howard-hou / UniqueClassCount

Pytorch implementation for ICLR 2020 paper "Weakly Supervised Clustering by Exploiting Unique Class Count"

Unique Class Count

Description

Pytorch implementation for ICLR 2020 paper "Weakly Supervised Clustering by Exploiting Unique Class Count"

Installation

# clone project
git clone https://github.com/howard-hou/UniqueClassCount
cd UniqueClassCount

# [OPTIONAL] create conda environment
conda create -n myenv python=3.9
conda activate myenv

# install pytorch according to instructions
# https://pytorch.org/get-started/

# install requirements
pip install -r requirements.txt

How to run

Download dataset first

mnist dataset is small, so it has been uploaded to github with the project.

for camelyon dataset, download the pre-processed dataset first, and put the dataset in data dir.

To see reproducing result

run notebook in notebooks dir

Train

Train mnist model with default configuration

# train
python ucc/train.py

Train camelyon model with chosen experiment configuration from configs/experiment/

python ucc/train.py experiment=camelyon.yaml