Skip to content

Commit e482e3a

Browse files
hhaAndroidhhaAndroid
committed
Add changelog for v0.1.1 (#111)
* add changelog * update * update * update * update * update
1 parent 0d97ffe commit e482e3a

12 files changed

Lines changed: 116 additions & 36 deletions

File tree

‎README.md‎

Lines changed: 8 additions & 7 deletions
Original file line numberDiff line numberDiff line change
@@ -69,12 +69,10 @@ The master branch works with **PyTorch 1.6+**.
6969

7070
## What's New
7171

72-
**v0.1.0** was released on 21/9/2022:
72+
**v0.1.1** was released on 29/9/2022:
7373

74-
- Unified component interfaces based on [OpenMMLab 2.0](https://github.com/open-mmlab) and [MMDetection 3.0](https://github.com/open-mmlab/mmdetection/tree/3.x)
75-
- Support YOLOv5/YOLOX training, support YOLOv6 inference. Deployment will be supported soon.
76-
- Refactored YOLOX from MMDetection to accelerate training and inference.
77-
- Detailed introduction and advanced tutorials are provided, see the [English tutorial](https://mmyolo.readthedocs.io/en/latest).
74+
- Support [RTMDet](https://github.com/open-mmlab/mmyolo/blob/main/configs/rtmdet).
75+
- Support for backbone customization plugins and update How-to documentation.
7876

7977
For release history and update details, please refer to [changelog](https://mmyolo.readthedocs.io/en/latest/notes/changelog.html).
8078

@@ -87,8 +85,8 @@ conda create -n open-mmlab python=3.8 pytorch==1.10.1 torchvision==0.11.2 cudato
8785
conda activate open-mmlab
8886
pip install openmim
8987
mim install mmengine
90-
mim install "mmcv>=2.0.0rc1"
91-
mim install "mmdet>=3.0.0rc0"
88+
mim install "mmcv>=2.0.0rc1,<2.1.0"
89+
mim install "mmdet>=3.0.0rc1,<3.1.0"
9290
git clone https://github.com/open-mmlab/mmyolo.git
9391
cd mmyolo
9492
# Install albumentations
@@ -135,6 +133,7 @@ Results and models are available in the [model zoo](docs/en/model_zoo.md).
135133

136134
- [x] [YOLOv5](configs/yolov5)
137135
- [x] [YOLOX](configs/yolox)
136+
- [x] [RTMDet](configs/rtmdet)
138137
- [ ] [YOLOv6](configs/yolov6)(Inference only)
139138

140139
</details>
@@ -165,13 +164,15 @@ Results and models are available in the [model zoo](docs/en/model_zoo.md).
165164
<li>YOLOv5CSPDarknet</li>
166165
<li>YOLOXCSPDarknet</li>
167166
<li>EfficientRep</li>
167+
<li>CSPNeXt</li>
168168
</ul>
169169
</td>
170170
<td>
171171
<ul>
172172
<li>YOLOv5PAFPN</li>
173173
<li>YOLOv6RepPAFPN</li>
174174
<li>YOLOXPAFPN</li>
175+
<li>CSPNeXtPAFPN</li>
175176
</ul>
176177
</td>
177178
<td>

‎README_zh-CN.md‎

Lines changed: 9 additions & 7 deletions
Original file line numberDiff line numberDiff line change
@@ -69,12 +69,11 @@ MMYOLO 是一个基于 PyTorch 和 MMDetection 的 YOLO 系列算法开源工具
6969

7070
## 最新进展
7171

72-
**v0.1.0** 版本已经在 2022.9.21 发布：
72+
**v0.1.1** 版本已经在 2022.9.29 发布：
7373

74-
- 基于 [OpenMMLab 2.0](https://github.com/open-mmlab) 和 [MMDetection 3.0](https://github.com/open-mmlab/mmdetection/tree/3.x) 统一了各组件接口。
75-
- 支持 YOLOv5/YOLOX 训练，支持 YOLOv6 推理。即将支持部署。
76-
- 重构了 MMDetection 的 YOLOX，提供了更快的训练和推理速度。
77-
- 提供了详细入门和进阶教程，详见 [中文教程](https://mmyolo.readthedocs.io/zh_CN/latest)。
74+
- 支持了 [RTMDet](https://github.com/open-mmlab/mmyolo/blob/main/configs/rtmdet)
75+
- 新增了 [RTMDet 原理和实现全解析中文文档](https://github.com/open-mmlab/mmyolo/blob/main/docs/zh_cn/algorithm_descriptions/rtmdet_description.md)
76+
- 支持对 backbone 自定义插件，并且更新了 How-to 文档
7877

7978
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)
8079

@@ -87,8 +86,8 @@ conda create -n open-mmlab python=3.8 pytorch==1.10.1 torchvision==0.11.2 cudato
8786
conda activate open-mmlab
8887
pip install openmim
8988
mim install mmengine
90-
mim install "mmcv>=2.0.0rc1"
91-
mim install "mmdet>=3.0.0rc0"
89+
mim install "mmcv>=2.0.0rc1,<2.1.0"
90+
mim install "mmdet>=3.0.0rc1,<3.1.0"
9291
git clone https://github.com/open-mmlab/mmyolo.git
9392
cd mmyolo
9493
# Install albumentations
@@ -136,6 +135,7 @@ MMYOLO 用法和 MMDetection 几乎一致，所有教程都是通用的，你也
136135

137136
- [x] [YOLOv5](configs/yolov5)
138137
- [x] [YOLOX](configs/yolox)
138+
- [x] [RTMDet](configs/rtmdet)
139139
- [ ] [YOLOv6](configs/yolov6)(仅推理)
140140

141141
</details>
@@ -166,13 +166,15 @@ MMYOLO 用法和 MMDetection 几乎一致，所有教程都是通用的，你也
166166
<li>YOLOv5CSPDarknet</li>
167167
<li>YOLOXCSPDarknet</li>
168168
<li>EfficientRep</li>
169+
<li>CSPNeXt</li>
169170
</ul>
170171
</td>
171172
<td>
172173
<ul>
173174
<li>YOLOv5PAFPN</li>
174175
<li>YOLOv6RepPAFPN</li>
175176
<li>YOLOXPAFPN</li>
177+
<li>CSPNeXtPAFPN</li>
176178
</ul>
177179
</td>
178180
<td>

‎docker/Dockerfile‎

Lines changed: 1 addition & 1 deletion
Original file line numberDiff line numberDiff line change
@@ -26,7 +26,7 @@ RUN apt-get update \
2626

2727
# Install MMEngine , MMCV and MMDet
2828
RUN pip install --no-cache-dir openmim && \
29-
mim install --no-cache-dir mmengine "mmcv>=2.0.0rc1" "mmdet>=3.0.0rc0"
29+
mim install --no-cache-dir mmengine "mmcv>=2.0.0rc1,<2.1.0" "mmdet>=3.0.0rc1,<3.1.0"
3030

3131
# Install MMYOLO
3232
RUN git clone https://github.com/open-mmlab/mmyolo.git /mmyolo && \

‎docs/en/get_started.md‎

Lines changed: 11 additions & 3 deletions
Original file line numberDiff line numberDiff line change
@@ -2,6 +2,14 @@
22

33
## Prerequisites
44

5+
Compatible MMEngine, MMCV and MMDetection versions are shown as below. Please install the correct version to avoid installation issues.
6+
7+
| MMYOLO version | MMDetection version | MMEngine version | MMCV version |
8+
| :------------: | :----------------------: | :----------------------: | :---------------------: |
9+
| main | mmdet>=3.0.0rc1, \<3.1.0 | mmengine>=0.1.0, \<0.2.0 | mmcv>=2.0.0rc0, \<2.1.0 |
10+
| 0.1.1 | mmdet>=3.0.0rc1, \<3.1.0 | mmengine>=0.1.0, \<0.2.0 | mmcv>=2.0.0rc0, \<2.1.0 |
11+
| 0.1.0 | mmdet>=3.0.0rc0, \<3.1.0 | mmengine>=0.1.0, \<0.2.0 | mmcv>=2.0.0rc0, \<2.1.0 |
12+
513
In this section, we demonstrate how to prepare an environment with PyTorch.
614

715
MMDetection works on Linux, Windows, and macOS. It requires Python 3.6+, CUDA 9.2+, and PyTorch 1.7+.
@@ -42,8 +50,8 @@ conda install pytorch torchvision cpuonly -c pytorch
4250
```shell
4351
pip install -U openmim
4452
mim install mmengine
45-
mim install "mmcv>=2.0.0rc1"
46-
mim install "mmdet>=3.0.0rc0"
53+
mim install "mmcv>=2.0.0rc1,<2.1.0"
54+
mim install "mmdet>=3.0.0rc1,<3.1.0"
4755
```
4856

4957
**Note:**
@@ -188,7 +196,7 @@ thus we only need to install MMEngine, MMCV, MMDetection, and MMYOLO with the fo
188196
!pip3 install openmim
189197
!mim install mmengine
190198
!mim install "mmcv>=2.0.0rc1,<2.1.0"
191-
!mim install "mmdet>=3.0.0.rc0"
199+
!mim install "mmdet>=3.0.0.rc1"
192200
```
193201

194202
**Step 2.** Install MMYOLO from the source.

‎docs/en/model_zoo.md‎

Lines changed: 4 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -18,3 +18,7 @@ Please refer to [YOLOv6](https://github.com/open-mmlab/mmyolo/blob/main/configs/
1818
### YOLOX
1919

2020
Please refer to [YOLOX](https://github.com/open-mmlab/mmyolo/blob/main/configs/yolox)。
21+
22+
### RTMDet
23+
24+
Please refer to [RTMDet](https://github.com/open-mmlab/mmyolo/blob/main/configs/rtmdet)。

‎docs/en/notes/changelog.md‎

Lines changed: 27 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -1,5 +1,32 @@
11
# Changelog
22

3+
## v0.1.1（29/9/2022)
4+
5+
Based on MMDetection's RTMDet high precision and low latency object detection algorithm, we have also released RTMDet and provided a Chinese document on the principle and implementation of RTMDet.
6+
7+
### Highlights
8+
9+
1. Support [RTMDet](https://github.com/open-mmlab/mmyolo/blob/main/configs/rtmdet)
10+
2. Support for backbone customization plugins and update How-to documentation (#75)
11+
12+
### Bug Fixes
13+
14+
1. Fix some documentation errors (#66, #72, #76, #83, #86)
15+
2. Fix checkpoints link error (#63)
16+
3. Fix the bug that the output of `LetterResize` does not meet the expectation when using `imscale` (#105)
17+
18+
### Improvements
19+
20+
1. Reducing the size of docker images (#67)
21+
2. Simplifying `Compose` Logic in `BaseMixImageTransform` (#71)
22+
3. Supports dump results in `test.py` (#84)
23+
24+
#### Contributors
25+
26+
A total of 13 developers contributed to this release.
27+
28+
Thank @wanghonglie, @hhaAndroid, @yang-0201, @PeterH0323, @RangeKing, @satuoqaq, @Zheng-LinXiao, @xin-li-67, @suibe-qingtian, @MambaWong, @MichaelCai0912, @rimoire, @Nioolek
29+
330
## v0.1.0（21/9/2022)
431

532
We have released MMYOLO open source library, which is based on MMEngine, MMCV 2.x and MMDetection 3.x libraries. At present, the object detection has been realized, and it will be expanded to multi-task in the future.

‎docs/zh_cn/algorithm_descriptions/rtmdet_description.md‎

Lines changed: 8 additions & 10 deletions
Original file line numberDiff line numberDiff line change
@@ -76,10 +76,9 @@ MMDetection 开源库中已经对单图数据增强进行了封装，用户通
7676
| MixUp | | 19.3 |
7777
| MixUp | √ | **12.4** |
7878

79-
| | RTMDet-s | RTMDet-l |
80-
| --------------------------------------------- | -------- | -------- |
81-
| Mosaic + MixUp + 20e finetune | 43.9 | **51.3** |
82-
| **Small-cache** Mosaic + MixUp + 20e finetune | **44.2** | 51.1 |
79+
| | RTMDet-s | RTMDet-l |
80+
| ----------------------------- | -------- | -------- |
81+
| Mosaic + MixUp + 20e finetune | 43.9 | **51.3** |
8382

8483
#### 1.1.1 为图像混合数据增强引入 Cache
8584

@@ -180,12 +179,11 @@ Mosaic+MixUp 失真度比较高，持续用太强的数据增强对模型并不
180179

181180
为了使数据增强的方式更为通用，RTMDet 在前 280 epoch 使用不带旋转的 Mosaic+MixUp, 且通过混入 8 张图片来提升强度以及正样本数。后 20 epoch 使用比较小的学习率在比较弱的增强下进行微调，同时在 EMA 的作用下将参数缓慢更新至模型，能够得到比较大的提升。
182181

183-
| | RTMDet-s | RTMDet-l |
184-
| ----------------------------------------- | -------- | -------- |
185-
| LSJ + rand crop | 42.3 | 46.7 |
186-
| Mosaic+MixUp | 41.9 | 49.8 |
187-
| Mosaic + MixUp + 20e finetune | 43.9 | **51.3** |
188-
| Small-cache Mosaic + MixUp + 20e finetune | **44.2** | 51.1 |
182+
| | RTMDet-s | RTMDet-l |
183+
| ----------------------------- | -------- | -------- |
184+
| LSJ + rand crop | 42.3 | 46.7 |
185+
| Mosaic+MixUp | 41.9 | 49.8 |
186+
| Mosaic + MixUp + 20e finetune | 43.9 | **51.3** |
189187

190188
### 1.2 模型结构
191189

‎docs/zh_cn/get_started.md‎

Lines changed: 11 additions & 3 deletions
Original file line numberDiff line numberDiff line change
@@ -2,6 +2,14 @@
22

33
## 依赖
44

5+
下表为 MMYOLO 和 MMEngine, MMCV, MMDetection 依赖库的版本要求，请安装正确的版本以避免安装问题。
6+
7+
| MMYOLO version || MMDetection version | MMEngine version | MMCV version |
8+
|:--------------:|:-------------------:|:-------------------:|:-------------------:|
9+
| main | mmdet>=3.0.0rc1, \<3.1.0 | mmengine>=0.1.0, \<0.2.0 | mmcv>=2.0.0rc0, \<2.1.0|
10+
| 0.1.1 | mmdet>=3.0.0rc1, \<3.1.0 | mmengine>=0.1.0, \<0.2.0 |mmcv>=2.0.0rc0, \<2.1.0|
11+
| 0.1.0 | mmdet>=3.0.0rc0, \<3.1.0 | mmengine>=0.1.0, \<0.2.0 |mmcv>=2.0.0rc0, \<2.1.0|
12+
513
本节中，我们将演示如何用 PyTorch 准备一个环境。
614

715
MMYOLO 支持在 Linux，Windows 和 macOS 上运行。它需要 Python 3.6 以上，CUDA 9.2 以上和 PyTorch 1.7 以上。
@@ -42,8 +50,8 @@ conda install pytorch torchvision cpuonly -c pytorch
4250
```shell
4351
pip install -U openmim
4452
mim install mmengine
45-
mim install "mmcv>=2.0.0rc1"
46-
mim install "mmdet>=3.0.0rc0"
53+
mim install "mmcv>=2.0.0rc1,<2.1.0"
54+
mim install "mmdet>=3.0.0rc1,<3.1.0"
4755
```
4856

4957
**注意：**
@@ -193,7 +201,7 @@ pip install "mmcv>=2.0.0rc1" -f https://download.openmmlab.com/mmcv/dist/cu116/t
193201
!pip3 install openmim
194202
!mim install mmengine
195203
!mim install "mmcv>=2.0.0rc1,<2.1.0"
196-
!mim install "mmdet>=3.0.0.rc0"
204+
!mim install "mmdet>=3.0.0.rc1"
197205
```
198206

199207
**步骤 2.** 使用源码安装 MMYOLO：

‎docs/zh_cn/model_zoo.md‎

Lines changed: 4 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -18,3 +18,7 @@
1818
### YOLOX
1919

2020
请参考 [YOLOX](https://github.com/open-mmlab/mmyolo/blob/main/configs/yolox)。
21+
22+
### RTMDet
23+
24+
请参考 [RTMDet](https://github.com/open-mmlab/mmyolo/blob/main/configs/rtmdet)。

‎docs/zh_cn/notes/changelog.md‎

Lines changed: 28 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -1,5 +1,33 @@
11
# 更新日志
22

3+
## v0.1.1（29/9/2022)
4+
5+
基于 MMDetection 的 RTMDet 高精度低延时目标检测算法，我们也同步发布了 RTMDet，并提供了 RTMDet 原理和实现全解析中文文档
6+
7+
### 亮点
8+
9+
1. 支持了 [RTMDet](https://github.com/open-mmlab/mmyolo/blob/main/configs/rtmdet)
10+
2. 新增了 [RTMDet 原理和实现全解析中文文档](https://github.com/open-mmlab/mmyolo/blob/main/docs/zh_cn/algorithm_descriptions/rtmdet_description.md)
11+
3. 支持对 backbone 自定义插件，并更新了 How-to 文档 (#75)
12+
13+
### Bug 修复
14+
15+
1. 修复一些文档错误 (#66, #72, #76, #83, #86)
16+
2. 修复权重链接错误 (#63)
17+
3. 修复 `LetterResize` 使用 `imscale` api 时候输出不符合预期的 bug (#105)
18+
19+
### 完善
20+
21+
1. 缩减 docker 镜像尺寸 (#67)
22+
2. 简化 BaseMixImageTransform 中 Compose 逻辑 (#71)
23+
3. test 脚本支持 dump 结果 (#84)
24+
25+
#### 贡献者
26+
27+
总共 13 位开发者参与了本次版本
28+
29+
谢谢 @wanghonglie, @hhaAndroid, @yang-0201, @PeterH0323, @RangeKing, @satuoqaq, @Zheng-LinXiao, @xin-li-67, @suibe-qingtian, @MambaWong, @MichaelCai0912, @rimoire, @Nioolek
30+
331
## v0.1.0（21/9/2022)
432

533
我们发布了 MMYOLO 开源库，其基于 MMEngine, MMCV 2.x 和 MMDetection 3.x 库. 目前实现了目标检测功能，后续会扩展为多任务。

0 commit comments

Comments
 (0)