All metric (train/test acc/auc) yields 0.5 when training from scratch with Logistic Regression - Githubissues

MinChen00 / UnlearningLeaks

Official implementation of "When Machine Unlearning Jeopardizes Privacy" (ACM CCS 2021)

GNU General Public License v3.0

44 stars 5 forks source link

All metric (train/test acc/auc) yields 0.5 when training from scratch with Logistic Regression #9

Open fengxijia opened 2 months ago

fengxijia commented 2 months ago

Hi, I tried

python main.py --exp mem_inf --unlearning_method scratch --dataset_name xxx --original_model LR

and all the parameters remained the default. The final results (train & test, accuracy & auc) always show 0.5 no matter what categorical dataset I use, I can't figure out why. Can you tell me why this might be happening? Thank you very much.