Skip to content

Latest commit

 

History

2 Commits

Folders and files

NameName
Last commit message
Last commit date
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

Unsupervised Image Classification PDF

Weijie Chen, Shiliang Pu, Di Xie, Shicai Yang, Yilu Guo, Luojun Lin. In ECCVW 2020.

Prerequisites

  • python3.6
  • pytorch1.1

UIC Pipeline

TLDR: UIC is a very simple self-supervised learning framework for joint image classification and representation learning. It utilizes the forward result at epoch t-1 as pseudo label to drive unsupervised training at epoch t.

Getting Started

Data Preparation

Link your own ImageNet dataset to ./data/ImageNet/train and ./data/ImageNet/val with the format of datasets.ImageFolder in Pytorch.

Unsupervised Training

$ sh ./main.sh

Linear Evaluation

$ sh ./eval_linear.sh

License

You can view in License.

Citation

If you find our code useful, please consider citing our paper:

@InProceedings{chen2020unsupervised,
  title={Unsupervised Image Classification for Deep Representation Learning},
  author={Chen, Weijie and Pu, Shiliang and Xie, Di and Yang, Shicai and Guo, Yilu and Lin, Luojun},
  booktitle={European Conference on Computer Vision Workshop},
  pages={430-446}
  year={2020},
  organization={Springer}
}

Acknowledgements

Our code is implemented based on DeepCluster.

Contact

If you are interested in internship, or applied researcher / developer positions in Hikvision Research Institute, please feel free to seed an email to chenweijie5 -at- hikvision.com.

About

A very simple self-supervised image classification framework!

Topics

Resources

Stars

52 stars

Watchers

1 watching

Forks

Releases

Packages

Used by

Contributors

Languages