Unoffical code for paper Maintaining Discrimination and Fairness in Class Incremental Learning(CVPR 2019).
This code is my reproduction of this paper.Thanks to the code base FACIL.
I am just a green hand in incremental learning.This code is probably wrong. If you are checking my code, please forgive me for wasting your time and point out my mistakes.
Dataset
- Cifar-100
!!! you need to change _BASE_DATA_PATH (Variables under folder: datasets/dataset_config.py) to your dataset path.
- Python 3.7.10
- torch 1.11.0
- torchvision 0.4.2
- CUDA 11.6
python main.py
experience | 0 | 1 | 2 | 3 | 4 | 5 | 6 | 7 | 8 | 9 | avg acc |
---|---|---|---|---|---|---|---|---|---|---|---|
acc | 0.828 | 0.8165 | 0.777 | 0.7135 | 0.6754 | 0.6266 | 0.6021 | 0.5706 | 0.5372 | 0.5218 | 0.66687 |
Code base:FACIL.
I just simplified the framework and implemented the code on this basis.