-
-
Notifications
You must be signed in to change notification settings - Fork 108
/
README.md
56 lines (39 loc) · 9.54 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
## Introduction
<a href="https://github.com/junfu1115/DANet/">Official Repo</a>
<a href="https://github.com/SegmentationBLWX/sssegmentation/blob/main/ssseg/modules/models/segmentors/danet/danet.py">Code Snippet</a>
<details>
<summary align="left"><a href="https://arxiv.org/pdf/1809.02983.pdf">DANet (CVPR'2019)</a></summary>
```latex
@article{fu2018dual,
title={Dual Attention Network for Scene Segmentation},
author={Jun Fu, Jing Liu, Haijie Tian, Yong Li, Yongjun Bao, Zhiwei Fang,and Hanqing Lu},
booktitle={The IEEE Conference on Computer Vision and Pattern Recognition (CVPR)},
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 | 76.39% | [cfg](https://raw.githubusercontent.com/SegmentationBLWX/sssegmentation/main/ssseg/configs/danet/danet_resnet50os8_voc.py) | [model](https://github.com/SegmentationBLWX/modelstore/releases/download/ssseg_danet/danet_resnet50os8_voc.pth) | [log](https://github.com/SegmentationBLWX/modelstore/releases/download/ssseg_danet/danet_resnet50os8_voc.log) |
| R-50-D16 | ImageNet-1k-224x224 | 512x512 | LR/POLICY/BS/EPOCH: 0.01/poly/16/60 | trainaug/val | 75.04% | [cfg](https://raw.githubusercontent.com/SegmentationBLWX/sssegmentation/main/ssseg/configs/danet/danet_resnet50os16_voc.py) | [model](https://github.com/SegmentationBLWX/modelstore/releases/download/ssseg_danet/danet_resnet50os16_voc.pth) | [log](https://github.com/SegmentationBLWX/modelstore/releases/download/ssseg_danet/danet_resnet50os16_voc.log) |
| R-101-D8 | ImageNet-1k-224x224 | 512x512 | LR/POLICY/BS/EPOCH: 0.01/poly/16/60 | trainaug/val | 77.97% | [cfg](https://raw.githubusercontent.com/SegmentationBLWX/sssegmentation/main/ssseg/configs/danet/danet_resnet101os8_voc.py) | [model](https://github.com/SegmentationBLWX/modelstore/releases/download/ssseg_danet/danet_resnet101os8_voc.pth) | [log](https://github.com/SegmentationBLWX/modelstore/releases/download/ssseg_danet/danet_resnet101os8_voc.log) |
| R-101-D16 | ImageNet-1k-224x224 | 512x512 | LR/POLICY/BS/EPOCH: 0.01/poly/16/60 | trainaug/val | 76.99% | [cfg](https://raw.githubusercontent.com/SegmentationBLWX/sssegmentation/main/ssseg/configs/danet/danet_resnet101os16_voc.py) | [model](https://github.com/SegmentationBLWX/modelstore/releases/download/ssseg_danet/danet_resnet101os16_voc.pth) | [log](https://github.com/SegmentationBLWX/modelstore/releases/download/ssseg_danet/danet_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 | 42.90% | [cfg](https://raw.githubusercontent.com/SegmentationBLWX/sssegmentation/main/ssseg/configs/danet/danet_resnet50os8_ade20k.py) | [model](https://github.com/SegmentationBLWX/modelstore/releases/download/ssseg_danet/danet_resnet50os8_ade20k.pth) | [log](https://github.com/SegmentationBLWX/modelstore/releases/download/ssseg_danet/danet_resnet50os8_ade20k.log) |
| R-50-D16 | ImageNet-1k-224x224 | 512x512 | LR/POLICY/BS/EPOCH: 0.01/poly/16/130 | train/val | 40.85% | [cfg](https://raw.githubusercontent.com/SegmentationBLWX/sssegmentation/main/ssseg/configs/danet/danet_resnet50os16_ade20k.py) | [model](https://github.com/SegmentationBLWX/modelstore/releases/download/ssseg_danet/danet_resnet50os16_ade20k.pth) | [log](https://github.com/SegmentationBLWX/modelstore/releases/download/ssseg_danet/danet_resnet50os16_ade20k.log) |
| R-101-D8 | ImageNet-1k-224x224 | 512x512 | LR/POLICY/BS/EPOCH: 0.01/poly/16/130 | train/val | 44.37% | [cfg](https://raw.githubusercontent.com/SegmentationBLWX/sssegmentation/main/ssseg/configs/danet/danet_resnet101os8_ade20k.py) | [model](https://github.com/SegmentationBLWX/modelstore/releases/download/ssseg_danet/danet_resnet101os8_ade20k.pth) | [log](https://github.com/SegmentationBLWX/modelstore/releases/download/ssseg_danet/danet_resnet101os8_ade20k.log) |
| R-101-D16 | ImageNet-1k-224x224 | 512x512 | LR/POLICY/BS/EPOCH: 0.01/poly/16/130 | train/val | 42.58% | [cfg](https://raw.githubusercontent.com/SegmentationBLWX/sssegmentation/main/ssseg/configs/danet/danet_resnet101os16_ade20k.py) | [model](https://github.com/SegmentationBLWX/modelstore/releases/download/ssseg_danet/danet_resnet101os16_ade20k.pth) | [log](https://github.com/SegmentationBLWX/modelstore/releases/download/ssseg_danet/danet_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 | 79.47% | [cfg](https://raw.githubusercontent.com/SegmentationBLWX/sssegmentation/main/ssseg/configs/danet/danet_resnet50os8_cityscapes.py) | [model](https://github.com/SegmentationBLWX/modelstore/releases/download/ssseg_danet/danet_resnet50os8_cityscapes.pth) | [log](https://github.com/SegmentationBLWX/modelstore/releases/download/ssseg_danet/danet_resnet50os8_cityscapes.log) |
| R-50-D16 | ImageNet-1k-224x224 | 512x1024 | LR/POLICY/BS/EPOCH: 0.01/poly/8/220 | train/val | 77.60% | [cfg](https://raw.githubusercontent.com/SegmentationBLWX/sssegmentation/main/ssseg/configs/danet/danet_resnet50os16_cityscapes.py) | [model](https://github.com/SegmentationBLWX/modelstore/releases/download/ssseg_danet/danet_resnet50os16_cityscapes.pth) | [log](https://github.com/SegmentationBLWX/modelstore/releases/download/ssseg_danet/danet_resnet50os16_cityscapes.log) |
| R-101-D8 | ImageNet-1k-224x224 | 512x1024 | LR/POLICY/BS/EPOCH: 0.01/poly/8/220 | train/val | 80.55% | [cfg](https://raw.githubusercontent.com/SegmentationBLWX/sssegmentation/main/ssseg/configs/danet/danet_resnet101os8_cityscapes.py) | [model](https://github.com/SegmentationBLWX/modelstore/releases/download/ssseg_danet/danet_resnet101os8_cityscapes.pth) | [log](https://github.com/SegmentationBLWX/modelstore/releases/download/ssseg_danet/danet_resnet101os8_cityscapes.log) |
| R-101-D16 | ImageNet-1k-224x224 | 512x1024 | LR/POLICY/BS/EPOCH: 0.01/poly/8/220 | train/val | 78.23% | [cfg](https://raw.githubusercontent.com/SegmentationBLWX/sssegmentation/main/ssseg/configs/danet/danet_resnet101os16_cityscapes.py) | [model](https://github.com/SegmentationBLWX/modelstore/releases/download/ssseg_danet/danet_resnet101os16_cityscapes.pth) | [log](https://github.com/SegmentationBLWX/modelstore/releases/download/ssseg_danet/danet_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**