-
-
Notifications
You must be signed in to change notification settings - Fork 106
/
Copy pathREADME.md
57 lines (40 loc) · 9.8 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/hszhao/PSANet">Official Repo</a>
<a href="https://github.com/SegmentationBLWX/sssegmentation/blob/main/ssseg/modules/models/segmentors/psanet/psanet.py">Code Snippet</a>
<details>
<summary align="left"><a href="https://openaccess.thecvf.com/content_ECCV_2018/papers/Hengshuang_Zhao_PSANet_Point-wise_Spatial_ECCV_2018_paper.pdf">PSANet (ECCV'2018)</a></summary>
```latex
@inproceedings{zhao2018psanet,
title={Psanet: Point-wise spatial attention network for scene parsing},
author={Zhao, Hengshuang and Zhang, Yi and Liu, Shu and Shi, Jianping and Change Loy, Chen and Lin, Dahua and Jia, Jiaya},
booktitle={Proceedings of the European Conference on Computer Vision (ECCV)},
pages={267--283},
year={2018}
}
```
</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 | 77.06% | [cfg](https://raw.githubusercontent.com/SegmentationBLWX/sssegmentation/main/ssseg/configs/psanet/psanet_resnet50os8_voc.py) | [model](https://github.com/SegmentationBLWX/modelstore/releases/download/ssseg_psanet/psanet_resnet50os8_voc.pth) | [log](https://github.com/SegmentationBLWX/modelstore/releases/download/ssseg_psanet/psanet_resnet50os8_voc.log) |
| R-50-D16 | ImageNet-1k-224x224 | 512x512 | LR/POLICY/BS/EPOCH: 0.01/poly/16/60 | trainaug/val | 76.92% | [cfg](https://raw.githubusercontent.com/SegmentationBLWX/sssegmentation/main/ssseg/configs/psanet/psanet_resnet50os16_voc.py) | [model](https://github.com/SegmentationBLWX/modelstore/releases/download/ssseg_psanet/psanet_resnet50os16_voc.pth) | [log](https://github.com/SegmentationBLWX/modelstore/releases/download/ssseg_psanet/psanet_resnet50os16_voc.log) |
| R-101-D8 | ImageNet-1k-224x224 | 512x512 | LR/POLICY/BS/EPOCH: 0.01/poly/16/60 | trainaug/val | 78.97% | [cfg](https://raw.githubusercontent.com/SegmentationBLWX/sssegmentation/main/ssseg/configs/psanet/psanet_resnet101os8_voc.py) | [model](https://github.com/SegmentationBLWX/modelstore/releases/download/ssseg_psanet/psanet_resnet101os8_voc.pth) | [log](https://github.com/SegmentationBLWX/modelstore/releases/download/ssseg_psanet/psanet_resnet101os8_voc.log) |
| R-101-D16 | ImageNet-1k-224x224 | 512x512 | LR/POLICY/BS/EPOCH: 0.01/poly/16/60 | trainaug/val | 77.90% | [cfg](https://raw.githubusercontent.com/SegmentationBLWX/sssegmentation/main/ssseg/configs/psanet/psanet_resnet101os16_voc.py) | [model](https://github.com/SegmentationBLWX/modelstore/releases/download/ssseg_psanet/psanet_resnet101os16_voc.pth) | [log](https://github.com/SegmentationBLWX/modelstore/releases/download/ssseg_psanet/psanet_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 | 41.99% | [cfg](https://raw.githubusercontent.com/SegmentationBLWX/sssegmentation/main/ssseg/configs/psanet/psanet_resnet50os8_ade20k.py) | [model](https://github.com/SegmentationBLWX/modelstore/releases/download/ssseg_psanet/psanet_resnet50os8_ade20k.pth) | [log](https://github.com/SegmentationBLWX/modelstore/releases/download/ssseg_psanet/psanet_resnet50os8_ade20k.log) |
| R-50-D16 | ImageNet-1k-224x224 | 512x512 | LR/POLICY/BS/EPOCH: 0.01/poly/16/130 | train/val | 40.26% | [cfg](https://raw.githubusercontent.com/SegmentationBLWX/sssegmentation/main/ssseg/configs/psanet/psanet_resnet50os16_ade20k.py) | [model](https://github.com/SegmentationBLWX/modelstore/releases/download/ssseg_psanet/psanet_resnet50os16_ade20k.pth) | [log](https://github.com/SegmentationBLWX/modelstore/releases/download/ssseg_psanet/psanet_resnet50os16_ade20k.log) |
| R-101-D8 | ImageNet-1k-224x224 | 512x512 | LR/POLICY/BS/EPOCH: 0.01/poly/16/130 | train/val | 43.85% | [cfg](https://raw.githubusercontent.com/SegmentationBLWX/sssegmentation/main/ssseg/configs/psanet/psanet_resnet101os8_ade20k.py) | [model](https://github.com/SegmentationBLWX/modelstore/releases/download/ssseg_psanet/psanet_resnet101os8_ade20k.pth) | [log](https://github.com/SegmentationBLWX/modelstore/releases/download/ssseg_psanet/psanet_resnet101os8_ade20k.log) |
| R-101-D16 | ImageNet-1k-224x224 | 512x512 | LR/POLICY/BS/EPOCH: 0.01/poly/16/130 | train/val | 42.05% | [cfg](https://raw.githubusercontent.com/SegmentationBLWX/sssegmentation/main/ssseg/configs/psanet/psanet_resnet101os16_ade20k.py) | [model](https://github.com/SegmentationBLWX/modelstore/releases/download/ssseg_psanet/psanet_resnet101os16_ade20k.pth) | [log](https://github.com/SegmentationBLWX/modelstore/releases/download/ssseg_psanet/psanet_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 | 78.88% | [cfg](https://raw.githubusercontent.com/SegmentationBLWX/sssegmentation/main/ssseg/configs/psanet/psanet_resnet50os8_cityscapes.py) | [model](https://github.com/SegmentationBLWX/modelstore/releases/download/ssseg_psanet/psanet_resnet50os8_cityscapes.pth) | [log](https://github.com/SegmentationBLWX/modelstore/releases/download/ssseg_psanet/psanet_resnet50os8_cityscapes.log) |
| R-50-D16 | ImageNet-1k-224x224 | 512x1024 | LR/POLICY/BS/EPOCH: 0.01/poly/8/220 | train/val | 76.66% | [cfg](https://raw.githubusercontent.com/SegmentationBLWX/sssegmentation/main/ssseg/configs/psanet/psanet_resnet50os16_cityscapes.py) | [model](https://github.com/SegmentationBLWX/modelstore/releases/download/ssseg_psanet/psanet_resnet50os16_cityscapes.pth) | [log](https://github.com/SegmentationBLWX/modelstore/releases/download/ssseg_psanet/psanet_resnet50os16_cityscapes.log) |
| R-101-D8 | ImageNet-1k-224x224 | 512x1024 | LR/POLICY/BS/EPOCH: 0.01/poly/8/220 | train/val | 79.65% | [cfg](https://raw.githubusercontent.com/SegmentationBLWX/sssegmentation/main/ssseg/configs/psanet/psanet_resnet101os8_cityscapes.py) | [model](https://github.com/SegmentationBLWX/modelstore/releases/download/ssseg_psanet/psanet_resnet101os8_cityscapes.pth) | [log](https://github.com/SegmentationBLWX/modelstore/releases/download/ssseg_psanet/psanet_resnet101os8_cityscapes.log) |
| R-101-D16 | ImageNet-1k-224x224 | 512x1024 | LR/POLICY/BS/EPOCH: 0.01/poly/8/220 | train/val | 77.04% | [cfg](https://raw.githubusercontent.com/SegmentationBLWX/sssegmentation/main/ssseg/configs/psanet/psanet_resnet101os16_cityscapes.py) | [model](https://github.com/SegmentationBLWX/modelstore/releases/download/ssseg_psanet/psanet_resnet101os16_cityscapes.pth) | [log](https://github.com/SegmentationBLWX/modelstore/releases/download/ssseg_psanet/psanet_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**