tufts-ml / InterLUDE

Code for ICML 2024 paper "InterLUDE: Interactions between Labeled and Unlabeled Data to Enhance Semi-supervised Learning"

Repository from Github https://github.comtufts-ml/InterLUDERepository from Github https://github.comtufts-ml/InterLUDE

InterLUDE

Code for ICML 2024 paper "InterLUDE: Interactions between Labeled and Unlabeled Data to Enhance Semi-supervised Learning"

Setup

Install Anaconda

Follow the instructions here: https://conda.io/projects/conda/en/latest/user-guide/install/index.html

Environment

pytorch 1.11.0

Running experiments

To run the code, you can use the bash file and pass in the hyperparameter settings. (Example bash files are provided in runs/ folder, you will need to modify the paths to point to the data/the folder you want to store the output/the path to your script etc. The examples are based on CNN experiments, for ViT experiments you can adapt them correspondingly.)

Hyperparameters

please refer to the paper for hyperparameter details.

About

Code for ICML 2024 paper "InterLUDE: Interactions between Labeled and Unlabeled Data to Enhance Semi-supervised Learning"

License:MIT License


Languages

Language:Python 86.1%Language:Shell 13.9%