Code for the paper presented at ECCV2018
Contents:
./logs
best logs for the tables reported in the paper
./scripts
sh scripts to reproduce experiments
./modules
contains the attention module proposed in the paper.
./models
contains the different networks used on cifar.
Usage:
- git clone this repository && cd this repository
- ./scripts/your_script.sh (edit to set dataset paths, etc)
Requirements:
- The code requires pytorch 0.4