# FGD

**Repository Path**: bear_happy/FGD

## Basic Information

- **Project Name**: FGD
- **Description**: 此代码实现对目标检测算法的蒸馏代码
- **Primary Language**: Python
- **License**: Apache-2.0
- **Default Branch**: master
- **Homepage**: None
- **GVP Project**: No

## Statistics

- **Stars**: 0
- **Forks**: 0
- **Created**: 2022-07-19
- **Last Updated**: 2022-07-20

## Categories & Tags

**Categories**: Uncategorized

**Tags**: None

## README

# FGD
CVPR 2022 Paper: [Focal and Global Knowledge Distillation for Detectors](https://arxiv.org/abs/2111.11837)
## Install MMDetection and MS COCO2017
  - Our codes are based on [MMDetection](https://github.com/open-mmlab/mmdetection). Please follow the installation of MMDetection and make sure you can run it successfully.
  - This repo uses mmdet==2.11.0 and mmcv-full==1.2.4
  - If you want to use higher mmdet version, you may have to change the optimizer in apis/train.py and build_detector in tools/train.py.
  - For mmdet>=2.12.0, if you want to use inheriting strategy, you have to initalize the student with teacher's parameters after model.init_weights().
## Higher mmdet and mmcv-full version
  - You can refer [MGD](https://github.com/yzd-v/MGD) to change model.init_weights() in [train.py](https://github.com/yzd-v/FGD/tree/master/tools/train.py) and self.student.init_weights() in [distiller.py](https://github.com/yzd-v/FGD/tree/master/mmdet/distillation/distillers/detection_distiller.py).
## Add and Replace the codes
  - Add the configs/. in our codes to the configs/ in mmdetectin's codes.
  - Add the mmdet/distillation/. in our codes to the mmdet/ in mmdetectin's codes.
  - Replace the mmdet/apis/train.py and tools/train.py in mmdetection's codes with mmdet/apis/train.py and tools/train.py in our codes.
  - Add pth_transfer.py to mmdetection's codes.
  - Unzip COCO dataset into data/coco/
## Train

```
#single GPU
python tools/train.py configs/distillers/fgd/fgd_retina_rx101_64x4d_distill_retina_r50_fpn_2x_coco.py

#multi GPU
bash tools/dist_train.sh configs/distillers/fgd/fgd_retina_rx101_64x4d_distill_retina_r50_fpn_2x_coco.py 8
```

## Transfer
```
# Tansfer the FGD model into mmdet model
python pth_transfer.py --fgd_path $fgd_ckpt --output_path $new_mmdet_ckpt
```
## Test

```
#single GPU
python tools/test.py configs/retinanet/retinanet_r50_fpn_2x_coco.py $new_mmdet_ckpt --eval bbox

#multi GPU
bash tools/dist_test.sh configs/retinanet/retinanet_r50_fpn_2x_coco.py $new_mmdet_ckpt 8 --eval bbox
```
## Results

|    Model    |  Backbone  | Baseline(mAP) | +FGD(mAP) |                            config                            |                          weight                          | code |
| :---------: | :--------: | :-----------: | :-------: | :----------------------------------------------------------: | :------------------------------------------------------: | :--: |
|  RetinaNet  | ResNet-50  |     37.4      |   40.7    | [config](https://github.com/open-mmlab/mmdetection/tree/master/configs/retinanet/retinanet_r50_fpn_2x_coco.py) | [baidu](https://pan.baidu.com/s/1TwF9W13eHg6Sxkrr-4VTqg) | wsfw |
|  RetinaNet  | ResNet-101 |     38.9      |   41.7    | [config](https://github.com/open-mmlab/mmdetection/tree/master/configs/retinanet/retinanet_r101_fpn_2x_coco.py) |                                                          |      |
| Faster RCNN | ResNet-50  |     38.4      |   42.0    | [config](https://github.com/open-mmlab/mmdetection/tree/master/configs/faster_rcnn/faster_rcnn_r50_fpn_2x_coco.py) | [baidu](https://pan.baidu.com/s/14WjoMqxILoPaKfY5QsCK8w) | dgpf |
| Faster RCNN | ResNet-101 |     39.8      |   44.1    | [config](https://github.com/open-mmlab/mmdetection/tree/master/configs/faster_rcnn/faster_rcnn_r101_fpn_2x_coco.py) |                                                          |      |
|  RepPoints  | ResNet-50  |     38.6      |   42.0    | [config](https://github.com/open-mmlab/mmdetection/tree/master/configs/reppoints/reppoints_moment_r50_fpn_gn-neck+head_2x_coco.py) | [baidu](https://pan.baidu.com/s/1EJo9uQuZhimm7HI92TNThQ) | qx5d |
|  RepPoints  | ResNet-101 |     40.5      |   43.8    | [config](https://github.com/open-mmlab/mmdetection/tree/master/configs/reppoints/reppoints_moment_r101_fpn_gn-neck+head_2x_coco.py) |                                                          |      |
|    FCOS     | ResNet-50  |     38.5      |   42.7    | [config](https://github.com/yzd-v/FGD/blob/master/configs/fcos/fcos_center-normbbox-giou_r50_caffe_fpn_gn-head_mstrain_1x_coco.py) | [baidu](https://pan.baidu.com/s/16uCTa81ZzG7EoizdfnXhzQ) | sedt |
|  MaskRCNN   | ResNet-50  |     39.2      |   42.1    | [config](https://github.com/open-mmlab/mmdetection/tree/master/configs/mask_rcnn/mask_rcnn_r50_fpn_2x_coco.py) | [baidu](https://pan.baidu.com/s/101eOFcD8JDwqrFuYcxcBIA) | sv8m |
|     GFL     | ResNet-50  |     40.2      |   43.5    | [config](https://github.com/open-mmlab/mmdetection/blob/master/configs/gfl/gfl_r50_fpn_1x_coco.py) |                                                          |      |

|  Model   | Backbone  | Baseline(Mask mAP) | +FGD(Mask mAP) |                            config                            |                          weight                          | code |
| :------: | :-------: | :----------------: | :------------: | :----------------------------------------------------------: | :------------------------------------------------------: | :--: |
|   SOLO   | ResNet-50 |        33.1        |      36.0      | [config](https://github.com/open-mmlab/mmdetection/blob/master/configs/solo/solo_r50_fpn_1x_coco.py) |                                                          |      |
| MaskRCNN | ResNet-50 |        35.4        |      37.8      | [config](https://github.com/open-mmlab/mmdetection/tree/master/configs/mask_rcnn/mask_rcnn_r50_fpn_2x_coco.py) | [baidu](https://pan.baidu.com/s/101eOFcD8JDwqrFuYcxcBIA) | sv8m |

| Student | Teacher | Baseline(mAP) | +FGD(mAP) |                            config                            | weight | code |
| :-----: | :-----: | :-----------: | :-------: | :----------------------------------------------------------: | :----: | :--: |
| YOLOX-m | YOLOX-l |     45.9      |   46.6    | [config](https://github.com/open-mmlab/mmdetection/blob/master/configs/yolox/yolox_m_8x8_300e_coco.py) |   [baidu](https://pan.baidu.com/s/1oagBUUV9RJReRdd4O-PV6Q?pwd=af9g)     |   af9g   |
1. Please refer branch yolox 

## Citation
```
@inproceedings{yang2022focal,
  title={Focal and global knowledge distillation for detectors},
  author={Yang, Zhendong and Li, Zhe and Jiang, Xiaohu and Gong, Yuan and Yuan, Zehuan and Zhao, Danpei and Yuan, Chun},
  booktitle={Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition},
  pages={4643--4652},
  year={2022}
}
```


## Acknowledgement

Our code is based on the project [MMDetection](https://github.com/open-mmlab/mmdetection).

Thanks to the work [GCNet](https://github.com/xvjiarui/GCNet) and [mmetection-distiller](https://github.com/pppppM/mmdetection-distiller).