Learning Semi-supervised Gaussian Mixture Models for Generalized Category Discovery

This repo contains code for our paper: Learning Semi-supervised Gaussian Mixture Models for Generalized Category Discovery

Contents

:running: 1. Running

:clipboard: 2. Citation

:running: Running

Dependencies

pip install -r requirements.txt

Config

Set paths to datasets, pre-trained models and desired log directories in config.py. Also set the experiment paths in bash_scripts/run.sh.

Datasets

We use fine-grained benchmarks in this paper, including:

• The Semantic Shift Benchmark (SSB) and Herbarium19

We also use generic object recognition datasets, including:

• CIFAR-10/100 and ImageNet

Please follow this repo or this repo to set up the data.

Scripts

Train representation:

bash bash_scripts/run.sh

:clipboard: Citation

If you use this code in your research, please consider citing our paper:

@InProceedings{Zhao_2023_ICCV,
    author    = {Zhao, Bingchen and Wen, Xin and Han, Kai},
    title     = {Learning Semi-supervised Gaussian Mixture Models for Generalized Category Discovery},
    booktitle = {Proceedings of the IEEE/CVF International Conference on Computer Vision (ICCV)},
    month     = {October},
    year      = {2023},
    pages     = {16623-16633}
}

Acknowledgements

The codebase is largely built on this repo: https://github.com/sgvaze/generalized-category-discovery.