Questions about the labels of Gallant dataset

I found that the labels provided for this dataset are more accurate than the output of pre-trained ResNet152 in PyTorch.

I am interested in the neuronal activation of the fine-tuned network, so I was wondering whether the implementation code and fine-tuned pth model file can be provided. Thanks in advance!

Sorry for the delay, yes I can provide these, but I have them as mxnet weights. Is that still useful? There are some tools for converting weights from one model to another, but I haven’t really used them before.

Yes, the code and saved weights based on mxnet are also useful. Thank you!