mirror of https://github.com/open-mmlab/mmyolo.git
Add v0.3.0 changelog (#440)
* bump version to v0.3.0 * add v7 cat demo config * update mmdet version * update * fix topk out of range error * update * add changelog.md * update * update * update * update * updatepull/443/head
Haian Huang(深度眸)
GitHub
parent
fc107b8997
commit
9c6b4fedc8
40
README.md
40
README.md
|
|
@ -71,14 +71,32 @@ And the figure of P6 model is in [model_design.md](docs/en/algorithm_description
|
|||
|
||||
## What's New
|
||||
|
||||
💎 **v0.2.0** was released on 1/12/2022:
|
||||
### Highlight
|
||||
|
||||
1. Support [YOLOv7](https://github.com/open-mmlab/mmyolo/tree/dev/configs/yolov7) P5 and P6 model
|
||||
2. Support [YOLOv6](https://github.com/open-mmlab/mmyolo/blob/dev/configs/yolov6/README.md) ML model
|
||||
3. Support [Grad-Based CAM and Grad-Free CAM](https://github.com/open-mmlab/mmyolo/blob/dev/demo/boxam_vis_demo.py)
|
||||
4. Support [large image inference](https://github.com/open-mmlab/mmyolo/blob/dev/demo/large_image_demo.py) based on sahi
|
||||
5. Add [easydeploy](https://github.com/open-mmlab/mmyolo/blob/dev/projects/easydeploy/README.md) project under the projects folder
|
||||
6. Add [custom dataset guide](https://github.com/open-mmlab/mmyolo/blob/dev/docs/zh_cn/user_guides/custom_dataset.md)
|
||||
We are excited to announce our latest work on real-time object recognition tasks, **RTMDet**, a family of fully convolutional single-stage detectors. RTMDet not only achieves the best parameter-accuracy trade-off on object detection from tiny to extra-large model sizes but also obtains new state-of-the-art performance on instance segmentation and rotated object detection tasks. Details can be found in the [technical report](https://arxiv.org/abs/2212.07784). Pre-trained models are [here](configs/rtmdet).
|
||||
|
||||
[](https://paperswithcode.com/sota/real-time-instance-segmentation-on-mscoco?p=rtmdet-an-empirical-study-of-designing-real)
|
||||
[](https://paperswithcode.com/sota/object-detection-in-aerial-images-on-dota-1?p=rtmdet-an-empirical-study-of-designing-real)
|
||||
[](https://paperswithcode.com/sota/object-detection-in-aerial-images-on-hrsc2016?p=rtmdet-an-empirical-study-of-designing-real)
|
||||
|
||||
| Task | Dataset | AP | FPS(TRT FP16 BS1 3090) |
|
||||
| ------------------------ | ------- | ------------------------------------ | ---------------------- |
|
||||
| Object Detection | COCO | 52.8 | 322 |
|
||||
| Instance Segmentation | COCO | 44.6 | 188 |
|
||||
| Rotated Object Detection | DOTA | 78.9(single-scale)/81.3(multi-scale) | 121 |
|
||||
|
||||
<div align=center>
|
||||
<img src="https://user-images.githubusercontent.com/12907710/208044554-1e8de6b5-48d8-44e4-a7b5-75076c7ebb71.png"/>
|
||||
</div>
|
||||
|
||||
MMYOLO currently only implements the object detection algorithm, but it has a significant training acceleration compared to the MMDeteciton version. The training speed is 2.6 times faster than the previous version.
|
||||
|
||||
💎 **v0.3.0** was released on 8/1/2023:
|
||||
|
||||
1. Implement fast version of [RTMDet](https://github.com/open-mmlab/mmyolo/blob/dev/configs/rtmdet/README.md). RTMDet-s 8xA100 training takes only 14 hours. The training speed is 2.6 times faster than the previous version.
|
||||
2. Support [PPYOLOE](https://github.com/open-mmlab/mmyolo/blob/dev/configs/ppyoloe/README.md) training
|
||||
3. Support `iscrowd` attribute training in [YOLOv5](https://github.com/open-mmlab/mmyolo/blob/dev/configs/yolov5/crowdhuman/yolov5_s-v61_8xb16-300e_ignore_crowdhuman.py)
|
||||
4. Support [YOLOv5 assigner result visualization](https://github.com/open-mmlab/mmyolo/blob/dev/projects/assigner_visualization/README.md)
|
||||
|
||||
For release history and update details, please refer to [changelog](https://mmyolo.readthedocs.io/en/latest/notes/changelog.html).
|
||||
|
||||
|
|
@ -92,7 +110,7 @@ conda activate open-mmlab
|
|||
pip install openmim
|
||||
mim install "mmengine>=0.3.1"
|
||||
mim install "mmcv>=2.0.0rc1,<2.1.0"
|
||||
mim install "mmdet>=3.0.0rc3,<3.1.0"
|
||||
mim install "mmdet>=3.0.0rc5,<3.1.0"
|
||||
git clone https://github.com/open-mmlab/mmyolo.git
|
||||
cd mmyolo
|
||||
# Install albumentations
|
||||
|
|
@ -152,7 +170,7 @@ Results and models are available in the [model zoo](docs/en/model_zoo.md).
|
|||
- [x] [RTMDet](configs/rtmdet)
|
||||
- [x] [YOLOv6](configs/yolov6)
|
||||
- [x] [YOLOv7](configs/yolov7)
|
||||
- [ ] [PPYOLOE](configs/ppyoloe)(Inference only)
|
||||
- [x] [PPYOLOE](configs/ppyoloe)
|
||||
|
||||
</details>
|
||||
|
||||
|
|
@ -183,6 +201,8 @@ Results and models are available in the [model zoo](docs/en/model_zoo.md).
|
|||
<li>YOLOXCSPDarknet</li>
|
||||
<li>EfficientRep</li>
|
||||
<li>CSPNeXt</li>
|
||||
<li>YOLOv7Backbone</li>
|
||||
<li>PPYOLOECSPResNet</li>
|
||||
</ul>
|
||||
</td>
|
||||
<td>
|
||||
|
|
@ -191,6 +211,8 @@ Results and models are available in the [model zoo](docs/en/model_zoo.md).
|
|||
<li>YOLOv6RepPAFPN</li>
|
||||
<li>YOLOXPAFPN</li>
|
||||
<li>CSPNeXtPAFPN</li>
|
||||
<li>YOLOv7PAFPN</li>
|
||||
<li>PPYOLOECSPPAFPN</li>
|
||||
</ul>
|
||||
</td>
|
||||
<td>
|
||||
|
|
|
|||
|
|
@ -71,25 +71,46 @@ P6 模型图详见 [model_design.md](docs/zh_CN/algorithm_descriptions/model_des
|
|||
|
||||
## 最新进展
|
||||
|
||||
💎 **v0.2.0** 版本已经在 2022.12.1 发布：
|
||||
### 亮点
|
||||
|
||||
1. 支持 [YOLOv7](https://github.com/open-mmlab/mmyolo/tree/dev/configs/yolov7) P5 和 P6 模型
|
||||
2. 支持 [YOLOv6](https://github.com/open-mmlab/mmyolo/blob/dev/configs/yolov6/README.md) 中的 ML 大模型
|
||||
3. 支持 [Grad-Based CAM 和 Grad-Free CAM](https://github.com/open-mmlab/mmyolo/blob/dev/demo/boxam_vis_demo.py)
|
||||
4. 基于 sahi 支持 [大图推理](https://github.com/open-mmlab/mmyolo/blob/dev/demo/large_image_demo.py)
|
||||
5. projects 文件夹下新增 [easydeploy](https://github.com/open-mmlab/mmyolo/blob/dev/projects/easydeploy/README.md) 项目
|
||||
6. 新增 [自定义数据集教程](https://github.com/open-mmlab/mmyolo/blob/dev/docs/zh_cn/user_guides/custom_dataset.md)
|
||||
我们很高兴向大家介绍我们在实时目标识别任务方面的最新成果 RTMDet，包含了一系列的全卷积单阶段检测模型。 RTMDet 不仅在从 tiny 到 extra-large 尺寸的目标检测模型上实现了最佳的参数量和精度的平衡，而且在实时实例分割和旋转目标检测任务上取得了最先进的成果。 更多细节请参阅[技术报告](https://arxiv.org/abs/2212.07784)。 预训练模型可以在[这里](configs/rtmdet)找到。
|
||||
|
||||
[](https://paperswithcode.com/sota/real-time-instance-segmentation-on-mscoco?p=rtmdet-an-empirical-study-of-designing-real)
|
||||
[](https://paperswithcode.com/sota/object-detection-in-aerial-images-on-dota-1?p=rtmdet-an-empirical-study-of-designing-real)
|
||||
[](https://paperswithcode.com/sota/object-detection-in-aerial-images-on-hrsc2016?p=rtmdet-an-empirical-study-of-designing-real)
|
||||
|
||||
| Task | Dataset | AP | FPS(TRT FP16 BS1 3090) |
|
||||
| ------------------------ | ------- | ------------------------------------ | ---------------------- |
|
||||
| Object Detection | COCO | 52.8 | 322 |
|
||||
| Instance Segmentation | COCO | 44.6 | 188 |
|
||||
| Rotated Object Detection | DOTA | 78.9(single-scale)/81.3(multi-scale) | 121 |
|
||||
|
||||
<div align=center>
|
||||
<img src="https://user-images.githubusercontent.com/12907710/208044554-1e8de6b5-48d8-44e4-a7b5-75076c7ebb71.png"/>
|
||||
</div>
|
||||
|
||||
MMYOLO 中目前仅仅实现了目标检测算法，但是相比 MMDeteciton 版本有显著训练加速，训练速度相比原先版本提升 2.6 倍。
|
||||
|
||||
💎 **v0.3.0** 版本已经在 2023.1.8 发布：
|
||||
|
||||
1. 实现了 [RTMDet](https://github.com/open-mmlab/mmyolo/blob/dev/configs/rtmdet/README.md) 的快速版本。RTMDet-s 8xA100 训练只需要 14 个小时，训练速度相比原先版本提升 2.6 倍。
|
||||
2. 支持 [PPYOLOE](https://github.com/open-mmlab/mmyolo/blob/dev/configs/ppyoloe/README.md) 训练。
|
||||
3. 支持 [YOLOv5](https://github.com/open-mmlab/mmyolo/blob/dev/configs/yolov5/crowdhuman/yolov5_s-v61_8xb16-300e_ignore_crowdhuman.py) 的 `iscrowd` 属性训练。
|
||||
4. 支持 [YOLOv5 正样本分配结果可视化](https://github.com/open-mmlab/mmyolo/blob/dev/projects/assigner_visualization/README.md)
|
||||
5. 新增 [YOLOv6 原理和实现全解析文档](https://github.com/open-mmlab/mmyolo/blob/dev/docs/zh_cn/algorithm_descriptions/yolov6_description.md)
|
||||
|
||||
同时我们也推出了解读视频：
|
||||
|
||||
| | 内容 | 视频 | 课程中的代码 |
|
||||
| :-: | :------------------------: | :------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------: | :---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------: |
|
||||
| 🌟 | 特征图可视化 | [](https://www.bilibili.com/video/BV188411s7o8) [](https://www.bilibili.com/video/BV188411s7o8) | [特征图可视化.ipynb](https://github.com/open-mmlab/OpenMMLabCourse/blob/main/codes/MMYOLO_tutorials/%5B%E5%B7%A5%E5%85%B7%E7%B1%BB%E7%AC%AC%E4%B8%80%E6%9C%9F%5D%E7%89%B9%E5%BE%81%E5%9B%BE%E5%8F%AF%E8%A7%86%E5%8C%96.ipynb) |
|
||||
| 🌟 | 特征图可视化 Demo | [](https://www.bilibili.com/video/BV1je4y1478R/) [](https://www.bilibili.com/video/BV1je4y1478R/) | |
|
||||
| 🌟 | 配置全解读 | [](https://www.bilibili.com/video/BV1214y157ck) [](https://www.bilibili.com/video/BV1214y157ck) | [配置全解读文档](https://zhuanlan.zhihu.com/p/577715188) |
|
||||
| 🌟 | 源码阅读和调试「必备」技巧 | [](https://www.bilibili.com/video/BV1N14y1V7mB) [](https://www.bilibili.com/video/BV1N14y1V7mB) | [源码阅读和调试「必备」技巧文档](https://zhuanlan.zhihu.com/p/580885852) |
|
||||
| 🌟 | 工程文件结构简析 | [](https://www.bilibili.com/video/BV1LP4y117jS)[](https://www.bilibili.com/video/BV1LP4y117jS) | [工程文件结构简析文档](https://zhuanlan.zhihu.com/p/584807195) |
|
||||
| 🌟 | 10分钟换遍主干网络 | [](https://www.bilibili.com/video/BV1JG4y1d7GC) [](https://www.bilibili.com/video/BV1JG4y1d7GC) | [10分钟换遍主干网络文档](https://zhuanlan.zhihu.com/p/585641598)<br>[10分钟换遍主干网络.ipynb](https://github.com/open-mmlab/OpenMMLabCourse/blob/main/codes/MMYOLO_tutorials/[实用类第二期]10分钟换遍主干网络.ipynb) |
|
||||
| | 内容 | 视频 | 课程中的代码 |
|
||||
| :-: | :--------------------------------: | :-------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------: | :---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------: |
|
||||
| 🌟 | 特征图可视化 | [](https://www.bilibili.com/video/BV188411s7o8) [](https://www.bilibili.com/video/BV188411s7o8) | [特征图可视化.ipynb](https://github.com/open-mmlab/OpenMMLabCourse/blob/main/codes/MMYOLO_tutorials/%5B%E5%B7%A5%E5%85%B7%E7%B1%BB%E7%AC%AC%E4%B8%80%E6%9C%9F%5D%E7%89%B9%E5%BE%81%E5%9B%BE%E5%8F%AF%E8%A7%86%E5%8C%96.ipynb) |
|
||||
| 🌟 | 特征图可视化 Demo | [](https://www.bilibili.com/video/BV1je4y1478R/) [](https://www.bilibili.com/video/BV1je4y1478R/) | |
|
||||
| 🌟 | 配置全解读 | [](https://www.bilibili.com/video/BV1214y157ck) [](https://www.bilibili.com/video/BV1214y157ck) | [配置全解读文档](https://zhuanlan.zhihu.com/p/577715188) |
|
||||
| 🌟 | 源码阅读和调试「必备」技巧 | [](https://www.bilibili.com/video/BV1N14y1V7mB) [](https://www.bilibili.com/video/BV1N14y1V7mB) | [源码阅读和调试「必备」技巧文档](https://zhuanlan.zhihu.com/p/580885852) |
|
||||
| 🌟 | 工程文件结构简析 | [](https://www.bilibili.com/video/BV1LP4y117jS)[](https://www.bilibili.com/video/BV1LP4y117jS) | [工程文件结构简析文档](https://zhuanlan.zhihu.com/p/584807195) |
|
||||
| 🌟 | 10分钟换遍主干网络 | [](https://www.bilibili.com/video/BV1JG4y1d7GC) [](https://www.bilibili.com/video/BV1JG4y1d7GC) | [10分钟换遍主干网络文档](https://zhuanlan.zhihu.com/p/585641598)<br>[10分钟换遍主干网络.ipynb](https://github.com/open-mmlab/OpenMMLabCourse/blob/main/codes/MMYOLO_tutorials/[实用类第二期]10分钟换遍主干网络.ipynb) |
|
||||
| 🌟 | 基于 sahi 的大图推理 | [](https://www.bilibili.com/video/BV1EK411R7Ws/) [](https://www.bilibili.com/video/BV1EK411R7Ws/) | [10分钟轻松掌握大图推理.ipynb](https://github.com/open-mmlab/OpenMMLabCourse/blob/main/codes/MMYOLO_tutorials/[工具类第二期]10分钟轻松掌握大图推理.ipynb) |
|
||||
| 🌟 | 自定义数据集从标注到部署保姆级教程 | [](https://www.bilibili.com/video/BV1RG4y137i5) [](https://www.bilibili.com/video/BV1JG4y1d7GC) | [自定义数据集从标注到部署保姆级教程](https://github.com/open-mmlab/mmyolo/blob/dev/docs/zh_cn/user_guides/custom_dataset.md) |
|
||||
|
||||
发布历史和更新细节请参考 [更新日志](https://mmyolo.readthedocs.io/zh_CN/latest/notes/changelog.html)
|
||||
|
||||
|
|
@ -103,7 +124,7 @@ conda activate open-mmlab
|
|||
pip install openmim
|
||||
mim install "mmengine>=0.3.1"
|
||||
mim install "mmcv>=2.0.0rc1,<2.1.0"
|
||||
mim install "mmdet>=3.0.0rc3,<3.1.0"
|
||||
mim install "mmdet>=3.0.0rc5,<3.1.0"
|
||||
git clone https://github.com/open-mmlab/mmyolo.git
|
||||
cd mmyolo
|
||||
# Install albumentations
|
||||
|
|
@ -149,6 +170,7 @@ MMYOLO 用法和 MMDetection 几乎一致，所有教程都是通用的，你也
|
|||
|
||||
- 进阶指南
|
||||
|
||||
- [模块组合](docs/zh_cn/advanced_guides/module_combination.md)
|
||||
- [数据流](docs/zh_cn/advanced_guides/data_flow.md)
|
||||
- [How to](docs/zh_cn/advanced_guides/how_to.md)
|
||||
- [插件](docs/zh_cn/advanced_guides/plugins.md)
|
||||
|
|
@ -167,7 +189,7 @@ MMYOLO 用法和 MMDetection 几乎一致，所有教程都是通用的，你也
|
|||
- [x] [RTMDet](configs/rtmdet)
|
||||
- [x] [YOLOv6](configs/yolov6)
|
||||
- [x] [YOLOv7](configs/yolov7)
|
||||
- [ ] [PPYOLOE](configs/ppyoloe)(仅推理)
|
||||
- [x] [PPYOLOE](configs/ppyoloe)
|
||||
|
||||
</details>
|
||||
|
||||
|
|
@ -198,6 +220,8 @@ MMYOLO 用法和 MMDetection 几乎一致，所有教程都是通用的，你也
|
|||
<li>YOLOXCSPDarknet</li>
|
||||
<li>EfficientRep</li>
|
||||
<li>CSPNeXt</li>
|
||||
<li>YOLOv7Backbone</li>
|
||||
<li>PPYOLOECSPResNet</li>
|
||||
</ul>
|
||||
</td>
|
||||
<td>
|
||||
|
|
@ -206,6 +230,8 @@ MMYOLO 用法和 MMDetection 几乎一致，所有教程都是通用的，你也
|
|||
<li>YOLOv6RepPAFPN</li>
|
||||
<li>YOLOXPAFPN</li>
|
||||
<li>CSPNeXtPAFPN</li>
|
||||
<li>YOLOv7PAFPN</li>
|
||||
<li>PPYOLOECSPPAFPN</li>
|
||||
</ul>
|
||||
</td>
|
||||
<td>
|
||||
|
|
|
|||
|
|
@ -26,7 +26,7 @@ RUN apt-get update \
|
|||
|
||||
# Install MMEngine , MMCV and MMDet
|
||||
RUN pip install --no-cache-dir openmim && \
|
||||
mim install --no-cache-dir "mmengine>=0.3.1" "mmcv>=2.0.0rc1,<2.1.0" "mmdet>=3.0.0rc3,<3.1.0"
|
||||
mim install --no-cache-dir "mmengine>=0.3.1" "mmcv>=2.0.0rc1,<2.1.0" "mmdet>=3.0.0rc5,<3.1.0"
|
||||
|
||||
# Install MMYOLO
|
||||
RUN git clone https://github.com/open-mmlab/mmyolo.git /mmyolo && \
|
||||
|
|
|
|||
|
|
@ -30,7 +30,7 @@ RUN wget -q https://github.com/microsoft/onnxruntime/releases/download/v${ONNXRU
|
|||
|
||||
# Install OPENMIM MMENGINE MMDET
|
||||
RUN pip install --no-cache-dir openmim \
|
||||
&& mim install --no-cache-dir "mmengine>=0.3.1" "mmdet>=3.0.0rc3,<3.1.0" \
|
||||
&& mim install --no-cache-dir "mmengine>=0.3.1" "mmdet>=3.0.0rc5,<3.1.0" \
|
||||
&& mim install --no-cache-dir opencv-python==4.5.5.64 opencv-python-headless==4.5.5.64
|
||||
|
||||
RUN git clone https://github.com/open-mmlab/mmcv.git -b 2.x mmcv \
|
||||
|
|
|
|||
|
|
@ -6,7 +6,8 @@ Compatible MMEngine, MMCV and MMDetection versions are shown as below. Please in
|
|||
|
||||
| MMYOLO version | MMDetection version | MMEngine version | MMCV version |
|
||||
| :------------: | :----------------------: | :----------------------: | :---------------------: |
|
||||
| main | mmdet>=3.0.0rc3, \<3.1.0 | mmengine>=0.3.1, \<1.0.0 | mmcv>=2.0.0rc0, \<2.1.0 |
|
||||
| main | mmdet>=3.0.0rc5, \<3.1.0 | mmengine>=0.3.1, \<1.0.0 | mmcv>=2.0.0rc0, \<2.1.0 |
|
||||
| 0.3.0 | mmdet>=3.0.0rc5, \<3.1.0 | mmengine>=0.3.1, \<1.0.0 | mmcv>=2.0.0rc0, \<2.1.0 |
|
||||
| 0.2.0 | mmdet>=3.0.0rc3, \<3.1.0 | mmengine>=0.3.1, \<1.0.0 | mmcv>=2.0.0rc0, \<2.1.0 |
|
||||
| 0.1.3 | mmdet>=3.0.0rc3, \<3.1.0 | mmengine>=0.3.1, \<1.0.0 | mmcv>=2.0.0rc0, \<2.1.0 |
|
||||
| 0.1.2 | mmdet>=3.0.0rc2, \<3.1.0 | mmengine>=0.3.0, \<1.0.0 | mmcv>=2.0.0rc0, \<2.1.0 |
|
||||
|
|
@ -54,7 +55,7 @@ conda install pytorch torchvision cpuonly -c pytorch
|
|||
pip install -U openmim
|
||||
mim install "mmengine>=0.3.1"
|
||||
mim install "mmcv>=2.0.0rc1,<2.1.0"
|
||||
mim install "mmdet>=3.0.0rc3,<3.1.0"
|
||||
mim install "mmdet>=3.0.0rc5,<3.1.0"
|
||||
```
|
||||
|
||||
**Note:**
|
||||
|
|
@ -213,7 +214,7 @@ thus we only need to install MMEngine, MMCV, MMDetection, and MMYOLO with the fo
|
|||
!pip3 install openmim
|
||||
!mim install "mmengine==0.1.0"
|
||||
!mim install "mmcv>=2.0.0rc1,<2.1.0"
|
||||
!mim install "mmdet>=3.0.0.rc1"
|
||||
!mim install "mmdet>=3.0.0rc5,<3.1.0"
|
||||
```
|
||||
|
||||
**Step 2.** Install MMYOLO from the source.
|
||||
|
|
|
|||
|
|
@ -1,5 +1,67 @@
|
|||
# Changelog
|
||||
|
||||
## v0.3.0 (8/1/2023)
|
||||
|
||||
### Highlights
|
||||
|
||||
1. Implement fast version of [RTMDet](https://github.com/open-mmlab/mmyolo/blob/dev/configs/rtmdet/README.md). RTMDet-s 8xA100 training takes only 14 hours. The training speed is 2.6 times faster than the previous version.
|
||||
2. Support [PPYOLOE](https://github.com/open-mmlab/mmyolo/blob/dev/configs/ppyoloe/README.md) training
|
||||
3. Support `iscrowd` attribute training in [YOLOv5](https://github.com/open-mmlab/mmyolo/blob/dev/configs/yolov5/crowdhuman/yolov5_s-v61_8xb16-300e_ignore_crowdhuman.py)
|
||||
4. Support [YOLOv5 assigner result visualization](https://github.com/open-mmlab/mmyolo/blob/dev/projects/assigner_visualization/README.md)
|
||||
|
||||
### New Features
|
||||
|
||||
01. Add `crowdhuman` dataset (#368)
|
||||
02. Easydeploy support TensorRT inference (#377)
|
||||
03. Add `YOLOX` structure description (#402)
|
||||
04. Add a feature for the video demo (#392)
|
||||
05. Support `YOLOv7` easy deploy (#427)
|
||||
06. Add resume from specific checkpoint in CLI (#393)
|
||||
07. Set `metainfo` fields to lower case (#362, #412)
|
||||
08. Add module combination doc (#349, #352, #345)
|
||||
09. Add docs about how to freeze the weight of backbone or neck (#418)
|
||||
10. Add don't used pre-training weights doc in `how_to.md` (#404)
|
||||
11. Add docs about how to set the random seed (#386)
|
||||
12. Translate `rtmdet_description.md` document to English (#353)
|
||||
13. Add doc of `yolov6_description.md` (#382, #372)
|
||||
|
||||
### Bug Fixes
|
||||
|
||||
01. Fix bugs in the output annotation file when `--class-id-txt` is set (#430)
|
||||
02. Fix batch inference bug in `YOLOv5` head (#413)
|
||||
03. Fix typehint in some heads (#415, #416, #443)
|
||||
04. Fix RuntimeError of `torch.cat()` expected a non-empty list of Tensors (#376)
|
||||
05. Fix the device inconsistency error in `YOLOv7` training (#397)
|
||||
06. Fix the `scale_factor` and `pad_param` value in `LetterResize` (#387)
|
||||
07. Fix docstring graph rendering error of readthedocs (#400)
|
||||
08. Fix AssertionError when `YOLOv6` from training to val (#378)
|
||||
09. Fix CI error due to `np.int` and legacy builder.py (#389)
|
||||
10. Fix MMDeploy rewriter (#366)
|
||||
11. Fix MMYOLO unittest scope bug (#351)
|
||||
12. Fix `pad_param` error (#354)
|
||||
13. Fix twice head inference bug (#342)
|
||||
14. Fix customize dataset training (#428)
|
||||
|
||||
### Improvements
|
||||
|
||||
01. Update `useful_tools.md` (#384)
|
||||
02. update the English version of `custom_dataset.md` (#381)
|
||||
03. Remove context argument from the rewriter function (#395)
|
||||
04. deprecating `np.bool` type alias (#396)
|
||||
05. Add new video link for custom dataset (#365)
|
||||
06. Export onnx for model only (#361)
|
||||
07. Add MMYOLO regression test yml (#359)
|
||||
08. Update video tutorials in `article.md` (#350)
|
||||
09. Add deploy demo (#343)
|
||||
10. Optimize the vis results of large images in debug mode (#346)
|
||||
11. Improve args for `browse_dataset` and support `RepeatDataset` (#340, #338)
|
||||
|
||||
### Contributors
|
||||
|
||||
A total of 28 developers contributed to this release.
|
||||
|
||||
Thank @RangeKing, @PeterH0323, @Nioolek, @triple-Mu, @matrixgame2018, @xin-li-67, @tang576225574, @kitecats, @Seperendity, @diplomatist, @vaew, @wzr-skn, @VoyagerXvoyagerx, @MambaWong, @tianleiSHI, @caj-github, @zhubochao, @lvhan028, @dsghaonan, @lyviva, @yuewangg, @wang-tf, @satuoqaq, @grimoire, @RunningLeon, @hanrui1sensetime, @RangiLyu, @hhaAndroid
|
||||
|
||||
## v0.2.0（1/12/2022)
|
||||
|
||||
### Highlights
|
||||
|
|
|
|||
|
|
@ -12,7 +12,7 @@ conda install pytorch torchvision -c pytorch
|
|||
pip install -U openmim
|
||||
mim install "mmengine>=0.3.1"
|
||||
mim install "mmcv>=2.0.0rc1,<2.1.0"
|
||||
mim install "mmdet>=3.0.0rc3,<3.1.0"
|
||||
mim install "mmdet>=3.0.0rc5,<3.1.0"
|
||||
git clone https://github.com/open-mmlab/mmyolo.git
|
||||
cd mmyolo
|
||||
# Install albumentations
|
||||
|
|
|
|||
|
|
@ -1,3 +1,11 @@
|
|||
模块组合
|
||||
************************
|
||||
|
||||
.. toctree::
|
||||
:maxdepth: 1
|
||||
|
||||
module_combination.md
|
||||
|
||||
数据流
|
||||
************************
|
||||
|
||||
|
|
|
|||
|
|
@ -7,18 +7,14 @@
|
|||
### 文章
|
||||
|
||||
- [社区协作，简洁易用，快来开箱新一代 YOLO 系列开源库](https://zhuanlan.zhihu.com/p/575615805)
|
||||
|
||||
- [MMYOLO 社区倾情贡献，RTMDet 原理社区开发者解读来啦！](https://zhuanlan.zhihu.com/p/569777684)
|
||||
|
||||
- [玩转 MMYOLO 基础类第一期： 配置文件太复杂？继承用法看不懂？配置全解读来了](https://zhuanlan.zhihu.com/p/577715188)
|
||||
|
||||
- [玩转 MMYOLO 工具类第一期： 特征图可视化](https://zhuanlan.zhihu.com/p/578141381?)
|
||||
|
||||
- [玩转 MMYOLO 实用类第二期：源码阅读和调试「必备」技巧文档](https://zhuanlan.zhihu.com/p/580885852)
|
||||
|
||||
- [玩转 MMYOLO 基础类第二期：工程文件结构简析](https://zhuanlan.zhihu.com/p/584807195)
|
||||
|
||||
- [玩转 MMYOLO 实用类第二期：10分钟换遍主干网络文档](https://zhuanlan.zhihu.com/p/585641598)
|
||||
- [MMYOLO 自定义数据集从标注到部署保姆级教程](https://zhuanlan.zhihu.com/p/595497726)
|
||||
- [满足一切需求的 MMYOLO 可视化：测试过程可视化](https://zhuanlan.zhihu.com/p/593179372)
|
||||
|
||||
### 视频
|
||||
|
||||
|
|
|
|||
|
|
@ -6,7 +6,8 @@
|
|||
|
||||
| MMYOLO version | MMDetection version | MMEngine version | MMCV version |
|
||||
| :------------: | :----------------------: | :----------------------: | :---------------------: |
|
||||
| main | mmdet>=3.0.0rc3, \<3.1.0 | mmengine>=0.3.1, \<1.0.0 | mmcv>=2.0.0rc0, \<2.1.0 |
|
||||
| main | mmdet>=3.0.0rc5, \<3.1.0 | mmengine>=0.3.1, \<1.0.0 | mmcv>=2.0.0rc0, \<2.1.0 |
|
||||
| 0.3.0 | mmdet>=3.0.0rc5, \<3.1.0 | mmengine>=0.3.1, \<1.0.0 | mmcv>=2.0.0rc0, \<2.1.0 |
|
||||
| 0.2.0 | mmdet>=3.0.0rc3, \<3.1.0 | mmengine>=0.3.1, \<1.0.0 | mmcv>=2.0.0rc0, \<2.1.0 |
|
||||
| 0.1.3 | mmdet>=3.0.0rc3, \<3.1.0 | mmengine>=0.3.1, \<1.0.0 | mmcv>=2.0.0rc0, \<2.1.0 |
|
||||
| 0.1.2 | mmdet>=3.0.0rc2, \<3.1.0 | mmengine>=0.3.0, \<1.0.0 | mmcv>=2.0.0rc0, \<2.1.0 |
|
||||
|
|
@ -54,7 +55,7 @@ conda install pytorch torchvision cpuonly -c pytorch
|
|||
pip install -U openmim
|
||||
mim install "mmengine>=0.3.1"
|
||||
mim install "mmcv>=2.0.0rc1,<2.1.0"
|
||||
mim install "mmdet>=3.0.0rc3,<3.1.0"
|
||||
mim install "mmdet>=3.0.0rc5,<3.1.0"
|
||||
```
|
||||
|
||||
**注意：**
|
||||
|
|
@ -214,7 +215,7 @@ pip install "mmcv>=2.0.0rc1" -f https://download.openmmlab.com/mmcv/dist/cu116/t
|
|||
!pip3 install openmim
|
||||
!mim install "mmengine==0.1.0"
|
||||
!mim install "mmcv>=2.0.0rc1,<2.1.0"
|
||||
!mim install "mmdet>=3.0.0.rc1"
|
||||
!mim install "mmdet>=3.0.0rc5,<3.1.0"
|
||||
```
|
||||
|
||||
**步骤 2.** 使用源码安装 MMYOLO：
|
||||
|
|
|
|||
|
|
@ -1,5 +1,72 @@
|
|||
# 更新日志
|
||||
|
||||
## v0.3.0 (8/1/2023)
|
||||
|
||||
### 亮点
|
||||
|
||||
1. 实现了 [RTMDet](https://github.com/open-mmlab/mmyolo/blob/dev/configs/rtmdet/README.md) 的快速版本。RTMDet-s 8xA100 训练只需要 14 个小时，训练速度相比原先版本提升 2.6 倍。
|
||||
2. 支持 [PPYOLOE](https://github.com/open-mmlab/mmyolo/blob/dev/configs/ppyoloe/README.md) 训练。
|
||||
3. 支持 [YOLOv5](https://github.com/open-mmlab/mmyolo/blob/dev/configs/yolov5/crowdhuman/yolov5_s-v61_8xb16-300e_ignore_crowdhuman.py) 的 `iscrowd` 属性训练。
|
||||
4. 支持 [YOLOv5 正样本分配结果可视化](https://github.com/open-mmlab/mmyolo/blob/dev/projects/assigner_visualization/README.md)
|
||||
5. 新增 [YOLOv6 原理和实现全解析文档](https://github.com/open-mmlab/mmyolo/blob/dev/docs/zh_cn/algorithm_descriptions/yolov6_description.md)
|
||||
|
||||
### 新特性
|
||||
|
||||
01. 新增 `crowdhuman` 数据集 (#368)
|
||||
02. EasyDeploy 中支持 TensorRT 推理 (#377)
|
||||
03. 新增 `YOLOX` 结构图描述 (#402)
|
||||
04. 新增视频推理脚本 (#392)
|
||||
05. EasyDeploy 中支持 `YOLOv7` 部署 (#427)
|
||||
06. 支持从 CLI 中的特定检查点恢复训练 (#393)
|
||||
07. 将元信息字段设置为小写（#362、#412）
|
||||
08. 新增模块组合文档 (#349, #352, #345)
|
||||
09. 新增关于如何冻结 backbone 或 neck 权重的文档 (#418)
|
||||
10. 在 `how_to.md` 中添加不使用预训练权重的文档 (#404)
|
||||
11. 新增关于如何设置随机种子的文档 (#386)
|
||||
12. 将 `rtmdet_description.md` 文档翻译成英文 (#353)
|
||||
|
||||
### Bug 修复
|
||||
|
||||
01. 修复设置 `--class-id-txt` 时输出注释文件中的错误 (#430)
|
||||
02. 修复 `YOLOv5` head 中的批量推理错误 (#413)
|
||||
03. 修复某些 head 的类型提示（#415、#416、#443）
|
||||
04. 修复 expected a non-empty list of Tensors 错误 (#376)
|
||||
05. 修复 `YOLOv7` 训练中的设备不一致错误（#397）
|
||||
06. 修复 `LetterResize` 中的 `scale_factor` 和 `pad_param` 值 (#387)
|
||||
07. 修复 readthedocs 的 docstring 图形渲染错误 (#400)
|
||||
08. 修复 `YOLOv6` 从训练到验证时的断言错误 (#378)
|
||||
09. 修复 `np.int` 和旧版 builder.py 导致的 CI 错误 (#389)
|
||||
10. 修复 MMDeploy 重写器 (#366)
|
||||
11. 修复 MMYOLO 单元测试错误 (#351)
|
||||
12. 修复 `pad_param` 错误 (#354)
|
||||
13. 修复 head 推理两次的错误（#342）
|
||||
14. 修复自定义数据集训练 (#428)
|
||||
|
||||
### 完善
|
||||
|
||||
01. 更新 `useful_tools.md` (#384)
|
||||
02. 更新英文版 `custom_dataset.md` (#381)
|
||||
03. 重写函数删除上下文参数 (#395)
|
||||
04. 弃用 `np.bool` 类型别名 (#396)
|
||||
05. 为自定义数据集添加新的视频链接 (#365)
|
||||
06. 仅为模型导出 onnx (#361)
|
||||
07. 添加 MMYOLO 回归测试 yml (#359)
|
||||
08. 更新 `article.md` 中的视频教程 (#350)
|
||||
09. 添加部署 demo (#343)
|
||||
10. 优化 debug 模式下大图的可视化效果(#346)
|
||||
11. 改进 `browse_dataset` 的参数并支持 `RepeatDataset` (#340, #338)
|
||||
|
||||
### 视频
|
||||
|
||||
1. 发布了 [基于 sahi 的大图推理](https://www.bilibili.com/video/BV1EK411R7Ws/)
|
||||
2. 发布了 [自定义数据集从标注到部署保姆级教程](https://www.bilibili.com/video/BV1RG4y137i5)
|
||||
|
||||
### 贡献者
|
||||
|
||||
总共 28 位开发者参与了本次版本
|
||||
|
||||
谢谢 @RangeKing, @PeterH0323, @Nioolek, @triple-Mu, @matrixgame2018, @xin-li-67, @tang576225574, @kitecats, @Seperendity, @diplomatist, @vaew, @wzr-skn, @VoyagerXvoyagerx, @MambaWong, @tianleiSHI, @caj-github, @zhubochao, @lvhan028, @dsghaonan, @lyviva, @yuewangg, @wang-tf, @satuoqaq, @grimoire, @RunningLeon, @hanrui1sensetime, @RangiLyu, @hhaAndroid
|
||||
|
||||
## v0.2.0（1/12/2022)
|
||||
|
||||
### 亮点
|
||||
|
|
|
|||
|
|
@ -51,8 +51,9 @@ MMYOLO 文件结构和 MMDetection 完全一致。为了能够充分复用 MMDet
|
|||
|
||||
5. 参考以下教程深入了解：
|
||||
|
||||
- [数据流](https://mmyolo.readthedocs.io/zh_CN/latest/advanced_guides/index.html#id1)
|
||||
- [模块组合](https://mmyolo.readthedocs.io/zh_CN/latest/advanced_guides/index.html#id1)
|
||||
- [数据流](https://mmyolo.readthedocs.io/zh_CN/latest/advanced_guides/index.html#id2)
|
||||
- [How to](https://mmyolo.readthedocs.io/zh_CN/latest/advanced_guides/index.html#how-to)
|
||||
- [插件](https://mmyolo.readthedocs.io/zh_CN/latest/advanced_guides/index.html#id3)
|
||||
- [插件](https://mmyolo.readthedocs.io/zh_CN/latest/advanced_guides/index.html#id4)
|
||||
|
||||
6. [解读文章和资源汇总](article.md)
|
||||
|
|
|
|||
|
|
@ -12,7 +12,7 @@ conda install pytorch torchvision -c pytorch
|
|||
pip install -U openmim
|
||||
mim install "mmengine>=0.3.1"
|
||||
mim install "mmcv>=2.0.0rc1,<2.1.0"
|
||||
mim install "mmdet>=3.0.0rc3,<3.1.0"
|
||||
mim install "mmdet>=3.0.0rc5,<3.1.0"
|
||||
git clone https://github.com/open-mmlab/mmyolo.git
|
||||
cd mmyolo
|
||||
# Install albumentations
|
||||
|
|
|
|||
|
|
@ -14,7 +14,7 @@ mmengine_minimum_version = '0.3.1'
|
|||
mmengine_maximum_version = '1.0.0'
|
||||
mmengine_version = digit_version(mmengine.__version__)
|
||||
|
||||
mmdet_minimum_version = '3.0.0rc3'
|
||||
mmdet_minimum_version = '3.0.0rc5'
|
||||
mmdet_maximum_version = '3.1.0'
|
||||
mmdet_version = digit_version(mmdet.__version__)
|
||||
|
||||
|
|
|
|||
|
|
@ -254,6 +254,9 @@ class BatchYOLOv7Assigner(nn.Module):
|
|||
_mlvl_decoderd_bboxes = torch.cat(_mlvl_decoderd_bboxes, dim=0)
|
||||
num_pred_positive = _mlvl_decoderd_bboxes.shape[0]
|
||||
|
||||
if num_pred_positive == 0:
|
||||
continue
|
||||
|
||||
# scaled xywh
|
||||
batch_input_shape_wh = pred_results[0].new_tensor(
|
||||
batch_input_shape[::-1]).repeat((1, 2))
|
||||
|
|
|
|||
|
|
@ -1,6 +1,6 @@
|
|||
# Copyright (c) OpenMMLab. All rights reserved.
|
||||
|
||||
__version__ = '0.2.0'
|
||||
__version__ = '0.3.0'
|
||||
|
||||
from typing import Tuple
|
||||
|
||||
|
|
|
|||
|
|
@ -37,20 +37,51 @@ You should claim whether this is based on the pre-trained weights, which are con
|
|||
|
||||
| Method | Backbone | Pretrained Model | Training set | Test set | #epoch | box AP | Download |
|
||||
| :---------------------------------------------------------------------------: | :-------------------: | :--------------: | :------------: | :----------: | :----: | :----: | :----------------------: |
|
||||
| [YOLOv5 dummy](configs/yolov5_s_dummy-backbone_v61_syncbn_8xb16-300e_coco.py) | DummyYOLOv5CSPDarknet | - | COCO2017 Train | COCO2017 Val | 12 | 0.8853 | [model](<>) \| [log](<>) |
|
||||
| [YOLOv5 dummy](configs/yolov5_s_dummy-backbone_v61_syncbn_8xb16-300e_coco.py) | DummyYOLOv5CSPDarknet | - | COCO2017 Train | COCO2017 Val | 300 | 37.7 | [model](<>) \| [log](<>) |
|
||||
|
||||
## Citation
|
||||
|
||||
<!-- You may remove this section if not applicable. -->
|
||||
|
||||
```latex
|
||||
@article{Ren_2017,
|
||||
title={Faster R-CNN: Towards Real-Time Object Detection with Region Proposal Networks},
|
||||
journal={IEEE Transactions on Pattern Analysis and Machine Intelligence},
|
||||
publisher={Institute of Electrical and Electronics Engineers (IEEE)},
|
||||
author={Ren, Shaoqing and He, Kaiming and Girshick, Ross and Sun, Jian},
|
||||
year={2017},
|
||||
month={Jun},
|
||||
@software{glenn_jocher_2022_7002879,
|
||||
author = {Glenn Jocher and
|
||||
Ayush Chaurasia and
|
||||
Alex Stoken and
|
||||
Jirka Borovec and
|
||||
NanoCode012 and
|
||||
Yonghye Kwon and
|
||||
TaoXie and
|
||||
Kalen Michael and
|
||||
Jiacong Fang and
|
||||
imyhxy and
|
||||
Lorna and
|
||||
Colin Wong and
|
||||
曾逸夫(Zeng Yifu) and
|
||||
Abhiram V and
|
||||
Diego Montes and
|
||||
Zhiqiang Wang and
|
||||
Cristi Fati and
|
||||
Jebastin Nadar and
|
||||
Laughing and
|
||||
UnglvKitDe and
|
||||
tkianai and
|
||||
yxNONG and
|
||||
Piotr Skalski and
|
||||
Adam Hogan and
|
||||
Max Strobel and
|
||||
Mrinal Jain and
|
||||
Lorenzo Mammana and
|
||||
xylieong},
|
||||
title = {{ultralytics/yolov5: v6.2 - YOLOv5 Classification
|
||||
Models, Apple M1, Reproducibility, ClearML and
|
||||
Deci.ai integrations}},
|
||||
month = aug,
|
||||
year = 2022,
|
||||
publisher = {Zenodo},
|
||||
version = {v6.2},
|
||||
doi = {10.5281/zenodo.7002879},
|
||||
url = {https://doi.org/10.5281/zenodo.7002879}
|
||||
}
|
||||
```
|
||||
|
||||
|
|
|
|||
|
|
@ -1,4 +1,4 @@
|
|||
_base_ = ['../../../configs/yolov5/yolov5_s-v61_syncbn_8xb16-300e_coco.py']
|
||||
_base_ = '../../../configs/yolov5/yolov5_s-v61_syncbn_8xb16-300e_coco.py'
|
||||
|
||||
custom_imports = dict(imports=['projects.example_project.dummy'])
|
||||
|
||||
|
|
|
|||
|
|
@ -0,0 +1,78 @@
|
|||
_base_ = '../yolov7/yolov7_tiny_syncbn_fast_8x16b-300e_coco.py'
|
||||
|
||||
max_epochs = 100
|
||||
data_root = './data/cat/'
|
||||
|
||||
work_dir = './work_dirs/yolov7_tiny_syncbn_fast_1xb32-100e_cat'
|
||||
|
||||
load_from = 'https://download.openmmlab.com/mmyolo/v0/yolov7/yolov7_tiny_syncbn_fast_8x16b-300e_coco/yolov7_tiny_syncbn_fast_8x16b-300e_coco_20221126_102719-0ee5bbdf.pth' # noqa
|
||||
|
||||
train_batch_size_per_gpu = 32
|
||||
train_num_workers = 4 # train_num_workers = nGPU x 4
|
||||
|
||||
save_epoch_intervals = 2
|
||||
|
||||
# base_lr_default * (your_bs / default_bs)
|
||||
base_lr = 0.01 / 4
|
||||
|
||||
anchors = [
|
||||
[(68, 69), (154, 91), (143, 162)], # P3/8
|
||||
[(242, 160), (189, 287), (391, 207)], # P4/16
|
||||
[(353, 337), (539, 341), (443, 432)] # P5/32
|
||||
]
|
||||
|
||||
class_name = ('cat', )
|
||||
num_classes = len(class_name)
|
||||
metainfo = dict(classes=class_name, palette=[(220, 20, 60)])
|
||||
|
||||
train_cfg = dict(
|
||||
max_epochs=max_epochs,
|
||||
val_begin=20,
|
||||
val_interval=save_epoch_intervals,
|
||||
dynamic_intervals=[(max_epochs - 10, 1)])
|
||||
|
||||
model = dict(
|
||||
bbox_head=dict(
|
||||
head_module=dict(num_classes=num_classes),
|
||||
prior_generator=dict(base_sizes=anchors),
|
||||
loss_cls=dict(loss_weight=0.5 *
|
||||
(num_classes / 80 * 3 / _base_.num_det_layers))))
|
||||
|
||||
train_dataloader = dict(
|
||||
batch_size=train_batch_size_per_gpu,
|
||||
num_workers=train_num_workers,
|
||||
dataset=dict(
|
||||
_delete_=True,
|
||||
type='RepeatDataset',
|
||||
times=5,
|
||||
dataset=dict(
|
||||
type=_base_.dataset_type,
|
||||
data_root=data_root,
|
||||
metainfo=metainfo,
|
||||
ann_file='annotations/trainval.json',
|
||||
data_prefix=dict(img='images/'),
|
||||
filter_cfg=dict(filter_empty_gt=False, min_size=32),
|
||||
pipeline=_base_.train_pipeline)))
|
||||
|
||||
val_dataloader = dict(
|
||||
dataset=dict(
|
||||
metainfo=metainfo,
|
||||
data_root=data_root,
|
||||
ann_file='annotations/trainval.json',
|
||||
data_prefix=dict(img='images/')))
|
||||
|
||||
test_dataloader = val_dataloader
|
||||
|
||||
val_evaluator = dict(ann_file=data_root + 'annotations/trainval.json')
|
||||
test_evaluator = val_evaluator
|
||||
|
||||
optim_wrapper = dict(optimizer=dict(lr=base_lr))
|
||||
|
||||
default_hooks = dict(
|
||||
checkpoint=dict(
|
||||
type='CheckpointHook',
|
||||
interval=save_epoch_intervals,
|
||||
max_keep_ckpts=2,
|
||||
save_best='auto'),
|
||||
param_scheduler=dict(max_epochs=max_epochs),
|
||||
logger=dict(type='LoggerHook', interval=10))
|
||||
|
|
@ -18,4 +18,4 @@ SPLIT_BEFORE_EXPRESSION_AFTER_OPENING_PAREN = true
|
|||
[codespell]
|
||||
skip = *.ipynb
|
||||
quiet-level = 3
|
||||
ignore-words-list = patten,nd,ty,mot,hist,formating,winn,gool,datas,wan,confids,TOOD,tood,ba,warmup,elease
|
||||
ignore-words-list = patten,nd,ty,mot,hist,formating,winn,gool,datas,wan,confids,tood,ba,warmup,elease,dota
|
||||
|
|
|
|||
Loading…
Reference in New Issue