codes-kzhan / GHM_Detection

The implementation of “Gradient Harmonized Single-stage Detector” published on AAAI 2019.

GHM_Detection

The implementation of Gradient Harmonized Single-stage Detector published on AAAI 2019 (Oral).

Loss Functions

• The GHM-C and GHM-R loss functions are available in ghm_loss.py.
• The code works for pytorch 0.4.1 and later version. If you want to run it with pytorch 0.3.x, please checkout to the pytorch-0.3 branch.

Training Code

• The main training code is based on mmdetection. Please see this for installation issues (note that you do not need to clone the mmdetection repo again).
• We provide training and testing scripts and configuration files for both GHM and baseline (focal loss and smooth L1 loss) in the experiments directory. You need specify the path of your own pre-trained model in the config files.

Result

Training using the Res50-FPN backbone and testing on COCO minival.

Method	AP
FL + SL1	35.6%
GHM-C + SL1	35.8%
GHM-C + GHM-R	37.0%

License and Citation

The use of this code is RESTRICTED to non-commercial research and educational purposes.

@article{li2019ghm,
  title={Gradient Harmonized Single-stage Detector},
  author={Buyu Li, Yu Liu, Xiaogang Wang},
  booktitle={AAAI},
  year={2019}
}