Skip to content

Commit

Permalink
revert readme
Browse files Browse the repository at this point in the history
  • Loading branch information
nemonameless committed Jul 17, 2022
1 parent 857e407 commit a9c22a6
Showing 1 changed file with 28 additions and 3 deletions.
31 changes: 28 additions & 3 deletions README_cn.md
Original file line number Diff line number Diff line change
@@ -1,13 +1,38 @@
## 简介

此repository是基于[PaddleDetection](https://github.com/PaddlePaddle/PaddleDetection)`YOLO集合`,其upstream为PaddleDetection的[develop](https://github.com/PaddlePaddle/PaddleDetection/tree/develop)分支,是精简版本的代码,只保留YOLO系列模型(YOLOv3、PP-YOLO、PP-YOLOv2、PP-YOLOE、YOLOX、YOLOv5),以及保留了YOLO相关的多目标跟踪算法。
此repository是基于[PaddleDetection](https://github.com/PaddlePaddle/PaddleDetection)`YOLOv5`,其upstream为PaddleDetection的[develop](https://github.com/PaddlePaddle/PaddleDetection/tree/develop)分支,在保持与PaddleDetection同步更新的同时,也更新关于YOLOv5的一些改进,并同步更新github和gitee的代码,欢迎一起使用和建设!

**注意:**
- github链接为:https://github.com/nemonameless/PaddleDetection_YOLOset
- gitee链接为:https://gitee.com/nemonameless/PaddleDetection_YOLOset
- github链接为:https://github.com/nemonameless/PaddleDetection_YOLOv5
- gitee链接为:https://gitee.com/nemonameless/PaddleDetection_YOLOv5


## Updates!
* 【2022/05/15】新增[YOLOv5](configs/yolov5)目标检测模型,支持n/s/m/l/x版本。


## [ModelZoo](configs/yolov5) on COCO

| 网络网络 | 输入尺寸 | 图片数/GPU | 学习率策略 | 模型推理耗时(ms) | mAP | AP50 | 下载链接 | 配置文件 |
| :------------- | :------- | :-------: | :------: | :---------: | :-----: |:-----: | :-------------: | :-----: |
| YOLOv5-n | 640 | 16 | 300e | 2.6 | 28.0 | 45.7 | [下载链接](https://paddledet.bj.bcebos.com/models/yolov5_n_300e_coco.pdparams) | [配置文件](configs/yolov5/yolov5_n_300e_coco.yml) |
| YOLOv5-s | 640 | 8 | 300e | 3.2 | 37.0 | 55.9 | [下载链接](https://paddledet.bj.bcebos.com/models/yolov5_s_300e_coco.pdparams) | [配置文件](configs/yolov5/yolov5_s_300e_coco.yml) |
| YOLOv5-m | 640 | 8 | 300e | 5.2 | 45.3 | 63.8 | [下载链接](https://paddledet.bj.bcebos.com/models/yolov5_m_300e_coco.pdparams) | [配置文件](configs/yolov5/yolov5_m_300e_coco.yml) |
| YOLOv5-l | 640 | 8 | 300e | 7.9 | 48.6 | 66.9 | [下载链接](https://paddledet.bj.bcebos.com/models/yolov5_l_300e_coco.pdparams) | [配置文件](configs/yolov5/yolov5_l_300e_coco.yml) |
| YOLOv5-x | 640 | 8 | 300e | 13.7 | 50.6 | 68.7 | [下载链接](https://paddledet.bj.bcebos.com/models/yolov5_x_300e_coco.pdparams) | [配置文件](configs/yolov5/yolov5_x_300e_coco.yml) |

**注意:**
- 具体精度和速度细节请查看[yolov5](configs/yolov5)


## TODO
- [x] YOLOv5-P6、YOLOv5-P2
- [x] 换主干网络:CSP-ResNet
- [ ] [PaddleDetection_YOLOv7](https://github.com/nemonameless/PaddleDetection_YOLOv7)(Doing)
- [ ] Transformer modules
- [ ] 更多Data Augmentation、Attention Module、Loss
- [ ] 精简代码只保留YOLO系列模型(YOLOv3、PP-YOLO、PP-YOLOv2、PP-YOLOE、YOLOX、YOLOv5)
- [ ] Objects365 pretrain


======================================================
Expand Down

0 comments on commit a9c22a6

Please sign in to comment.