Closed crazyn2 closed 8 months ago
Hey @crazyn2 thanks for the post. We do have tests against Sklearn that are passing and have passed a bunch of times since auroc landed. Can you give a specific set of inputs that yields a different output?
Thank you for your prompt response. I've provide my code repository of github https://github.com/crazyn2/mini-ad which contained the whole procedures including dataset inputs ? If you have time, you can reproduce the result utilizing those codes.
I've identified the reason for the discrepancy in AUC-ROC values with sklearn. I realized that I forgot to reset the metric class in my loop. Apologies for any inconvenience.
π Describe the bug
Torcheval gives a different answer from sklearn for roc auc
here is my code repository https://github.com/crazyn2/mini-ad please set download=True in datamodules/cifar10.py and run this command
Versions