Skip to content

Commit

Permalink
update-project
Browse files Browse the repository at this point in the history
  • Loading branch information
z1069614715 committed Nov 19, 2023
1 parent c69308b commit 4b10815
Show file tree
Hide file tree
Showing 3 changed files with 16 additions and 6 deletions.
10 changes: 5 additions & 5 deletions yolo-improve/readme.md
Original file line number Diff line number Diff line change
Expand Up @@ -303,27 +303,27 @@
视频教学地址:[哔哩哔哩](https://www.bilibili.com/video/BV1PH4y1S7mf/).
reference: [链接](https://github.com/huawei-noah/Efficient-Computing/tree/master/Detection/Gold-YOLO)
paper: [链接](https://arxiv.org/abs/2309.11331)
- **yolov7-GOLDYOLO(文件在yolov5-GOLDYOLO的文件夹中)**
- **yolov7-GOLDYOLO(文件在yolov5-GOLDYOLO的文件夹中)**
利用华为2023最新GOLD-YOLO中的Gatherand-Distribute进行改进YOLOV7中的特征融合模块.
视频教学地址:[哔哩哔哩](https://www.bilibili.com/video/BV14V411c7H1/).
reference: [链接](https://github.com/huawei-noah/Efficient-Computing/tree/master/Detection/Gold-YOLO)
paper: [链接](https://arxiv.org/abs/2309.11331)
- **yolov5-DySnakeConv**
- **yolov5-DySnakeConv**
利用动态蛇形卷积改进YOLOV5.
视频教学地址:[哔哩哔哩](https://www.bilibili.com/video/BV1Qu411K7Hw/).
reference: [链接](https://github.com/YaoleiQi/DSCNet)
paper: [链接](https://arxiv.org/abs/2307.08388)
- **yolov7-DySnakeConv**
- **yolov7-DySnakeConv**
利用动态蛇形卷积改进YOLOV7.
视频教学地址:[哔哩哔哩](https://www.bilibili.com/video/BV1Wj411x7fq/).
reference: [链接](https://github.com/YaoleiQi/DSCNet)
paper: [链接](https://arxiv.org/abs/2307.08388)
- **yolov5-AIFI**
- **yolov5-AIFI**
利用带有位置信息编码的AIFI自注意力机制改进YOLOV5.
视频教学地址:[哔哩哔哩](https://www.bilibili.com/video/BV14V411c7H1/).
reference: [链接](https://github.com/lyuwenyu/RT-DETR)
paper: [链接](https://arxiv.org/pdf/2304.08069.pdf)
- **yolov7-AIFI**
- **yolov7-AIFI**
利用带有位置信息编码的AIFI自注意力机制改进YOLOV7.
视频教学地址:[哔哩哔哩](https://www.bilibili.com/video/BV14V411c7H1/).
reference: [链接](https://github.com/lyuwenyu/RT-DETR)
Expand Down
10 changes: 9 additions & 1 deletion yolo-improve/rtdetr-project.md
Original file line number Diff line number Diff line change
Expand Up @@ -131,4 +131,12 @@
3. 增加以yolov5和yolov8为基准模型更换为RTDETR的Head,后续也会提供yolov5-detr,yolov8-detr相关的改进.
4. 新增百度云视频-20231109更新说明视频和替换主干说明视频.
5. 新增GhostHGNetV2,RepHGNetV2,详细请看使用教程中的RT-DETR改进方案.
6. 新增使用DWRSeg中的Dilation-wise Residual(DWR)模块,加强从网络高层的可扩展感受野中提取特征,详细请看使用教程中的RT-DETR改进方案.
6. 新增使用DWRSeg中的Dilation-wise Residual(DWR)模块,加强从网络高层的可扩展感受野中提取特征,详细请看使用教程中的RT-DETR改进方案.

- **20231119-rtdetr-v1.2**
1. 增加DCNV2,DCNV3,DCNV2-Dynamic,并以RTDETR-R18,RTDETR-R50,YOLOV5-Detr,YOLOV8-Detr多个基准模型进行改进,详细请看使用教程中的RT-DETR改进方案.
2. 使用CVPR2022-OrthoNets中的正交通道注意力改进resnet18-backbone中的BasicBlock,resnet50-backbone中的BottleNeck,yolov8-C2f,yolov5-C3,详细请看使用教程中的RT-DETR改进方案.
3. 使用LearnedPositionalEncoding改进AIFI中的位置编码信息生成,详细请看使用教程中的RT-DETR改进方案.
4. 增加EMO模型中的iRMB模块,并使用(EfficientViT-CVPR2023)中的CascadedAttention对其二次创新得到iRMB_Cascaded,详细请看使用教程中的RT-DETR改进方案.
5. 百度云视频增加1119更新说明和手把手添加注意力机制视频教学.
6. 更新使用教程.
2 changes: 2 additions & 0 deletions yolo-improve/yolov5v7-light.md
Original file line number Diff line number Diff line change
Expand Up @@ -23,6 +23,8 @@
6. LAMP
7. GroupSL
8. GroupReg
9. GroupHessian
10. GroupTaylor

### 其中prune系列还有一些细节:
1. 支持稀疏训练时候可视化BN稀疏程度和数值。
Expand Down

0 comments on commit 4b10815

Please sign in to comment.