-
-
Notifications
You must be signed in to change notification settings - Fork 108
/
Copy pathREADME.md
65 lines (45 loc) · 11.2 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
58
59
60
61
62
63
64
65
## Introduction
<a href="https://github.com/SegmentationBLWX/sssegmentation">Official Repo</a>
<a href="https://github.com/SegmentationBLWX/sssegmentation/blob/main/ssseg/modules/models/segmentors/isnet/isnet.py">Code Snippet</a>
<details>
<summary align="left"><a href="https://arxiv.org/pdf/2108.12382.pdf">ISNet (ICCV'2021)</a></summary>
```latex
@inproceedings{jin2021isnet,
title={ISNet: Integrate Image-Level and Semantic-Level Context for Semantic Segmentation},
author={Jin, Zhenchao and Liu, Bin and Chu, Qi and Yu, Nenghai},
booktitle={Proceedings of the IEEE/CVF International Conference on Computer Vision},
pages={7189--7198},
year={2021}
}
```
</details>
## Results
#### COCOStuff-10k
| Backbone | Pretrain | Crop Size | Schedule | Train/Eval Set | mIoU/mIoU(ms+flip) | Download |
| :-: | :-: | :-: | :-: | :-: | :-: | :-: |
| R-50-D8 | ImageNet-1k-224x224 | 512x512 | LR/POLICY/BS/EPOCH: 0.001/poly/16/110 | train/test | 38.06%/40.39% | [cfg](https://raw.githubusercontent.com/SegmentationBLWX/sssegmentation/main/ssseg/configs/isnet/isnet_resnet50os8_cocostuff10k.py) | [model](https://github.com/SegmentationBLWX/modelstore/releases/download/ssseg_isnet/isnet_resnet50os8_cocostuff10k.pth) | [log](https://github.com/SegmentationBLWX/modelstore/releases/download/ssseg_isnet/isnet_resnet50os8_cocostuff10k.log) |
| R-101-D8 | ImageNet-1k-224x224 | 512x512 | LR/POLICY/BS/EPOCH: 0.001/poly/16/110 | train/test | 40.53%/41.74% | [cfg](https://raw.githubusercontent.com/SegmentationBLWX/sssegmentation/main/ssseg/configs/isnet/isnet_resnet101os8_cocostuff10k.py) | [model](https://github.com/SegmentationBLWX/modelstore/releases/download/ssseg_isnet/isnet_resnet101os8_cocostuff10k.pth) | [log](https://github.com/SegmentationBLWX/modelstore/releases/download/ssseg_isnet/isnet_resnet101os8_cocostuff10k.log) |
| S-101-D8 | ImageNet-1k-224x224 | 512x512 | LR/POLICY/BS/EPOCH: 0.001/poly/32/150 | train/test | 41.55%/42.53% | [cfg](https://raw.githubusercontent.com/SegmentationBLWX/sssegmentation/main/ssseg/configs/isnet/isnet_resnest101os8_cocostuff10k.py) | [model](https://github.com/SegmentationBLWX/modelstore/releases/download/ssseg_isnet/isnet_resnest101os8_cocostuff10k.pth) | [log](https://github.com/SegmentationBLWX/modelstore/releases/download/ssseg_isnet/isnet_resnest101os8_cocostuff10k.log) |
#### ADE20k
| Backbone | Pretrain | Crop Size | Schedule | Train/Eval Set | mIoU/mIoU(ms+flip) | Download |
| :-: | :-: | :-: | :-: | :-: | :-: | :-: |
| R-50-D8 | ImageNet-1k-224x224 | 512x512 | LR/POLICY/BS/EPOCH: 0.01/poly/16/130 | train/val | 44.22%/45.03% | [cfg](https://raw.githubusercontent.com/SegmentationBLWX/sssegmentation/main/ssseg/configs/isnet/isnet_resnet50os8_ade20k.py) | [model](https://github.com/SegmentationBLWX/modelstore/releases/download/ssseg_isnet/isnet_resnet50os8_ade20k.pth) | [log](https://github.com/SegmentationBLWX/modelstore/releases/download/ssseg_isnet/isnet_resnet50os8_ade20k.log) |
| R-101-D8 | ImageNet-1k-224x224 | 512x512 | LR/POLICY/BS/EPOCH: 0.01/poly/16/130 | train/val | 45.92%/47.29% | [cfg](https://raw.githubusercontent.com/SegmentationBLWX/sssegmentation/main/ssseg/configs/isnet/isnet_resnet101os8_ade20k.py) | [model](https://github.com/SegmentationBLWX/modelstore/releases/download/ssseg_isnet/isnet_resnet101os8_ade20k.pth) | [log](https://github.com/SegmentationBLWX/modelstore/releases/download/ssseg_isnet/isnet_resnet101os8_ade20k.log) |
| S-101-D8 | ImageNet-1k-224x224 | 512x512 | LR/POLICY/BS/EPOCH: 0.004/poly/16/180 | train/val | 46.65%/47.56% | [cfg](https://raw.githubusercontent.com/SegmentationBLWX/sssegmentation/main/ssseg/configs/isnet/isnet_resnest101os8_ade20k.py) | [model](https://github.com/SegmentationBLWX/modelstore/releases/download/ssseg_isnet/isnet_resnest101os8_ade20k.pth) | [log](https://github.com/SegmentationBLWX/modelstore/releases/download/ssseg_isnet/isnet_resnest101os8_ade20k.log) |
#### CityScapes
| Backbone | Pretrain | Crop Size | Schedule | Train/Eval Set | mIoU/mIoU(ms+flip) | Download |
| :-: | :-: | :-: | :-: | :-: | :-: | :-: |
| R-50-D8 | ImageNet-1k-224x224 | 512x1024 | LR/POLICY/BS/EPOCH: 0.01/poly/16/440 | train/val | 79.32%/81.31% | [cfg](https://raw.githubusercontent.com/SegmentationBLWX/sssegmentation/main/ssseg/configs/isnet/isnet_resnet50os8_cityscapes.py) | [model](https://github.com/SegmentationBLWX/modelstore/releases/download/ssseg_isnet/isnet_resnet50os8_cityscapes.pth) | [log](https://github.com/SegmentationBLWX/modelstore/releases/download/ssseg_isnet/isnet_resnet50os8_cityscapes.log) |
| R-101-D8 | ImageNet-1k-224x224 | 512x1024 | LR/POLICY/BS/EPOCH: 0.01/poly/16/440 | train/val | 80.56%/81.96% | [cfg](https://raw.githubusercontent.com/SegmentationBLWX/sssegmentation/main/ssseg/configs/isnet/isnet_resnet101os8_cityscapes.py) | [model](https://github.com/SegmentationBLWX/modelstore/releases/download/ssseg_isnet/isnet_resnet101os8_cityscapes.pth) | [log](https://github.com/SegmentationBLWX/modelstore/releases/download/ssseg_isnet/isnet_resnet101os8_cityscapes.log) |
| S-101-D8 | ImageNet-1k-224x224 | 512x1024 | LR/POLICY/BS/EPOCH: 0.01/poly/16/440 | train/val | 78.78%/81.33% | [cfg](https://raw.githubusercontent.com/SegmentationBLWX/sssegmentation/main/ssseg/configs/isnet/isnet_resnest101os8_cityscapes.py) | [model](https://github.com/SegmentationBLWX/modelstore/releases/download/ssseg_isnet/isnet_resnest101os8_cityscapes.pth) | [log](https://github.com/SegmentationBLWX/modelstore/releases/download/ssseg_isnet/isnet_resnest101os8_cityscapes.log) |
#### LIP
| Backbone | Pretrain | Crop Size | Schedule | Train/Eval Set | mIoU/mIoU(flip) | Download |
| :-: | :-: | :-: | :-: | :-: | :-: | :-: |
| R-50-D8 | ImageNet-1k-224x224 | 473x473 | LR/POLICY/BS/EPOCH: 0.01/poly/32/150 | train/val | 53.14%/53.41% | [cfg](https://raw.githubusercontent.com/SegmentationBLWX/sssegmentation/main/ssseg/configs/isnet/isnet_resnet50os8_lip.py) | [model](https://github.com/SegmentationBLWX/modelstore/releases/download/ssseg_isnet/isnet_resnet50os8_lip.pth) | [log](https://github.com/SegmentationBLWX/modelstore/releases/download/ssseg_isnet/isnet_resnet50os8_lip.log) |
| R-101-D8 | ImageNet-1k-224x224 | 473x473 | LR/POLICY/BS/EPOCH: 0.01/poly/32/150 | train/val | 54.96%/55.41% | [cfg](https://raw.githubusercontent.com/SegmentationBLWX/sssegmentation/main/ssseg/configs/isnet/isnet_resnet101os8_lip.py) | [model](https://github.com/SegmentationBLWX/modelstore/releases/download/ssseg_isnet/isnet_resnet101os8_lip.pth) | [log](https://github.com/SegmentationBLWX/modelstore/releases/download/ssseg_isnet/isnet_resnet101os8_lip.log) |
| S-101-D8 | ImageNet-1k-224x224 | 473x473 | LR/POLICY/BS/EPOCH: 0.007/poly/40/150 | train/val | 56.52%/56.81% | [cfg](https://raw.githubusercontent.com/SegmentationBLWX/sssegmentation/main/ssseg/configs/isnet/isnet_resnest101os8_lip.py) | [model](https://github.com/SegmentationBLWX/modelstore/releases/download/ssseg_isnet/isnet_resnest101os8_lip.pth) | [log](https://github.com/SegmentationBLWX/modelstore/releases/download/ssseg_isnet/isnet_resnest101os8_lip.log) |
## More
You can also download the model weights from following sources:
- BaiduNetdisk: https://pan.baidu.com/s/1gD-NJJWOtaHCtB0qHE79rA with access code **s757**
Please note that, due to differences in computational precision, the numerical values obtained when testing model performance on different versions of PyTorch or graphics cards may vary slightly.
This is a normal phenomenon and the performance differences are generally within 0.1%.