CHNxindong / PML

[CVPR 2021] This repository is the official implementation of paper: "PML: Progressive Margin Loss for Long-tailed Age Classification"

PML: Progressive Margin Loss for Long-tailed Age Classification

This repository is the official implementation of paper: "PML: Progressive Margin Loss for Long-tailed Age Classification"(CVPR 2021) [Paper(CVF)] [Paper(arXiv)]

Datasets

[Morph II] | [FG-NET] | [ChaLearn LAP 2015] | [IMDB-WIKI]

Training

We use (SNMC) Single Node Multi-GPU Cards training (with DistributedDataParallel) to get better performance.

 python -m torch.distributed.launch --nproc_per_node=2  --master_port 29502 ./train.py  --config ./configs/chalearn/exp_margin.yml

Testing

We test while training to save the best model.

Experiments

Citation

If you found this code or our work useful, please cite our paper.

@InProceedings{Deng_2021_CVPR,
    author    = {Deng, Zongyong and Liu, Hao and Wang, Yaoxing and Wang, Chenyang and Yu, Zekuan and Sun, Xuehong},
    title     = {PML: Progressive Margin Loss for Long-Tailed Age Classification},
    booktitle = {Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR)},
    month     = {June},
    year      = {2021},
    pages     = {10503-10512}
}