/Adam-NSCL

PyTorch implementation of our Adam-NSCL algorithm from our CVPR2021 (oral) paper "Training Networks in Null Space for Continual Learning"

Primary LanguagePythonMIT LicenseMIT

Adam-NSCL

This is a PyTorch implementation of Adam-NSCL algorithm for continual learning from our CVPR2021 (oral) paper:

Title: Training Networks in Null Space of Feature Covariance for Continual Learning

Authors: Shipeng Wang, Xiaorong Li, Jian Sun, Zongben Xu

Email: wangshipeng8128@stu.xjtu.edu.cn; wangshipeng8128@gmail.com

Arxiv: https://arxiv.org/pdf/2103.07113

Usage

sh scripts_svd/adamnscl.sh

Requirements: Python 3.7, PyTorch=1.5,tensorboardX

Citation

@InProceedings{Wang_2021_CVPR,
    author    = {Wang, Shipeng and Li, Xiaorong and Sun, Jian and Xu, Zongben},
    title     = {Training Networks in Null Space of Feature Covariance for Continual Learning},
    booktitle = {Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR)},
    month     = {June},
    year      = {2021},
    pages     = {184-193}
}