-
-
Notifications
You must be signed in to change notification settings - Fork 108
/
Copy pathREADME.md
70 lines (49 loc) · 14.3 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
66
67
68
69
70
## Introduction
<a href="https://github.com/tensorflow/models/tree/master/research/deeplab">Official Repo</a>
<a href="https://github.com/SegmentationBLWX/sssegmentation/blob/main/ssseg/modules/models/segmentors/deeplabv3/deeplabv3.py">Code Snippet</a>
<details>
<summary align="left"><a href="https://arxiv.org/pdf/1706.05587.pdf">DeepLabV3 (ArXiv'2017)</a></summary>
```latex
@article{chen2017rethinking,
title={Rethinking atrous convolution for semantic image segmentation},
author={Chen, Liang-Chieh and Papandreou, George and Schroff, Florian and Adam, Hartwig},
journal={arXiv preprint arXiv:1706.05587},
year={2017}
}
```
</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 | 77.72% | [cfg](https://raw.githubusercontent.com/SegmentationBLWX/sssegmentation/main/ssseg/configs/deeplabv3/deeplabv3_resnet50os8_voc.py) | [model](https://github.com/SegmentationBLWX/modelstore/releases/download/ssseg_deeplabv3/deeplabv3_resnet50os8_voc.pth) | [log](https://github.com/SegmentationBLWX/modelstore/releases/download/ssseg_deeplabv3/deeplabv3_resnet50os8_voc.log) |
| R-50-D16 | 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/deeplabv3/deeplabv3_resnet50os16_voc.py) | [model](https://github.com/SegmentationBLWX/modelstore/releases/download/ssseg_deeplabv3/deeplabv3_resnet50os16_voc.pth) | [log](https://github.com/SegmentationBLWX/modelstore/releases/download/ssseg_deeplabv3/deeplabv3_resnet50os16_voc.log) |
| R-101-D8 | ImageNet-1k-224x224 | 512x512 | LR/POLICY/BS/EPOCH: 0.01/poly/16/60 | trainaug/val | 79.52% | [cfg](https://raw.githubusercontent.com/SegmentationBLWX/sssegmentation/main/ssseg/configs/deeplabv3/deeplabv3_resnet101os8_voc.py) | [model](https://github.com/SegmentationBLWX/modelstore/releases/download/ssseg_deeplabv3/deeplabv3_resnet101os8_voc.pth) | [log](https://github.com/SegmentationBLWX/modelstore/releases/download/ssseg_deeplabv3/deeplabv3_resnet101os8_voc.log) |
| R-101-D16 | ImageNet-1k-224x224 | 512x512 | LR/POLICY/BS/EPOCH: 0.01/poly/16/60 | trainaug/val | 78.55% | [cfg](https://raw.githubusercontent.com/SegmentationBLWX/sssegmentation/main/ssseg/configs/deeplabv3/deeplabv3_resnet101os16_voc.py) | [model](https://github.com/SegmentationBLWX/modelstore/releases/download/ssseg_deeplabv3/deeplabv3_resnet101os16_voc.pth) | [log](https://github.com/SegmentationBLWX/modelstore/releases/download/ssseg_deeplabv3/deeplabv3_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 | 43.19% | [cfg](https://raw.githubusercontent.com/SegmentationBLWX/sssegmentation/main/ssseg/configs/deeplabv3/deeplabv3_resnet50os8_ade20k.py) | [model](https://github.com/SegmentationBLWX/modelstore/releases/download/ssseg_deeplabv3/deeplabv3_resnet50os8_ade20k.pth) | [log](https://github.com/SegmentationBLWX/modelstore/releases/download/ssseg_deeplabv3/deeplabv3_resnet50os8_ade20k.log) |
| R-50-D16 | ImageNet-1k-224x224 | 512x512 | LR/POLICY/BS/EPOCH: 0.01/poly/16/130 | train/val | 41.41% | [cfg](https://raw.githubusercontent.com/SegmentationBLWX/sssegmentation/main/ssseg/configs/deeplabv3/deeplabv3_resnet50os16_ade20k.py) | [model](https://github.com/SegmentationBLWX/modelstore/releases/download/ssseg_deeplabv3/deeplabv3_resnet50os16_ade20k.pth) | [log](https://github.com/SegmentationBLWX/modelstore/releases/download/ssseg_deeplabv3/deeplabv3_resnet50os16_ade20k.log) |
| R-101-D8 | ImageNet-1k-224x224 | 512x512 | LR/POLICY/BS/EPOCH: 0.01/poly/16/130 | train/val | 45.16% | [cfg](https://raw.githubusercontent.com/SegmentationBLWX/sssegmentation/main/ssseg/configs/deeplabv3/deeplabv3_resnet101os8_ade20k.py) | [model](https://github.com/SegmentationBLWX/modelstore/releases/download/ssseg_deeplabv3/deeplabv3_resnet101os8_ade20k.pth) | [log](https://github.com/SegmentationBLWX/modelstore/releases/download/ssseg_deeplabv3/deeplabv3_resnet101os8_ade20k.log) |
| R-101-D16 | ImageNet-1k-224x224 | 512x512 | LR/POLICY/BS/EPOCH: 0.01/poly/16/130 | train/val | 43.45% | [cfg](https://raw.githubusercontent.com/SegmentationBLWX/sssegmentation/main/ssseg/configs/deeplabv3/deeplabv3_resnet101os16_ade20k.py) | [model](https://github.com/SegmentationBLWX/modelstore/releases/download/ssseg_deeplabv3/deeplabv3_resnet101os16_ade20k.pth) | [log](https://github.com/SegmentationBLWX/modelstore/releases/download/ssseg_deeplabv3/deeplabv3_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 | 79.62% | [cfg](https://raw.githubusercontent.com/SegmentationBLWX/sssegmentation/main/ssseg/configs/deeplabv3/deeplabv3_resnet50os8_cityscapes.py) | [model](https://github.com/SegmentationBLWX/modelstore/releases/download/ssseg_deeplabv3/deeplabv3_resnet50os8_cityscapes.pth) | [log](https://github.com/SegmentationBLWX/modelstore/releases/download/ssseg_deeplabv3/deeplabv3_resnet50os8_cityscapes.log) |
| R-50-D16 | ImageNet-1k-224x224 | 512x1024 | LR/POLICY/BS/EPOCH: 0.01/poly/8/220 | train/val | 78.19% | [cfg](https://raw.githubusercontent.com/SegmentationBLWX/sssegmentation/main/ssseg/configs/deeplabv3/deeplabv3_resnet50os16_cityscapes.py) | [model](https://github.com/SegmentationBLWX/modelstore/releases/download/ssseg_deeplabv3/deeplabv3_resnet50os16_cityscapes.pth) | [log](https://github.com/SegmentationBLWX/modelstore/releases/download/ssseg_deeplabv3/deeplabv3_resnet50os16_cityscapes.log) |
| R-101-D8 | ImageNet-1k-224x224 | 512x1024 | LR/POLICY/BS/EPOCH: 0.01/poly/8/220 | train/val | 80.28% | [cfg](https://raw.githubusercontent.com/SegmentationBLWX/sssegmentation/main/ssseg/configs/deeplabv3/deeplabv3_resnet101os8_cityscapes.py) | [model](https://github.com/SegmentationBLWX/modelstore/releases/download/ssseg_deeplabv3/deeplabv3_resnet101os8_cityscapes.pth) | [log](https://github.com/SegmentationBLWX/modelstore/releases/download/ssseg_deeplabv3/deeplabv3_resnet101os8_cityscapes.log) |
| R-101-D16 | ImageNet-1k-224x224 | 512x1024 | LR/POLICY/BS/EPOCH: 0.01/poly/8/220 | train/val | 78.03% | [cfg](https://raw.githubusercontent.com/SegmentationBLWX/sssegmentation/main/ssseg/configs/deeplabv3/deeplabv3_resnet101os16_cityscapes.py) | [model](https://github.com/SegmentationBLWX/modelstore/releases/download/ssseg_deeplabv3/deeplabv3_resnet101os16_cityscapes.pth) | [log](https://github.com/SegmentationBLWX/modelstore/releases/download/ssseg_deeplabv3/deeplabv3_resnet101os16_cityscapes.log) |
#### PASCAL Context
| Backbone | Pretrain | Crop Size | Schedule | Train/Eval Set | mIoU | Download |
| :-: | :-: | :-: | :-: | :-: | :-: | :-: |
| R-50-D8 | ImageNet-1k-224x224 | 480x480 | LR/POLICY/BS/EPOCH: 0.004/poly/16/260 | train/val | 46.31% | [cfg](https://raw.githubusercontent.com/SegmentationBLWX/sssegmentation/main/ssseg/configs/deeplabv3/deeplabv3_resnet50os8_pascalcontext.py) | [model](https://github.com/SegmentationBLWX/modelstore/releases/download/ssseg_deeplabv3/deeplabv3_resnet50os8_pascalcontext.pth) | [log](https://github.com/SegmentationBLWX/modelstore/releases/download/ssseg_deeplabv3/deeplabv3_resnet50os8_pascalcontext.log) |
| R-101-D8 | ImageNet-1k-224x224 | 480x480 | LR/POLICY/BS/EPOCH: 0.004/poly/16/260 | train/val | 48.43% | [cfg](https://raw.githubusercontent.com/SegmentationBLWX/sssegmentation/main/ssseg/configs/deeplabv3/deeplabv3_resnet101os8_pascalcontext.py) | [model](https://github.com/SegmentationBLWX/modelstore/releases/download/ssseg_deeplabv3/deeplabv3_resnet101os8_pascalcontext.pth) | [log](https://github.com/SegmentationBLWX/modelstore/releases/download/ssseg_deeplabv3/deeplabv3_resnet101os8_pascalcontext.log) |
#### PASCAL Context 59
| Backbone | Pretrain | Crop Size | Schedule | Train/Eval Set | mIoU | Download |
| :-: | :-: | :-: | :-: | :-: | :-: | :-: |
| R-50-D8 | ImageNet-1k-224x224 | 480x480 | LR/POLICY/BS/EPOCH: 0.004/poly/16/260 | train/val | 51.69% | [cfg](https://raw.githubusercontent.com/SegmentationBLWX/sssegmentation/main/ssseg/configs/deeplabv3/deeplabv3_resnet50os8_pascalcontext59.py) | [model](https://github.com/SegmentationBLWX/modelstore/releases/download/ssseg_deeplabv3/deeplabv3_resnet50os8_pascalcontext59.pth) | [log](https://github.com/SegmentationBLWX/modelstore/releases/download/ssseg_deeplabv3/deeplabv3_resnet50os8_pascalcontext59.log) |
| R-101-D8 | ImageNet-1k-224x224 | 480x480 | LR/POLICY/BS/EPOCH: 0.004/poly/16/260 | train/val | 53.81% | [cfg](https://raw.githubusercontent.com/SegmentationBLWX/sssegmentation/main/ssseg/configs/deeplabv3/deeplabv3_resnet101os8_pascalcontext59.py) | [model](https://github.com/SegmentationBLWX/modelstore/releases/download/ssseg_deeplabv3/deeplabv3_resnet101os8_pascalcontext59.pth) | [log](https://github.com/SegmentationBLWX/modelstore/releases/download/ssseg_deeplabv3/deeplabv3_resnet101os8_pascalcontext59.log) |
## More
You can also download the model weights from following sources:
- BaiduNetdisk: https://pan.baidu.com/s/1gD-NJJWOtaHCtB0qHE79rA with access code **s757**