Kaushalya/gnn-meta-attack-pytorch

Issues

• train accuracy keeps decreasing while train-loss also decrease?

  #4 opened 5 years ago by zhaohan-xi
  2

• Implement first order approximation for meta gradient

  #2 opened 5 years ago by Kaushalya
  1

• Inner loop optimization is inconsistent.

  #1 opened 5 years ago by Kaushalya
  2

• Update learning rate based on a momentum parameter.

  #3 opened 5 years ago by Kaushalya
  0