class_forgetting
[Deep Unlearning-PyTorch] Class Forgetting as in paper "Deep Unlearning: Fast and Efficient Training-free Approach to Controlled Forgetting"
Science Score: 13.0%
This score indicates how likely this project is to be science-related based on various indicators:
-
○CITATION.cff file
-
✓codemeta.json file
Found codemeta.json file -
○.zenodo.json file
-
○DOI references
-
○Academic publication links
-
○Academic email domains
-
○Institutional organization owner
-
○JOSS paper metadata
-
○Scientific vocabulary similarity
Low similarity (7.9%) to scientific vocabulary
Repository
[Deep Unlearning-PyTorch] Class Forgetting as in paper "Deep Unlearning: Fast and Efficient Training-free Approach to Controlled Forgetting"
Basic Info
- Host: GitHub
- Owner: sangamesh-kodge
- Language: Python
- Default Branch: master
- Homepage: https://arxiv.org/abs/2312.00761
- Size: 68.2 MB
Statistics
- Stars: 11
- Watchers: 2
- Forks: 0
- Open Issues: 0
- Releases: 0
Metadata Files
Readme.md
Introduction
This is official repository for the paper Deep Unlearning: Fast and Efficient Gradient-free Approach to Class Forgetting accepted at TMLR.
Setup Environment using yml
bash
conda env create -f env.yml
conda activate forget
Demo of unlearning algorithm
bash
python3 ./demo.py
Unlearning Single class on CIFAR10, CIFAR100 and ImageNet.
```bash
for CIFAR10
sh ./scripts/our_cifar10.sh ```
```bash
for CIFAR100
sh ./scripts/our_cifar100.sh ```
```bash
for ImageNet
sh ./scripts/our_imagenet.sh ```
Analysis
Scripts for analysis done in the paper can be found in scripts/analysis.
An older version of the repository can be found in the legacy branch of the repository.
Citation
Kindly cite the paper if you use the code. Thanks!
APA
Kodge, Sangamesh, Gobinda Saha, and Kaushik Roy. "Deep Unlearning: Fast and Efficient Gradient-Free Class Forgetting."
Bibtex
@article{
kodge2024deep,
title={Deep Unlearning: Fast and Efficient Gradient-free Class Forgetting},
author={Sangamesh Kodge and Gobinda Saha and Kaushik Roy},
journal={Transactions on Machine Learning Research},
issn={2835-8856},
year={2024},
url={https://openreview.net/forum?id=BmI5p6wBi0},
note={}
}
Owner
- Login: sangamesh-kodge
- Kind: user
- Repositories: 1
- Profile: https://github.com/sangamesh-kodge
GitHub Events
Total
- Watch event: 3
- Fork event: 1
Last Year
- Watch event: 3
- Fork event: 1