beierzhu/xerm

[AAAI 2022 Oral] This is a Pytorch implementation of the AAAI 2022 paper "Cross-Domain Empirical Risk Minimization for Unbiased Long-tailed Classification"

xerm's Issues

xERM-PC performance

Dear Beier,
I got a problem that I could not get your accuracy of 53.2 with xERM-PC code. My accuracy result is 0.527 with your config ce.yaml. The accuracies are :
- Imbalanced Acc 0.461 Balanced Acc 0.513 xERM Acc 0.527
Could you give some hints about how to reproduce your result?

Best

xERM-TDE ImageNet Performance

Dear Beier,
Thank you for your innovative work and I am really interested in it. However, when training the ImageNet_LT dataset with xERM-TDE method, I got low performance only 14.4 in the overall accuracy. I would like to ask, is it convenient to share the ImageNet_LT training code or configs there to make sure we could get your sota results?

how to get a balanced model

Dear Beier:
I'm very interested in your work！In your code，I can get a imbalanced model by TDE, but I don't how do you get the balanced model ?