Error(s) in loading state_dict for CRNN:

[Bilgee]

I downloaded and tested the pretrained densenet model, but it showed the error messages.

Error(s) in loading state_dict for CRNN:
Unexpected key(s) in state_dict: "features.1.denselayer1.norm1.num_batches_tracked", ...

[zhiqwang]

I guess you load the pretrained densenet121 model with the scripts python ./main.py --arch densenet121 --resume densenet121_pretrained.pth --test-only?

The pretrained model just contain the model.state_dict(), which is not the resume checkpoint of the script main.py (It also contain optimizer.state_dict()).

One way to use the pretrained model is demo.ipynb, or modify main.py to load the model with parameter model.load_state_dict(torch.load(PATH), strict=False).