[Feature] Support DDRNet #2855
[Feature] Support DDRNet #2855
Conversation
Codecov ReportPatch coverage has no change and project coverage change:
Additional details and impacted files@@ Coverage Diff @@
## dev-1.x #2855 +/- ##
===========================================
- Coverage 83.33% 82.92% -0.42%
===========================================
Files 143 158 +15
Lines 8127 9465 +1338
Branches 1211 1384 +173
===========================================
+ Hits 6773 7849 +1076
- Misses 1165 1374 +209
- Partials 189 242 +53
Flags with carried forward coverage won't be shown. Click here to find out more.
☔ View full report in Codecov by Sentry. |
|
The results look pretty good, the old pr was not merged mainly because I can not re-implement training metric. And it was not in developing schedule so I did not have time to figure it out. |
I have update the pretrained weight link, u can retry |
There was a problem hiding this comment.
Hi @whu-pzhang,
Thanks for your contribution!
I think we might add some unit tests for the backbone and head.
mmseg/models/backbones/ddrnet.py
Outdated
| Args: | ||
| in_channels | ||
| channels: | ||
| ppm_channels | ||
| align_corners | ||
| norm_cfg | ||
| act_cfg | ||
| init_cfg |
There was a problem hiding this comment.
We might complete the docstring.
|
|
||
| @MODELS.register_module() | ||
| class DDRHead(BaseDecodeHead): | ||
| """Decode head forDDRNet. |
There was a problem hiding this comment.
| """Decode head forDDRNet. | |
| """Decode head for DDRNet. |
Thanks for your contribution and we appreciate it a lot. The following instructions would make your pull request more healthy and more easily get feedback. If you do not understand some items, don't worry, just make the pull request and seek help from maintainers. ## Motivation Support DDRNet Paper: [Deep Dual-resolution Networks for Real-time and Accurate Semantic Segmentation of Road Scenes](https://arxiv.org/pdf/2101.06085) official Code: https://github.com/ydhongHIT/DDRNet There is already a PR open-mmlab#1722 , but it has been inactive for a long time. ## Current Result ### Cityscapes #### inference with converted official weights | Method | Backbone | mIoU(official) | mIoU(converted weight) | | ------ | ------------- | -------------- | ---------------------- | | DDRNet | DDRNet23-slim | 77.8 | 77.84 | | DDRNet | DDRNet23 | 79.5 | 79.53 | #### training with converted pretrained backbone | Method | Backbone | Crop Size | Lr schd | Inf time(fps) | Device | mIoU | mIoU(ms+flip) | config | download | | ------ | ------------- | --------- | ------- | ------- | -------- | ----- | ------------- | ------------ | ------------ | | DDRNet | DDRNet23-slim | 1024x1024 | 120000 | 85.85 | RTX 8000 | 77.85 | 79.80 | [config](https://github.com/whu-pzhang/mmsegmentation/blob/ddrnet/configs/ddrnet/ddrnet_23-slim_in1k-pre_2xb6-120k_cityscapes-1024x1024.py) | model \| log | | DDRNet | DDRNet23 | 1024x1024 | 120000 | 33.41 | RTX 8000 | 79.53 | 80.98 | [config](https://github.com/whu-pzhang/mmsegmentation/blob/ddrnet/configs/ddrnet/ddrnet_23_in1k-pre_2xb6-120k_cityscapes-1024x1024.py) | model \| log | The converted pretrained backbone weights download link: 1. [ddrnet23s_in1k_mmseg.pth](https://drive.google.com/file/d/1Ni4F1PMGGjuld-1S9fzDTmneLfpMuPTG/view?usp=sharing) 2. [ddrnet23_in1k_mmseg.pth](https://drive.google.com/file/d/11rsijC1xOWB6B0LgNQkAG-W6e1OdbCyJ/view?usp=sharing) ## To do - [x] support inference with converted official weights - [x] support training on cityscapes dataset --------- Co-authored-by: xiexinch <xiexinch@outlook.com>
Thanks for your contribution and we appreciate it a lot. The following instructions would make your pull request more healthy and more easily get feedback. If you do not understand some items, don't worry, just make the pull request and seek help from maintainers.
Motivation
Support DDRNet
Paper: Deep Dual-resolution Networks for Real-time and Accurate Semantic Segmentation of Road Scenes
official Code: https://github.com/ydhongHIT/DDRNet
There is already a PR #1722 , but it has been inactive for a long time.
Current Result
Cityscapes
inference with converted official weights
training with converted pretrained backbone
The converted pretrained backbone weights download link:
To do