Torch.save output is a zip archieve

I am extracting features from each patches in .pt files and using these files for training in the downstream of my project. At one point I concatenate the feature files using a code similar to following:

features1 = torch.load('File1.pt')
features2 = torch.load('File2.pt')
merged_features = torch.cat([features1, features2], dim=0)
torch.save(merged_features, 'merged_feature.pt')

However, I receive following error for some files when I try to load:

RuntimeError("merged_feature.pt is a zip archive (did you mean to use torch.jit.load()?)".format(f.name))

Previously, I run the feature extraction and merging codes again and magically the issue was solved. However, it became persistent and for all the files. I do not know how to solve. From the following location the working and not working files can be seen:
https://drive.google.com/drive/folders/1V2JLAfF9o2WjLUxwQPwYvxKfCr_bFuu1?usp=sharing

I would be glad if anyone can help to solve this issue.

pytorch: 1.3.1
Red Hat 7.9