FedLC
FedLC copied to clipboard
the dice value is nan
run python scripts/train_multi_lcfed.py ,but the bug is up.how to solve it?
thank you!
the dataset is already slice into 6 sites.
Maybe the preprocessing stage has bugs. You could check the generated data and the dataloader to guarantee that the data format is correct.
i use the pmr dataset and i use your prepare_dataset.py to handle the data.i think the preprocess is ok
i change to the fundus dataset,and it happens again! how could i solve it?
I also met the same issue, have you solved it?