Skip to content

Commit 5970e09

Browse files
chhluochhluo
authored
Add changelog of v2.25.0 (#8071)
* bump to v2.25.0 * update * update * update * update date * update * reset * reset * reset * reset * reset
1 parent 51ad7cb commit 5970e09

File tree

7 files changed

+179
-3
lines changed

7 files changed

+179
-3
lines changed

‎docker/serve/Dockerfile‎

Lines changed: 1 addition & 1 deletion
Original file line numberDiff line numberDiff line change
@@ -4,7 +4,7 @@ ARG CUDNN="7"
44
FROM pytorch/pytorch:${PYTORCH}-cuda${CUDA}-cudnn${CUDNN}-devel
55

66
ARG MMCV="1.3.17"
7-
ARG MMDET="2.24.1"
7+
ARG MMDET="2.25.0"
88

99
ENV PYTHONUNBUFFERED TRUE
1010

‎docs/en/changelog.md‎

Lines changed: 115 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -1,5 +1,120 @@
11
## Changelog
22

3+
### v2.25.0 (31/5/2022)
4+
5+
#### Highlights
6+
7+
- Support dedicated `WandbLogger` hook
8+
- Support [ConvNeXt](configs/convnext), [DDOD](configs/ddod), [SOLOv2](configs/solov2)
9+
- Support [Mask2Former](configs/mask2former) for instance segmentation
10+
- Rename [config files of Mask2Former](configs/mask2former)
11+
12+
#### Backwards incompatible changes
13+
14+
- Rename [config files of Mask2Former](configs/mask2former) (#7571)
15+
16+
<table align="center">
17+
<thead>
18+
<tr align='center'>
19+
<td>before v2.25.0</td>
20+
<td>after v2.25.0</td>
21+
</tr>
22+
</thead>
23+
<tbody><tr valign='top'>
24+
<th>
25+
26+
- `mask2former_xxx_coco.py` represents config files for **panoptic segmentation**.
27+
28+
</th>
29+
<th>
30+
31+
- `mask2former_xxx_coco.py` represents config files for **instance segmentation**.
32+
- `mask2former_xxx_coco-panoptic.py` represents config files for **panoptic segmentation**.
33+
34+
</th></tr>
35+
</tbody></table>
36+
37+
#### New Features
38+
39+
- Support [ConvNeXt](https://arxiv.org/abs/2201.03545) (#7281)
40+
- Support [DDOD](https://arxiv.org/abs/2107.02963) (#7279)
41+
- Support [SOLOv2](https://arxiv.org/abs/2003.10152) (#7441)
42+
- Support [Mask2Former](https://arxiv.org/abs/2112.01527) for instance segmentation (#7571, #8032)
43+
44+
#### Bug Fixes
45+
46+
- Enable YOLOX training on different devices (#7912)
47+
- Fix the log plot error when evaluation with `interval != 1` (#7784)
48+
- Fix RuntimeError of HTC (#8083)
49+
50+
#### Improvements
51+
52+
- Support dedicated `WandbLogger` hook (#7459)
53+
54+
Users can set
55+
56+
```python
57+
cfg.log_config.hooks = [
58+
dict(type='MMDetWandbHook',
59+
init_kwargs={'project': 'MMDetection-tutorial'},
60+
interval=10,
61+
log_checkpoint=True,
62+
log_checkpoint_metadata=True,
63+
num_eval_images=10)]
64+
```
65+
66+
in the config to use `MMDetWandbHook`. Example can be found in this [colab tutorial](https://colab.research.google.com/drive/1RCSXHZwDZvakFh3eo9RuNrJbCGqD0dru?usp=sharing#scrollTo=WTEdPDRaBz2C)
67+
68+
- Add `AvoidOOM` to avoid OOM (#7434, #8091)
69+
70+
Try to use `AvoidCUDAOOM` to avoid GPU out of memory. It will first retry after calling `torch.cuda.empty_cache()`. If it still fails, it will then retry by converting the type of inputs to FP16 format. If it still fails, it will try to copy inputs from GPUs to CPUs to continue computing. Try AvoidOOM in code to make the code continue to run when GPU memory runs out:
71+
72+
```python
73+
from mmdet.utils import AvoidCUDAOOM
74+
75+
output = AvoidCUDAOOM.retry_if_cuda_oom(some_function)(input1, input2)
76+
```
77+
78+
Users can also try `AvoidCUDAOOM` as a decorator to make the code continue to run when GPU memory runs out:
79+
80+
```python
81+
from mmdet.utils import AvoidCUDAOOM
82+
83+
@AvoidCUDAOOM.retry_if_cuda_oom
84+
def function(*args, **kwargs):
85+
...
86+
return xxx
87+
```
88+
89+
- Support reading `gpu_collect` from `cfg.evaluation.gpu_collect` (#7672)
90+
91+
- Speedup the Video Inference by Accelerating data-loading Stage (#7832)
92+
93+
- Support replacing the `${key}` with the value of `cfg.key` (#7492)
94+
95+
- Accelerate result analysis in `analyze_result.py`. The evaluation time is speedup by 10 ~ 15 times and only tasks 10 ~ 15 minutes now. (#7891)
96+
97+
- Support to set `block_dilations` in `DilatedEncoder` (#7812)
98+
99+
- Support panoptic segmentation result analysis (#7922)
100+
101+
- Release DyHead with Swin-Large backbone (#7733)
102+
103+
- Documentations updating and adding
104+
105+
- Fix wrong default type of `act_cfg` in `SwinTransformer` (#7794)
106+
- Fix text errors in the tutorials (#7959)
107+
- Rewrite the [installation guide](docs/en/get_started.md) (#7897)
108+
- [Useful hooks](docs/en/tutorials/useful_hooks.md) (#7810)
109+
- Fix heading anchor in documentation (#8006)
110+
- Replace `markdownlint` with `mdformat` for avoiding installing ruby (#8009)
111+
112+
#### Contributors
113+
114+
A total of 20 developers contributed to this release.
115+
116+
Thanks @ZwwWayne, @DarthThomas, @solyaH, @LutingWang, @chenxinfeng4, @Czm369, @Chenastron, @chhluo, @austinmw, @Shanyaliux @hellock, @Y-M-Y, @jbwang1997, @hhaAndroid, @Irvingao, @zhanggefan, @BIGWangYuDong, @Keiku, @PeterVennerstrom, @ayulockin
117+
3118
### v2.24.0 (26/4/2022)
4119

5120
#### Highlights

‎docs/en/compatibility.md‎

Lines changed: 29 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -1,5 +1,34 @@
11
# Compatibility of MMDetection 2.x
22

3+
## MMDetection 2.25.0
4+
5+
In order to support Mask2Former for instance segmentation, the original config files of Mask2Former for panpotic segmentation need to be renamed [PR #7571](https://github.com/open-mmlab/mmdetection/pull/7571).
6+
7+
<table align="center">
8+
<thead>
9+
<tr align='center'>
10+
<td>before v2.25.0</td>
11+
<td>after v2.25.0</td>
12+
</tr>
13+
</thead>
14+
<tbody><tr valign='top'>
15+
<th>
16+
17+
```
18+
'mask2former_xxx_coco.py' represents config files for **panoptic segmentation**.
19+
```
20+
21+
</th>
22+
<th>
23+
24+
```
25+
'mask2former_xxx_coco.py' represents config files for **instance segmentation**.
26+
'mask2former_xxx_coco-panoptic.py' represents config files for **panoptic segmentation**.
27+
```
28+
29+
</th></tr>
30+
</tbody></table>
31+
332
## MMDetection 2.21.0
433

534
In order to support CPU training, the logic of scatter in batch collating has been changed. We recommend to use

‎docs/en/faq.md‎

Lines changed: 2 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -11,6 +11,8 @@ We list some common troubles faced by many users and their corresponding solutio
1111
| MMDetection version | MMCV version |
1212
| :-----------------: | :------------------------: |
1313
| master | mmcv-full>=1.3.17, \<1.6.0 |
14+
| 2.25.0 | mmcv-full>=1.3.17, \<1.6.0 |
15+
| 2.24.1 | mmcv-full>=1.3.17, \<1.6.0 |
1416
| 2.24.0 | mmcv-full>=1.3.17, \<1.6.0 |
1517
| 2.23.0 | mmcv-full>=1.3.17, \<1.5.0 |
1618
| 2.22.0 | mmcv-full>=1.3.17, \<1.5.0 |

‎docs/zh_cn/compatibility.md‎

Lines changed: 30 additions & 1 deletion
Original file line numberDiff line numberDiff line change
@@ -1,6 +1,35 @@
11
# MMDetection v2.x 兼容性说明
22

3-
## MMDection 2.21.0
3+
## MMDetection 2.25.0
4+
5+
为了加入 Mask2Former 实例分割模型，对 Mask2Former 的配置文件进行了重命名 [PR #7571](https://github.com/open-mmlab/mmdetection/pull/7571)：
6+
7+
<table align="center">
8+
<thead>
9+
<tr align='center'>
10+
<td>在 v2.25.0 之前</td>
11+
<td>v2.25.0 及之后</td>
12+
</tr>
13+
</thead>
14+
<tbody><tr valign='top'>
15+
<th>
16+
17+
```
18+
'mask2former_xxx_coco.py' 代表全景分割的配置文件
19+
```
20+
21+
</th>
22+
<th>
23+
24+
```
25+
'mask2former_xxx_coco.py' 代表实例分割的配置文件
26+
'mask2former_xxx_coco-panoptic.py' 代表全景分割的配置文件
27+
```
28+
29+
</th></tr>
30+
</tbody></table>
31+
32+
## MMDetection 2.21.0
433

534
为了支持 CPU 训练，MMCV 中进行批处理的 scatter 的代码逻辑已经被修改。我们推荐使用 MMCV v1.4.4 或更高版本，
635
更多信息请参考 [MMCV PR #1621](https://github.com/open-mmlab/mmcv/pull/1621).

‎docs/zh_cn/get_started.md‎

Lines changed: 1 addition & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -12,6 +12,7 @@ MMDetection 和 MMCV 版本兼容性如下所示，需要安装正确的 MMCV
1212
| MMDetection 版本 | MMCV 版本 |
1313
| :------------: | :------------------------: |
1414
| master | mmcv-full>=1.3.17, \<1.6.0 |
15+
| 2.25.0 | mmcv-full>=1.3.17, \<1.6.0 |
1516
| 2.24.1 | mmcv-full>=1.3.17, \<1.6.0 |
1617
| 2.24.0 | mmcv-full>=1.3.17, \<1.6.0 |
1718
| 2.23.0 | mmcv-full>=1.3.17, \<1.5.0 |

‎mmdet/version.py‎

Lines changed: 1 addition & 1 deletion
Original file line numberDiff line numberDiff line change
@@ -1,6 +1,6 @@
11
# Copyright (c) OpenMMLab. All rights reserved.
22

3-
__version__ = '2.24.1'
3+
__version__ = '2.25.0'
44
short_version = __version__
55

66

0 commit comments

Comments
 (0)