This folder contains a simple Wide-ResNet implementation that can be trained on Cifar10/SVHN with SAM. Start the training by running
python3 train-(cifar/svhn).py
SVHN data loader is taken from: https://github.com/uoguelph-mlrg/Cutout
Files
This branch is 20 commits ahead of, 18 commits behind davda54/sam:main.
example
Folders and files
| Name | Name | Last commit date | ||
|---|---|---|---|---|
parent directory.. | ||||