alinlab / LfF

Learning from Failure: Training Debiased Classifier from Biased Classifier (NeurIPS 2020)
89 stars 9 forks source link

Unable to replicate numbers on CelebA #2

Open erobic opened 3 years ago

erobic commented 3 years ago

Hi,

After trying to run the code for CelebA (with ResNet18), I get the following results:

  1. valid_attrwise_accs_b tensor([[0.9578, 0.9966], [0.8382, 0.2747]])
  2. valid_attrwise_accs_d tensor([[0.0300, 0.0205], [0.6841, 0.9231]])

Could you please specify the exact way to replicate the results in the paper?

Thank you!

dfg-collab commented 11 months ago

Hello, I would like to know how your celeba dataset was downloaded. I don't seem to have seen the code for loading the celeba dataset