-
-
Notifications
You must be signed in to change notification settings - Fork 108
/
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/CSAILVision/unifiedparsing">Official Repo</a>
<a href="https://github.com/SegmentationBLWX/sssegmentation/blob/main/ssseg/modules/models/segmentors/upernet/upernet.py">Code Snippet</a>
<details>
<summary align="left"><a href="https://arxiv.org/pdf/1807.10221.pdf">UPerNet (ECCV'2018)</a></summary>
```latex
@inproceedings{xiao2018unified,
title={Unified perceptual parsing for scene understanding},
author={Xiao, Tete and Liu, Yingcheng and Zhou, Bolei and Jiang, Yuning and Sun, Jian},
booktitle={Proceedings of the European Conference on Computer Vision (ECCV)},
pages={418--434},
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 | 76.86% | [cfg](https://raw.githubusercontent.com/SegmentationBLWX/sssegmentation/main/ssseg/configs/upernet/upernet_resnet50os8_voc.py) | [model](https://github.com/SegmentationBLWX/modelstore/releases/download/ssseg_upernet/upernet_resnet50os8_voc.pth) | [log](https://github.com/SegmentationBLWX/modelstore/releases/download/ssseg_upernet/upernet_resnet50os8_voc.log) |
| R-50-D16 | ImageNet-1k-224x224 | 512x512 | LR/POLICY/BS/EPOCH: 0.01/poly/16/60 | trainaug/val | 77.48% | [cfg](https://raw.githubusercontent.com/SegmentationBLWX/sssegmentation/main/ssseg/configs/upernet/upernet_resnet50os16_voc.py) | [model](https://github.com/SegmentationBLWX/modelstore/releases/download/ssseg_upernet/upernet_resnet50os16_voc.pth) | [log](https://github.com/SegmentationBLWX/modelstore/releases/download/ssseg_upernet/upernet_resnet50os16_voc.log) |
| R-101-D8 | ImageNet-1k-224x224 | 512x512 | LR/POLICY/BS/EPOCH: 0.01/poly/16/60 | trainaug/val | 79.13% | [cfg](https://raw.githubusercontent.com/SegmentationBLWX/sssegmentation/main/ssseg/configs/upernet/upernet_resnet101os8_voc.py) | [model](https://github.com/SegmentationBLWX/modelstore/releases/download/ssseg_upernet/upernet_resnet101os8_voc.pth) | [log](https://github.com/SegmentationBLWX/modelstore/releases/download/ssseg_upernet/upernet_resnet101os8_voc.log) |
| R-101-D16 | ImageNet-1k-224x224 | 512x512 | LR/POLICY/BS/EPOCH: 0.01/poly/16/60 | trainaug/val | 77.88% | [cfg](https://raw.githubusercontent.com/SegmentationBLWX/sssegmentation/main/ssseg/configs/upernet/upernet_resnet101os16_voc.py) | [model](https://github.com/SegmentationBLWX/modelstore/releases/download/ssseg_upernet/upernet_resnet101os16_voc.pth) | [log](https://github.com/SegmentationBLWX/modelstore/releases/download/ssseg_upernet/upernet_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 | 43.02% | [cfg](https://raw.githubusercontent.com/SegmentationBLWX/sssegmentation/main/ssseg/configs/upernet/upernet_resnet50os8_ade20k.py) | [model](https://github.com/SegmentationBLWX/modelstore/releases/download/ssseg_upernet/upernet_resnet50os8_ade20k.pth) | [log](https://github.com/SegmentationBLWX/modelstore/releases/download/ssseg_upernet/upernet_resnet50os8_ade20k.log) |
| R-50-D16 | ImageNet-1k-224x224 | 512x512 | LR/POLICY/BS/EPOCH: 0.01/poly/16/130 | train/val | 42.87% | [cfg](https://raw.githubusercontent.com/SegmentationBLWX/sssegmentation/main/ssseg/configs/upernet/upernet_resnet50os16_ade20k.py) | [model](https://github.com/SegmentationBLWX/modelstore/releases/download/ssseg_upernet/upernet_resnet50os16_ade20k.pth) | [log](https://github.com/SegmentationBLWX/modelstore/releases/download/ssseg_upernet/upernet_resnet50os16_ade20k.log) |
| R-101-D8 | ImageNet-1k-224x224 | 512x512 | LR/POLICY/BS/EPOCH: 0.01/poly/16/130 | train/val | 44.92% | [cfg](https://raw.githubusercontent.com/SegmentationBLWX/sssegmentation/main/ssseg/configs/upernet/upernet_resnet101os8_ade20k.py) | [model](https://github.com/SegmentationBLWX/modelstore/releases/download/ssseg_upernet/upernet_resnet101os8_ade20k.pth) | [log](https://github.com/SegmentationBLWX/modelstore/releases/download/ssseg_upernet/upernet_resnet101os8_ade20k.log) |
| R-101-D16 | ImageNet-1k-224x224 | 512x512 | LR/POLICY/BS/EPOCH: 0.01/poly/16/130 | train/val | 44.77% | [cfg](https://raw.githubusercontent.com/SegmentationBLWX/sssegmentation/main/ssseg/configs/upernet/upernet_resnet101os16_ade20k.py) | [model](https://github.com/SegmentationBLWX/modelstore/releases/download/ssseg_upernet/upernet_resnet101os16_ade20k.pth) | [log](https://github.com/SegmentationBLWX/modelstore/releases/download/ssseg_upernet/upernet_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 | 79.08% | [cfg](https://raw.githubusercontent.com/SegmentationBLWX/sssegmentation/main/ssseg/configs/upernet/upernet_resnet50os8_cityscapes.py) | [model](https://github.com/SegmentationBLWX/modelstore/releases/download/ssseg_upernet/upernet_resnet50os8_cityscapes.pth) | [log](https://github.com/SegmentationBLWX/modelstore/releases/download/ssseg_upernet/upernet_resnet50os8_cityscapes.log) |
| R-50-D16 | ImageNet-1k-224x224 | 512x1024 | LR/POLICY/BS/EPOCH: 0.01/poly/8/220 | train/val | 78.94% | [cfg](https://raw.githubusercontent.com/SegmentationBLWX/sssegmentation/main/ssseg/configs/upernet/upernet_resnet50os16_cityscapes.py) | [model](https://github.com/SegmentationBLWX/modelstore/releases/download/ssseg_upernet/upernet_resnet50os16_cityscapes.pth) | [log](https://github.com/SegmentationBLWX/modelstore/releases/download/ssseg_upernet/upernet_resnet50os16_cityscapes.log) |
| R-101-D8 | ImageNet-1k-224x224 | 512x1024 | LR/POLICY/BS/EPOCH: 0.01/poly/8/220 | train/val | 80.39% | [cfg](https://raw.githubusercontent.com/SegmentationBLWX/sssegmentation/main/ssseg/configs/upernet/upernet_resnet101os8_cityscapes.py) | [model](https://github.com/SegmentationBLWX/modelstore/releases/download/ssseg_upernet/upernet_resnet101os8_cityscapes.pth) | [log](https://github.com/SegmentationBLWX/modelstore/releases/download/ssseg_upernet/upernet_resnet101os8_cityscapes.log) |
| R-101-D16 | ImageNet-1k-224x224 | 512x1024 | LR/POLICY/BS/EPOCH: 0.01/poly/8/220 | train/val | 79.64% | [cfg](https://raw.githubusercontent.com/SegmentationBLWX/sssegmentation/main/ssseg/configs/upernet/upernet_resnet101os16_cityscapes.py) | [model](https://github.com/SegmentationBLWX/modelstore/releases/download/ssseg_upernet/upernet_resnet101os16_cityscapes.pth) | [log](https://github.com/SegmentationBLWX/modelstore/releases/download/ssseg_upernet/upernet_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**