Skip to content

Offical Code for Paper "Exploring Inter-Channel Correlation for Diversity-preserved Knowledge Distillation"

Notifications You must be signed in to change notification settings

ADLab-AutoDrive/ICKD

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

4 Commits
 
 
 
 
 
 
 
 
 
 

Repository files navigation

ICKD

This repository provides code of our paper "Exploring Inter-Channel Correlation for Diversity-preserved Knowledge Distillation". We provide training code on Cifar100 and evaluation code on ImageNet & Pascal VOC. The remaining training code will be released after the paper is accepted.

Preparation

  1. Download all checkpoints from https://drive.google.com/drive/folders/1ZvwEAVJurTXSuPL_0HylHMNUGbG1zl3U?usp=sharing
  2. Build a docker image using the provided Dockerfile. All code should be run in the docker image.

Running

Go into directory of each task and following the README there.

License

The License will be updated after the code is released without anonymity.

Acknowledgement

This work is built on three different repository, RepDistiller(ICLR 2020), torchdistill(ICPR 2020) and OverHaul(ICCV 2019). Thanks to their great work.

About

Offical Code for Paper "Exploring Inter-Channel Correlation for Diversity-preserved Knowledge Distillation"

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published