facebookresearch / mixup-cifar10

mixup: Beyond Empirical Risk Minimization
Other
1.16k stars 225 forks source link

Unable to Reproduce Results for WRN-28-10 #26

Open humza909 opened 1 year ago

humza909 commented 1 year ago

Hi,

I am unable to reproduce results for Cifar-10 dataset with WRN-28-10. Accuracy in paper is 97.3% whereas for me it stay below 97% around 96.9%.

Any suggestions?

Regards,