/fn_playground

WRN 40-4 training from scratch. Best test accuracy on Fashion MNIST dataset is ~96.74%; best test accuracy on Cifar-10 dataset is ~98.03%.

Primary LanguagePython

No issues in this repository yet.