I can't modify patch_size

[QUSTwh]

Thank you a lot for your contribution to the code, however, when I try to adjust the patch size of training data, it sometimes display an error：

File "main.py", line 311, in
display=viz)
File "/home/WH/Hyperspectral-Classification-master/models.py", line 1045, in train
val_acc = val(net, val_loader, device=device, supervision=supervision)
File "/home/WH/Hyperspectral-Classification-master/models.py", line 1142, in val
return accuracy / total
ZeroDivisionError: float division by zero

How can I fix it？

[eecn]

I think this error should be independent of the size of path_size, maybe you can check other codes.