-
-
Notifications
You must be signed in to change notification settings - Fork 108
/
Copy pathREADME.md
40 lines (27 loc) · 4.84 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
## Introduction
<a href="https://github.com/Meituan-AutoML/Twins">Official Repo</a>
<a href="https://github.com/SegmentationBLWX/sssegmentation/blob/main/ssseg/modules/models/backbones/twins.py">Code Snippet</a>
<details>
<summary align="left"><a href="https://arxiv.org/pdf/2104.13840.pdf">Twins (NeurIPS'2021)</a></summary>
```latex
@article{chu2021twins,
title={Twins: Revisiting spatial attention design in vision transformers},
author={Chu, Xiangxiang and Tian, Zhi and Wang, Yuqing and Zhang, Bo and Ren, Haibing and Wei, Xiaolin and Xia, Huaxia and Shen, Chunhua},
journal={arXiv preprint arXiv:2104.13840},
year={2021}
}
```
</details>
## Results
#### ADE20k
| Segmentor | Pretrain | Backbone | Crop Size | Schedule | Train/Eval Set | mIoU | Download |
| :-: | :-: | :-: | :-: | :-: | :-: | :-: | :-: |
| UperNet | ImageNet-1k-224x224 | SVT-S | 512x512 | LR/POLICY/BS/EPOCH: 0.00006/poly/16/130 | train/val | 46.16% | [cfg](https://raw.githubusercontent.com/SegmentationBLWX/sssegmentation/main/ssseg/configs/twins/upernet_svtsmall_ade20k.py) | [model](https://github.com/SegmentationBLWX/modelstore/releases/download/ssseg_twins/upernet_svtsmall_ade20k.pth) | [log](https://github.com/SegmentationBLWX/modelstore/releases/download/ssseg_twins/upernet_svtsmall_ade20k.log) |
| UperNet | ImageNet-1k-224x224 | SVT-B | 512x512 | LR/POLICY/BS/EPOCH: 0.00006/poly/16/130 | train/val | 48.05% | [cfg](https://raw.githubusercontent.com/SegmentationBLWX/sssegmentation/main/ssseg/configs/twins/upernet_svtbase_ade20k.py) | [model](https://github.com/SegmentationBLWX/modelstore/releases/download/ssseg_twins/upernet_svtbase_ade20k.pth) | [log](https://github.com/SegmentationBLWX/modelstore/releases/download/ssseg_twins/upernet_svtbase_ade20k.log) |
| UperNet | ImageNet-1k-224x224 | SVT-L | 512x512 | LR/POLICY/BS/EPOCH: 0.00006/poly/16/130 | train/val | 49.80% | [cfg](https://raw.githubusercontent.com/SegmentationBLWX/sssegmentation/main/ssseg/configs/twins/upernet_svtlarge_ade20k.py) | [model](https://github.com/SegmentationBLWX/modelstore/releases/download/ssseg_twins/upernet_svtlarge_ade20k.pth) | [log](https://github.com/SegmentationBLWX/modelstore/releases/download/ssseg_twins/upernet_svtlarge_ade20k.log) |
| UperNet | ImageNet-1k-224x224 | PCPVT-S | 512x512 | LR/POLICY/BS/EPOCH: 0.00006/poly/16/130 | train/val | 46.07% | [cfg](https://raw.githubusercontent.com/SegmentationBLWX/sssegmentation/main/ssseg/configs/twins/upernet_pcpvtsmall_ade20k.py) | [model](https://github.com/SegmentationBLWX/modelstore/releases/download/ssseg_twins/upernet_pcpvtsmall_ade20k.pth) | [log](https://github.com/SegmentationBLWX/modelstore/releases/download/ssseg_twins/upernet_pcpvtsmall_ade20k.log) |
| UperNet | ImageNet-1k-224x224 | PCPVT-B | 512x512 | LR/POLICY/BS/EPOCH: 0.00006/poly/16/130 | train/val | 48.06% | [cfg](https://raw.githubusercontent.com/SegmentationBLWX/sssegmentation/main/ssseg/configs/twins/upernet_pcpvtbase_ade20k.py) | [model](https://github.com/SegmentationBLWX/modelstore/releases/download/ssseg_twins/upernet_pcpvtbase_ade20k.pth) | [log](https://github.com/SegmentationBLWX/modelstore/releases/download/ssseg_twins/upernet_pcpvtbase_ade20k.log) |
| UperNet | ImageNet-1k-224x224 | PCPVT-L | 512x512 | LR/POLICY/BS/EPOCH: 0.00006/poly/16/130 | train/val | 49.35% | [cfg](https://raw.githubusercontent.com/SegmentationBLWX/sssegmentation/main/ssseg/configs/twins/upernet_pcpvtlarge_ade20k.py) | [model](https://github.com/SegmentationBLWX/modelstore/releases/download/ssseg_twins/upernet_pcpvtlarge_ade20k.pth) | [log](https://github.com/SegmentationBLWX/modelstore/releases/download/ssseg_twins/upernet_pcpvtlarge_ade20k.log) |
## More
You can also download the model weights from following sources:
- BaiduNetdisk: https://pan.baidu.com/s/1gD-NJJWOtaHCtB0qHE79rA with access code **s757**