-
-
Notifications
You must be signed in to change notification settings - Fork 106
/
Copy pathREADME.md
57 lines (40 loc) · 9.66 KB
/
README.md
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
## Introduction
<a href="https://github.com/Junjun2016/DMNet">Official Repo</a>
<a href="https://github.com/SegmentationBLWX/sssegmentation/blob/main/ssseg/modules/models/segmentors/dmnet/dmnet.py">Code Snippet</a>
<details>
<summary align="left"><a href="https://openaccess.thecvf.com/content_ICCV_2019/papers/He_Dynamic_Multi-Scale_Filters_for_Semantic_Segmentation_ICCV_2019_paper.pdf">DMNet (ICCV'2019)</a></summary>
```latex
@InProceedings{He_2019_ICCV,
author = {He, Junjun and Deng, Zhongying and Qiao, Yu},
title = {Dynamic Multi-Scale Filters for Semantic Segmentation},
booktitle = {Proceedings of the IEEE/CVF International Conference on Computer Vision (ICCV)},
month = {October},
year = {2019}
}
```
</details>
## Results
#### PASCAL VOC
| Backbone | Pretrain | Crop Size | Schedule | Train/Eval Set | mIoU | Download |
| :-: | :-: | :-: | :-: | :-: | :-: | :-: |
| R-50-D8 | ImageNet-1k-224x224 | 512x512 | LR/POLICY/BS/EPOCH: 0.01/poly/16/60 | trainaug/val | 77.38% | [cfg](https://raw.githubusercontent.com/SegmentationBLWX/sssegmentation/main/ssseg/configs/dmnet/dmnet_resnet50os8_voc.py) | [model](https://github.com/SegmentationBLWX/modelstore/releases/download/ssseg_dmnet/dmnet_resnet50os8_voc.pth) | [log](https://github.com/SegmentationBLWX/modelstore/releases/download/ssseg_dmnet/dmnet_resnet50os8_voc.log) |
| R-50-D16 | ImageNet-1k-224x224 | 512x512 | LR/POLICY/BS/EPOCH: 0.01/poly/16/60 | trainaug/val | 76.70% | [cfg](https://raw.githubusercontent.com/SegmentationBLWX/sssegmentation/main/ssseg/configs/dmnet/dmnet_resnet50os16_voc.py) | [model](https://github.com/SegmentationBLWX/modelstore/releases/download/ssseg_dmnet/dmnet_resnet50os16_voc.pth) | [log](https://github.com/SegmentationBLWX/modelstore/releases/download/ssseg_dmnet/dmnet_resnet50os16_voc.log) |
| R-101-D8 | ImageNet-1k-224x224 | 512x512 | LR/POLICY/BS/EPOCH: 0.01/poly/16/60 | trainaug/val | 79.15% | [cfg](https://raw.githubusercontent.com/SegmentationBLWX/sssegmentation/main/ssseg/configs/dmnet/dmnet_resnet101os8_voc.py) | [model](https://github.com/SegmentationBLWX/modelstore/releases/download/ssseg_dmnet/dmnet_resnet101os8_voc.pth) | [log](https://github.com/SegmentationBLWX/modelstore/releases/download/ssseg_dmnet/dmnet_resnet101os8_voc.log) |
| R-101-D16 | ImageNet-1k-224x224 | 512x512 | LR/POLICY/BS/EPOCH: 0.01/poly/16/60 | trainaug/val | 77.76% | [cfg](https://raw.githubusercontent.com/SegmentationBLWX/sssegmentation/main/ssseg/configs/dmnet/dmnet_resnet101os16_voc.py) | [model](https://github.com/SegmentationBLWX/modelstore/releases/download/ssseg_dmnet/dmnet_resnet101os16_voc.pth) | [log](https://github.com/SegmentationBLWX/modelstore/releases/download/ssseg_dmnet/dmnet_resnet101os16_voc.log) |
#### ADE20k
| Backbone | Pretrain | Crop Size | Schedule | Train/Eval Set | mIoU | Download |
| :-: | :-: | :-: | :-: | :-: | :-: | :-: |
| R-50-D8 | ImageNet-1k-224x224 | 512x512 | LR/POLICY/BS/EPOCH: 0.01/poly/16/130 | train/val | 43.54% | [cfg](https://raw.githubusercontent.com/SegmentationBLWX/sssegmentation/main/ssseg/configs/dmnet/dmnet_resnet50os8_ade20k.py) | [model](https://github.com/SegmentationBLWX/modelstore/releases/download/ssseg_dmnet/dmnet_resnet50os8_ade20k.pth) | [log](https://github.com/SegmentationBLWX/modelstore/releases/download/ssseg_dmnet/dmnet_resnet50os8_ade20k.log) |
| R-50-D16 | ImageNet-1k-224x224 | 512x512 | LR/POLICY/BS/EPOCH: 0.01/poly/16/130 | train/val | 41.43% | [cfg](https://raw.githubusercontent.com/SegmentationBLWX/sssegmentation/main/ssseg/configs/dmnet/dmnet_resnet50os16_ade20k.py) | [model](https://github.com/SegmentationBLWX/modelstore/releases/download/ssseg_dmnet/dmnet_resnet50os16_ade20k.pth) | [log](https://github.com/SegmentationBLWX/modelstore/releases/download/ssseg_dmnet/dmnet_resnet50os16_ade20k.log) |
| R-101-D8 | ImageNet-1k-224x224 | 512x512 | LR/POLICY/BS/EPOCH: 0.01/poly/16/130 | train/val | 45.53% | [cfg](https://raw.githubusercontent.com/SegmentationBLWX/sssegmentation/main/ssseg/configs/dmnet/dmnet_resnet101os8_ade20k.py) | [model](https://github.com/SegmentationBLWX/modelstore/releases/download/ssseg_dmnet/dmnet_resnet101os8_ade20k.pth) | [log](https://github.com/SegmentationBLWX/modelstore/releases/download/ssseg_dmnet/dmnet_resnet101os8_ade20k.log) |
| R-101-D16 | ImageNet-1k-224x224 | 512x512 | LR/POLICY/BS/EPOCH: 0.01/poly/16/130 | train/val | 43.53% | [cfg](https://raw.githubusercontent.com/SegmentationBLWX/sssegmentation/main/ssseg/configs/dmnet/dmnet_resnet101os16_ade20k.py) | [model](https://github.com/SegmentationBLWX/modelstore/releases/download/ssseg_dmnet/dmnet_resnet101os16_ade20k.pth) | [log](https://github.com/SegmentationBLWX/modelstore/releases/download/ssseg_dmnet/dmnet_resnet101os16_ade20k.log) |
#### CityScapes
| Backbone | Pretrain | Crop Size | Schedule | Train/Eval Set | mIoU | Download |
| :-: | :-: | :-: | :-: | :-: | :-: | :-: |
| R-50-D8 | ImageNet-1k-224x224 | 512x1024 | LR/POLICY/BS/EPOCH: 0.01/poly/8/220 | train/val | 79.17% | [cfg](https://raw.githubusercontent.com/SegmentationBLWX/sssegmentation/main/ssseg/configs/dmnet/dmnet_resnet50os8_cityscapes.py) | [model](https://github.com/SegmentationBLWX/modelstore/releases/download/ssseg_dmnet/dmnet_resnet50os8_cityscapes.pth) | [log](https://github.com/SegmentationBLWX/modelstore/releases/download/ssseg_dmnet/dmnet_resnet50os8_cityscapes.log) |
| R-50-D16 | ImageNet-1k-224x224 | 512x1024 | LR/POLICY/BS/EPOCH: 0.01/poly/8/220 | train/val | 76.43% | [cfg](https://raw.githubusercontent.com/SegmentationBLWX/sssegmentation/main/ssseg/configs/dmnet/dmnet_resnet50os16_cityscapes.py) | [model](https://github.com/SegmentationBLWX/modelstore/releases/download/ssseg_dmnet/dmnet_resnet50os16_cityscapes.pth) | [log](https://github.com/SegmentationBLWX/modelstore/releases/download/ssseg_dmnet/dmnet_resnet50os16_cityscapes.log) |
| R-101-D8 | ImageNet-1k-224x224 | 512x1024 | LR/POLICY/BS/EPOCH: 0.01/poly/8/220 | train/val | 79.90% | [cfg](https://raw.githubusercontent.com/SegmentationBLWX/sssegmentation/main/ssseg/configs/dmnet/dmnet_resnet101os8_cityscapes.py) | [model](https://github.com/SegmentationBLWX/modelstore/releases/download/ssseg_dmnet/dmnet_resnet101os8_cityscapes.pth) | [log](https://github.com/SegmentationBLWX/modelstore/releases/download/ssseg_dmnet/dmnet_resnet101os8_cityscapes.log) |
| R-101-D16 | ImageNet-1k-224x224 | 512x1024 | LR/POLICY/BS/EPOCH: 0.01/poly/8/220 | train/val | 78.09% | [cfg](https://raw.githubusercontent.com/SegmentationBLWX/sssegmentation/main/ssseg/configs/dmnet/dmnet_resnet101os16_cityscapes.py) | [model](https://github.com/SegmentationBLWX/modelstore/releases/download/ssseg_dmnet/dmnet_resnet101os16_cityscapes.pth) | [log](https://github.com/SegmentationBLWX/modelstore/releases/download/ssseg_dmnet/dmnet_resnet101os16_cityscapes.log) |
## More
You can also download the model weights from following sources:
- BaiduNetdisk: https://pan.baidu.com/s/1gD-NJJWOtaHCtB0qHE79rA with access code **s757**