Skip to content
Learning a Unified Classifier Incrementally via Rebalancing
Python Shell
Branch: master
Clone or download
Fetching latest commit…
Cannot retrieve the latest commit at this time.
Permalink
Type Name Latest commit message Commit time
Failed to load latest commit information.
cifar100-class-incremental update run.sh Jun 11, 2019
imagenet-class-incremental minor Jun 14, 2019
utils_incremental init Jun 11, 2019
.gitignore init Jun 11, 2019
README.md bib Jun 11, 2019

README.md

CVPR19 Incremental Learning

Thi repository is for the paper "Learning a Unified Classifier Incrementally via Rebalancing".

[Paper] [Project Page]

Instructions

  1. Dependencies
    • Python 3.6 (Anaconda3 Recommended)
    • Pytorch 0.4.0
  2. Getting Started
    • the data for CIFAR100 and ImageNet are put in cifar100-class-incremental/data and imagenet-class-incremental/data, or you can make soft links to the directories which include the corresponding data
    • see cifar100-class-incremental/run.sh for the experiments on CIFAR100
    • see imagenet-class-incremental/run.sh for the experiments on ImageNet-Subset
    • see imagenet-class-incremental/run_all.sh for the experiments on ImageNet-Full

Citation

Please cite the following paper if you find this useful in your research:

@InProceedings{Hou_2019_CVPR,
author = {Hou, Saihui and Pan, Xinyu and Loy, Chen Change and Wang, Zilei and Lin, Dahua},
title = {Learning a Unified Classifier Incrementally via Rebalancing},
booktitle = {The IEEE Conference on Computer Vision and Pattern Recognition (CVPR)},
month = {June},
year = {2019}
}
You can’t perform that action at this time.