This is the official implementation of “Class Imbalanced Deep Learning via a Class-Balanced Ensemble”, our paper has been accepted by IEEE Transactions on Neural Networks and Learning Systems. If you use the codes in this repo, please cite the paper as follows:
[1] Zhi Chen, Jiang Duan, Li Kang, and G. Qiu, "Class Imbalanced Deep Learning via a Class-Balanced Ensemble," IEEE Transactions on Neural Networks and Learning Systems, p. 15, 2021.
Pytorch 1.2 or 1.3
Python 3.7
(1) You can call the function MyEnsemble() in MyEnsemble.py to train the deep Ensemble proposed in our paper.
(2) EnsembleV2.py is the deep Ensemble proposed in our paper.
(3) CifarImbalanced.py is the code used to create imbalanced CIFAR dataset.
(4) ForcalLossV5.py is the Loss Function proposed in our paper.