change it to pt.

I will fix that in the code in the following days.

Let me know if it worked for you.

[tchinchuthomas]

Hi,
Thank you for the quick response.

Actually I tried with label_classifier.pt and non_adapted_cnn.pt at the directory outputs/models.
Then I got the following error:
RuntimeError: Error(s) in loading state_dict for LabelClassifierAUDASC:
Missing key(s) in state_dict: "classifier.0.weight", "classifier.0.bias", "classifier.3.weight", "classifier.3.bias", "classifier.6.weight", "classifier.6.bias".
Unexpected key(s) in state_dict: "linear_layers.0.weight", "linear_layers.0.bias", "linear_layers.1.weight", "linear_layers.1.bias", "output_layer.weight", "output_layer.bias".

Could you please let me know the issue?

Thank you.

[dr-costas]

Hi,

I see! Thank you for reporting that. I will fix it and do a commit.

The problem that you see is caused because I changed the code from the original one that provided the classifier (from the repository of our previous paper). Now PyTorch cannot find the same names for class members, and this is the error message.

Thank you for reporting it! I thought that I had fix it before publishing this repo... it seems that I was wrong :)

Sorry!

[tchinchuthomas]

Hi,

Thank you for the response.