-
-
Notifications
You must be signed in to change notification settings - Fork 106
/
Copy pathREADME.md
126 lines (90 loc) · 35.8 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
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
## Introduction
<a href="https://github.com/SegmentationBLWX/sssegmentation">Official Repo</a>
<a href="https://github.com/SegmentationBLWX/sssegmentation/blob/main/ssseg/modules/models/segmentors/mcibiplusplus/mcibiplusplus.py">Code Snippet</a>
<details>
<summary align="left"><a href="https://arxiv.org/pdf/2209.04471.pdf">MCIBI++ (TPAMI'2022)</a></summary>
```latex
@article{jin2022mcibi++,
title={MCIBI++: Soft Mining Contextual Information Beyond Image for Semantic Segmentation},
author={Jin, Zhenchao and Yu, Dongdong and Yuan, Zehuan and Yu, Lequan},
journal={IEEE Transactions on Pattern Analysis and Machine Intelligence},
year={2022},
publisher={IEEE}
}
```
</details>
## Results of Different Frameworks
#### LIP
| Segmentor | Pretrain | Backbone | Crop Size | Schedule | Train/Eval Set | mIoU | Download |
| :-: | :-: | :-: | :-: | :-: | :-: | :-: | :-: |
| FCN | ImageNet-1k-224x224 | R-50-D8 | 473x473 | LR/POLICY/BS/EPOCH: 0.01/poly/32/150 | train/val | 51.13% | [cfg](https://raw.githubusercontent.com/SegmentationBLWX/sssegmentation/main/ssseg/configs/mcibiplusplus/mcibiplusplus_resnet50os8_lip.py) | [model](https://github.com/SegmentationBLWX/modelstore/releases/download/ssseg_mcibiplusplus/mcibiplusplus_resnet50os8_lip.pth) | [log](https://github.com/SegmentationBLWX/modelstore/releases/download/ssseg_mcibiplusplus/mcibiplusplus_resnet50os8_lip.log) |
| PSNet | ImageNet-1k-224x224 | R-50-D8 | 473x473 | LR/POLICY/BS/EPOCH: 0.01/poly/32/150 | train/val | 52.93% | [cfg](https://raw.githubusercontent.com/SegmentationBLWX/sssegmentation/main/ssseg/configs/mcibiplusplus/mcibiplusplus_ppm_resnet50os8_lip.py) | [model](https://github.com/SegmentationBLWX/modelstore/releases/download/ssseg_mcibiplusplus/mcibiplusplus_ppm_resnet50os8_lip.pth) | [log](https://github.com/SegmentationBLWX/modelstore/releases/download/ssseg_mcibiplusplus/mcibiplusplus_ppm_resnet50os8_lip.log) |
| UperNet | ImageNet-1k-224x224 | R-50-D8 | 473x473 | LR/POLICY/BS/EPOCH: 0.01/poly/32/150 | train/val | 53.92% | [cfg](https://raw.githubusercontent.com/SegmentationBLWX/sssegmentation/main/ssseg/configs/mcibiplusplus/mcibiplusplus_upernet_resnet50os8_lip.py) | [model](https://github.com/SegmentationBLWX/modelstore/releases/download/ssseg_mcibiplusplus/mcibiplusplus_upernet_resnet50os8_lip.pth) | [log](https://github.com/SegmentationBLWX/modelstore/releases/download/ssseg_mcibiplusplus/mcibiplusplus_upernet_resnet50os8_lip.log) |
| DeepLabV3 | ImageNet-1k-224x224 | R-50-D8 | 473x473 | LR/POLICY/BS/EPOCH: 0.01/poly/32/150 | train/val | 53.59% | [cfg](https://raw.githubusercontent.com/SegmentationBLWX/sssegmentation/main/ssseg/configs/mcibiplusplus/mcibiplusplus_aspp_resnet50os8_lip.py) | [model](https://github.com/SegmentationBLWX/modelstore/releases/download/ssseg_mcibiplusplus/mcibiplusplus_aspp_resnet50os8_lip.pth) | [log](https://github.com/SegmentationBLWX/modelstore/releases/download/ssseg_mcibiplusplus/mcibiplusplus_aspp_resnet50os8_lip.log) |
#### ADE20k
| Segmentor | Pretrain | Backbone | Crop Size | Schedule | Train/Eval Set | mIoU | Download |
| :-: | :-: | :-: | :-: | :-: | :-: | :-: | :-: |
| FCN | ImageNet-1k-224x224 | R-50-D8 | 512x512 | LR/POLICY/BS/EPOCH: 0.01/poly/16/130 | train/val | 43.39% | [cfg](https://raw.githubusercontent.com/SegmentationBLWX/sssegmentation/main/ssseg/configs/mcibiplusplus/mcibiplusplus_resnet50os8_ade20k.py) | [model](https://github.com/SegmentationBLWX/modelstore/releases/download/ssseg_mcibiplusplus/mcibiplusplus_resnet50os8_ade20k.pth) | [log](https://github.com/SegmentationBLWX/modelstore/releases/download/ssseg_mcibiplusplus/mcibiplusplus_resnet50os8_ade20k.log) |
| PSNet | ImageNet-1k-224x224 | R-50-D8 | 512x512 | LR/POLICY/BS/EPOCH: 0.01/poly/16/130 | train/val | 43.88% | [cfg](https://raw.githubusercontent.com/SegmentationBLWX/sssegmentation/main/ssseg/configs/mcibiplusplus/mcibiplusplus_ppm_resnet50os8_ade20k.py) | [model](https://github.com/SegmentationBLWX/modelstore/releases/download/ssseg_mcibiplusplus/mcibiplusplus_ppm_resnet50os8_ade20k.pth) | [log](https://github.com/SegmentationBLWX/modelstore/releases/download/ssseg_mcibiplusplus/mcibiplusplus_ppm_resnet50os8_ade20k.log) |
| UperNet | ImageNet-1k-224x224 | R-50-D8 | 512x512 | LR/POLICY/BS/EPOCH: 0.01/poly/16/130 | train/val | 44.30% | [cfg](https://raw.githubusercontent.com/SegmentationBLWX/sssegmentation/main/ssseg/configs/mcibiplusplus/mcibiplusplus_upernet_resnet50os8_ade20k.py) | [model](https://github.com/SegmentationBLWX/modelstore/releases/download/ssseg_mcibiplusplus/mcibiplusplus_upernet_resnet50os8_ade20k.pth) | [log](https://github.com/SegmentationBLWX/modelstore/releases/download/ssseg_mcibiplusplus/mcibiplusplus_upernet_resnet50os8_ade20k.log) |
| DeepLabV3 | ImageNet-1k-224x224 | R-50-D8 | 512x512 | LR/POLICY/BS/EPOCH: 0.01/poly/16/130 | train/val | 44.85% | [cfg](https://raw.githubusercontent.com/SegmentationBLWX/sssegmentation/main/ssseg/configs/mcibiplusplus/mcibiplusplus_aspp_resnet50os8_ade20k.py) | [model](https://github.com/SegmentationBLWX/modelstore/releases/download/ssseg_mcibiplusplus/mcibiplusplus_aspp_resnet50os8_ade20k.pth) | [log](https://github.com/SegmentationBLWX/modelstore/releases/download/ssseg_mcibiplusplus/mcibiplusplus_aspp_resnet50os8_ade20k.log) |
#### CityScapes
| Segmentor | Pretrain | Backbone | Crop Size | Schedule | Train/Eval Set | mIoU | Download |
| :-: | :-: | :-: | :-: | :-: | :-: | :-: | :-: |
| FCN | ImageNet-1k-224x224 | R-50-D8 | 512x1024 | LR/POLICY/BS/EPOCH: 0.01/poly/8/220 | train/val | 78.77% | [cfg](https://raw.githubusercontent.com/SegmentationBLWX/sssegmentation/main/ssseg/configs/mcibiplusplus/mcibiplusplus_resnet50os8_cityscapes.py) | [model](https://github.com/SegmentationBLWX/modelstore/releases/download/ssseg_mcibiplusplus/mcibiplusplus_resnet50os8_cityscapes.pth) | [log](https://github.com/SegmentationBLWX/modelstore/releases/download/ssseg_mcibiplusplus/mcibiplusplus_resnet50os8_cityscapes.log) |
| PSNet | ImageNet-1k-224x224 | R-50-D8 | 512x1024 | LR/POLICY/BS/EPOCH: 0.01/poly/8/220 | train/val | 79.91% | [cfg](https://raw.githubusercontent.com/SegmentationBLWX/sssegmentation/main/ssseg/configs/mcibiplusplus/mcibiplusplus_ppm_resnet50os8_cityscapes.py) | [model](https://github.com/SegmentationBLWX/modelstore/releases/download/ssseg_mcibiplusplus/mcibiplusplus_ppm_resnet50os8_cityscapes.pth) | [log](https://github.com/SegmentationBLWX/modelstore/releases/download/ssseg_mcibiplusplus/mcibiplusplus_ppm_resnet50os8_cityscapes.log) |
| UperNet | ImageNet-1k-224x224 | R-50-D8 | 512x1024 | LR/POLICY/BS/EPOCH: 0.01/poly/8/220 | train/val | 80.05% | [cfg](https://raw.githubusercontent.com/SegmentationBLWX/sssegmentation/main/ssseg/configs/mcibiplusplus/mcibiplusplus_upernet_resnet50os8_cityscapes.py) | [model](https://github.com/SegmentationBLWX/modelstore/releases/download/ssseg_mcibiplusplus/mcibiplusplus_upernet_resnet50os8_cityscapes.pth) | [log](https://github.com/SegmentationBLWX/modelstore/releases/download/ssseg_mcibiplusplus/mcibiplusplus_upernet_resnet50os8_cityscapes.log) |
| DeepLabV3 | ImageNet-1k-224x224 | R-50-D8 | 512x1024 | LR/POLICY/BS/EPOCH: 0.01/poly/8/220 | train/val | 80.72% | [cfg](https://raw.githubusercontent.com/SegmentationBLWX/sssegmentation/main/ssseg/configs/mcibiplusplus/mcibiplusplus_aspp_resnet50os8_cityscapes.py) | [model](https://github.com/SegmentationBLWX/modelstore/releases/download/ssseg_mcibiplusplus/mcibiplusplus_aspp_resnet50os8_cityscapes.pth) | [log](https://github.com/SegmentationBLWX/modelstore/releases/download/ssseg_mcibiplusplus/mcibiplusplus_aspp_resnet50os8_cityscapes.log) |
#### COCOStuff-10k
| Segmentor | Pretrain | Backbone | Crop Size | Schedule | Train/Eval Set | mIoU | Download |
| :-: | :-: | :-: | :-: | :-: | :-: | :-: | :-: |
| FCN | ImageNet-1k-224x224 | R-50-D8 | 512x512 | LR/POLICY/BS/EPOCH: 0.001/poly/16/110 | train/test | 37.38% | [cfg](https://raw.githubusercontent.com/SegmentationBLWX/sssegmentation/main/ssseg/configs/mcibiplusplus/mcibiplusplus_resnet50os8_cocostuff10k.py) | [model](https://github.com/SegmentationBLWX/modelstore/releases/download/ssseg_mcibiplusplus/mcibiplusplus_resnet50os8_cocostuff10k.pth) | [log](https://github.com/SegmentationBLWX/modelstore/releases/download/ssseg_mcibiplusplus/mcibiplusplus_resnet50os8_cocostuff10k.log) |
| PSNet | ImageNet-1k-224x224 | R-50-D8 | 512x512 | LR/POLICY/BS/EPOCH: 0.001/poly/16/110 | train/test | 38.47% | [cfg](https://raw.githubusercontent.com/SegmentationBLWX/sssegmentation/main/ssseg/configs/mcibiplusplus/mcibiplusplus_ppm_resnet50os8_cocostuff10k.py) | [model](https://github.com/SegmentationBLWX/modelstore/releases/download/ssseg_mcibiplusplus/mcibiplusplus_ppm_resnet50os8_cocostuff10k.pth) | [log](https://github.com/SegmentationBLWX/modelstore/releases/download/ssseg_mcibiplusplus/mcibiplusplus_ppm_resnet50os8_cocostuff10k.log) |
| UperNet | ImageNet-1k-224x224 | R-50-D8 | 512x512 | LR/POLICY/BS/EPOCH: 0.001/poly/16/110 | train/test | 39.20% | [cfg](https://raw.githubusercontent.com/SegmentationBLWX/sssegmentation/main/ssseg/configs/mcibiplusplus/mcibiplusplus_upernet_resnet50os8_cocostuff10k.py) | [model](https://github.com/SegmentationBLWX/modelstore/releases/download/ssseg_mcibiplusplus/mcibiplusplus_upernet_resnet50os8_cocostuff10k.pth) | [log](https://github.com/SegmentationBLWX/modelstore/releases/download/ssseg_mcibiplusplus/mcibiplusplus_upernet_resnet50os8_cocostuff10k.log) |
| DeepLabV3 | ImageNet-1k-224x224 | R-50-D8 | 512x512 | LR/POLICY/BS/EPOCH: 0.001/poly/16/110 | train/test | 38.94% | [cfg](https://raw.githubusercontent.com/SegmentationBLWX/sssegmentation/main/ssseg/configs/mcibiplusplus/mcibiplusplus_aspp_resnet50os8_cocostuff10k.py) | [model](https://github.com/SegmentationBLWX/modelstore/releases/download/ssseg_mcibiplusplus/mcibiplusplus_aspp_resnet50os8_cocostuff10k.pth) | [log](https://github.com/SegmentationBLWX/modelstore/releases/download/ssseg_mcibiplusplus/mcibiplusplus_aspp_resnet50os8_cocostuff10k.log) |
## SOTA Results
#### VSPW
| Segmentor | Pretrain | Backbone | Crop Size | Schedule | Train/Eval Set | mIoU | Download |
| :-: | :-: | :-: | :-: | :-: | :-: | :-: | :-: |
| UperNet | ImageNet-1k-224x224 | R-101-D8 | 512x512 | LR/POLICY/BS/EPOCH: 0.01/poly/16/240 | train/val | 43.21% | [cfg](https://raw.githubusercontent.com/SegmentationBLWX/sssegmentation/main/ssseg/configs/mcibiplusplus/mcibiplusplus_upernet_resnet101os8_vspw.py) | [model](https://github.com/SegmentationBLWX/modelstore/releases/download/ssseg_mcibiplusplus/mcibiplusplus_upernet_resnet101os8_vspw.pth) | [log](https://github.com/SegmentationBLWX/modelstore/releases/download/ssseg_mcibiplusplus/mcibiplusplus_upernet_resnet101os8_vspw.log) |
| UperNet | ImageNet-22k-384x384 | Swin-Large | 512x512 | LR/POLICY/BS/EPOCH: 0.00006/poly/16/240 | train/val | 56.04% | [cfg](https://raw.githubusercontent.com/SegmentationBLWX/sssegmentation/main/ssseg/configs/mcibiplusplus/mcibiplusplus_upernet_swinlarge_vspw.py) | [model](https://github.com/SegmentationBLWX/modelstore/releases/download/ssseg_mcibiplusplus/mcibiplusplus_upernet_swinlarge_vspw.pth) | [log](https://github.com/SegmentationBLWX/modelstore/releases/download/ssseg_mcibiplusplus/mcibiplusplus_upernet_swinlarge_vspw.log) |
#### PASCAL-VOC
| Segmentor | Pretrain | Backbone | Crop Size | Schedule | Train/Eval Set | mIoU | Download |
| :-: | :-: | :-: | :-: | :-: | :-: | :-: | :-: |
| UperNet | ImageNet-1k-224x224 | R-50-D8 | 512x512 | LR/POLICY/BS/EPOCH: 0.01/poly/16/60 | train/val | 79.48 | [cfg](https://raw.githubusercontent.com/SegmentationBLWX/sssegmentation/main/ssseg/configs/mcibiplusplus/mcibiplusplus_upernet_resnet50os8_voc.py) | [model](https://github.com/SegmentationBLWX/modelstore/releases/download/ssseg_mcibiplusplus/mcibiplusplus_upernet_resnet50os8_voc.pth) | [log](https://github.com/SegmentationBLWX/modelstore/releases/download/ssseg_mcibiplusplus/mcibiplusplus_upernet_resnet50os8_voc.log) |
| UperNet | ImageNet-1k-224x224 | R-101-D8 | 512x512 | LR/POLICY/BS/EPOCH: 0.01/poly/16/60 | train/val | 80.42 | [cfg](https://raw.githubusercontent.com/SegmentationBLWX/sssegmentation/main/ssseg/configs/mcibiplusplus/mcibiplusplus_upernet_resnet101os8_voc.py) | [model](https://github.com/SegmentationBLWX/modelstore/releases/download/ssseg_mcibiplusplus/mcibiplusplus_upernet_resnet101os8_voc.pth) | [log](https://github.com/SegmentationBLWX/modelstore/releases/download/ssseg_mcibiplusplus/mcibiplusplus_upernet_resnet101os8_voc.pth) |
#### PASCAL-Context-59
| Segmentor | Pretrain | Backbone | Crop Size | Schedule | Train/Eval Set | mIoU/mIoU (ms+flip) | Download |
| :-: | :-: | :-: | :-: | :-: | :-: | :-: | :-: |
| UperNet | ImageNet-1k-224x224 | R-101-D8 | 480x480 | LR/POLICY/BS/EPOCH: 0.004/poly/16/260 | train/val | 55.63%/56.82% | [cfg](https://raw.githubusercontent.com/SegmentationBLWX/sssegmentation/main/ssseg/configs/mcibiplusplus/mcibiplusplus_upernet_resnet101os8_pascalcontext59.py) | [model](https://github.com/SegmentationBLWX/modelstore/releases/download/ssseg_mcibiplusplus/mcibiplusplus_upernet_resnet101os8_pascalcontext59.pth) | [log](https://github.com/SegmentationBLWX/modelstore/releases/download/ssseg_mcibiplusplus/mcibiplusplus_upernet_resnet101os8_pascalcontext59.log) |
| UperNet | ImageNet-1k-224x224 | S-101-D8 | 480x480 | LR/POLICY/BS/EPOCH: 0.004/poly/16/260 | train/val | 56.83%/57.92% | [cfg](https://raw.githubusercontent.com/SegmentationBLWX/sssegmentation/main/ssseg/configs/mcibiplusplus/mcibiplusplus_upernet_resnest101os8_pascalcontext59.py) | [model](https://github.com/SegmentationBLWX/modelstore/releases/download/ssseg_mcibiplusplus/mcibiplusplus_upernet_resnest101os8_pascalcontext59.pth) | [log](https://github.com/SegmentationBLWX/modelstore/releases/download/ssseg_mcibiplusplus/mcibiplusplus_upernet_resnest101os8_pascalcontext59.log) |
| UperNet | ImageNet-22k-384x384 | Swin-Large | 480x480 | LR/POLICY/BS/EPOCH: 0.00006/poly/16/260 | train/val | 62.37%/64.01% | [cfg](https://raw.githubusercontent.com/SegmentationBLWX/sssegmentation/main/ssseg/configs/mcibiplusplus/mcibiplusplus_upernet_swinlarge_pascalcontext59.py) | [model](https://github.com/SegmentationBLWX/modelstore/releases/download/ssseg_mcibiplusplus/mcibiplusplus_upernet_swinlarge_pascalcontext59.pth) | [log](https://github.com/SegmentationBLWX/modelstore/releases/download/ssseg_mcibiplusplus/mcibiplusplus_upernet_swinlarge_pascalcontext59.log) |
#### LIP
| Segmentor | Pretrain | Backbone | Crop Size | Schedule | Train/Eval Set | mIoU/mIoU (flip)/mIoU (ms+flip) | Download |
| :-: | :-: | :-: | :-: | :-: | :-: | :-: | :-: |
| UperNet | ImageNet-1k-224x224 | R-101-D8 | 473x473 | LR/POLICY/BS/EPOCH: 0.01/poly/32/150 | train/val | 55.87%/56.26%/56.32% | [cfg](https://raw.githubusercontent.com/SegmentationBLWX/sssegmentation/main/ssseg/configs/mcibiplusplus/mcibiplusplus_upernet_resnet101os8_lip.py) | [model](https://github.com/SegmentationBLWX/modelstore/releases/download/ssseg_mcibiplusplus/mcibiplusplus_upernet_resnet101os8_lip.pth) | [log](https://github.com/SegmentationBLWX/modelstore/releases/download/ssseg_mcibiplusplus/mcibiplusplus_upernet_resnet101os8_lip.log) |
| UperNet | ImageNet-1k-224x224 | S-101-D8 | 473x473 | LR/POLICY/BS/EPOCH: 0.01/poly/32/150 | train/val | 56.57%/56.77%/57.08% | [cfg](https://raw.githubusercontent.com/SegmentationBLWX/sssegmentation/main/ssseg/configs/mcibiplusplus/mcibiplusplus_upernet_resnest101os8_lip.py) | [model](https://github.com/SegmentationBLWX/modelstore/releases/download/ssseg_mcibiplusplus/mcibiplusplus_upernet_resnest101os8_lip.pth) | [log](https://github.com/SegmentationBLWX/modelstore/releases/download/ssseg_mcibiplusplus/mcibiplusplus_upernet_resnest101os8_lip.log) |
| UperNet | ImageNet-22k-384x384 | Swin-Large | 473x473 | LR/POLICY/BS/EPOCH: 0.00006/poly/16/110 | train/val | 59.58%/59.89%/59.91% | [cfg](https://raw.githubusercontent.com/SegmentationBLWX/sssegmentation/main/ssseg/configs/mcibiplusplus/mcibiplusplus_upernet_swinlarge_lip.py) | [model](https://github.com/SegmentationBLWX/modelstore/releases/download/ssseg_mcibiplusplus/mcibiplusplus_upernet_swinlarge_lip.pth) | [log](https://github.com/SegmentationBLWX/modelstore/releases/download/ssseg_mcibiplusplus/mcibiplusplus_upernet_swinlarge_lip.log) |
| DeepLabV3 | ImageNet-1k-224x224 | HRNetV2p-W48 | 473x473 | LR/POLICY/BS/EPOCH: 0.007/poly/40/150 | train/val | 56.70%/57.27%/57.42% | [cfg](https://raw.githubusercontent.com/SegmentationBLWX/sssegmentation/main/ssseg/configs/mcibiplusplus/mcibiplusplus_aspp_hrnetv2w48_lip.py) | [model](https://github.com/SegmentationBLWX/modelstore/releases/download/ssseg_mcibiplusplus/mcibiplusplus_aspp_hrnetv2w48_lip.pth) | [log](https://github.com/SegmentationBLWX/modelstore/releases/download/ssseg_mcibiplusplus/mcibiplusplus_aspp_hrnetv2w48_lip.log) |
#### ADE20k
| Segmentor | Pretrain | Backbone | Crop Size | Schedule | Train/Eval Set | mIoU/mIoU (ms+flip) | Download |
| :-: | :-: | :-: | :-: | :-: | :-: | :-: | :-: |
| UperNet | ImageNet-1k-224x224 | R-101-D8 | 512x512 | LR/POLICY/BS/EPOCH: 0.01/poly/16/130 | train/val | 46.38%/47.93% | [cfg](https://raw.githubusercontent.com/SegmentationBLWX/sssegmentation/main/ssseg/configs/mcibiplusplus/mcibiplusplus_upernet_resnet101os8_ade20k.py) | [model](https://github.com/SegmentationBLWX/modelstore/releases/download/ssseg_mcibiplusplus/mcibiplusplus_upernet_resnet101os8_ade20k.pth) | [log](https://github.com/SegmentationBLWX/modelstore/releases/download/ssseg_mcibiplusplus/mcibiplusplus_upernet_resnet101os8_ade20k.log) |
| UperNet | ImageNet-1k-224x224 | S-101-D8 | 512x512 | LR/POLICY/BS/EPOCH: 0.004/poly/16/180 | train/val | 47.59%/48.56% | [cfg](https://raw.githubusercontent.com/SegmentationBLWX/sssegmentation/main/ssseg/configs/mcibiplusplus/mcibiplusplus_upernet_resnest101os8_ade20k.py) | [model](https://github.com/SegmentationBLWX/modelstore/releases/download/ssseg_mcibiplusplus/mcibiplusplus_upernet_resnest101os8_ade20k.pth) | [log](https://github.com/SegmentationBLWX/modelstore/releases/download/ssseg_mcibiplusplus/mcibiplusplus_upernet_resnest101os8_ade20k.log) |
| UperNet | ImageNet-22k-384x384 | Swin-Large | 640x640 | LR/POLICY/BS/EPOCH: 0.00006/poly/16/130 | train/val | 53.48%/54.50% | [cfg](https://raw.githubusercontent.com/SegmentationBLWX/sssegmentation/main/ssseg/configs/mcibiplusplus/mcibiplusplus_upernet_swinlarge_ade20k.py) | [model](https://github.com/SegmentationBLWX/modelstore/releases/download/ssseg_mcibiplusplus/mcibiplusplus_upernet_swinlarge_ade20k.pth) | [log](https://github.com/SegmentationBLWX/modelstore/releases/download/ssseg_mcibiplusplus/mcibiplusplus_upernet_swinlarge_ade20k.log) |
#### CityScapes
| Segmentor | Pretrain | Backbone | Crop Size | Schedule | Train/Eval Set | mIoU (ms+flip) | Download |
| :-: | :-: | :-: | :-: | :-: | :-: | :-: | :-: |
| DeepLabV3 | ImageNet-1k-224x224 | R-101-D8 | 512x1024 | LR/POLICY/BS/EPOCH: 0.01/poly/16/440 | trainval/test | 82.20% | [cfg](https://raw.githubusercontent.com/SegmentationBLWX/sssegmentation/main/ssseg/configs/mcibiplusplus/mcibiplusplus_aspp_resnet101os8_cityscapes.py) | [model](https://github.com/SegmentationBLWX/modelstore/releases/download/ssseg_mcibiplusplus/mcibiplusplus_aspp_resnet101os8_cityscapes.pth) | [log](https://github.com/SegmentationBLWX/modelstore/releases/download/ssseg_mcibiplusplus/mcibiplusplus_aspp_resnet101os8_cityscapes.log) |
| DeepLabV3 | ImageNet-1k-224x224 | S-101-D8 | 512x1024 | LR/POLICY/BS/EPOCH: 0.004/poly/16/440 | trainval/test | 81.70% | [cfg](https://raw.githubusercontent.com/SegmentationBLWX/sssegmentation/main/ssseg/configs/mcibiplusplus/mcibiplusplus_aspp_resnest101os8_cityscapes.py) | [model](https://github.com/SegmentationBLWX/modelstore/releases/download/ssseg_mcibiplusplus/mcibiplusplus_aspp_resnest101os8_cityscapes.pth) | [log](https://github.com/SegmentationBLWX/modelstore/releases/download/ssseg_mcibiplusplus/mcibiplusplus_aspp_resnest101os8_cityscapes.log) |
| DeepLabV3 | ImageNet-1k-224x224 | HRNetV2p-W48 | 512x1024 | LR/POLICY/BS/EPOCH: 0.01/poly/16/440 | trainval/test | 82.74% | [cfg](https://raw.githubusercontent.com/SegmentationBLWX/sssegmentation/main/ssseg/configs/mcibiplusplus/mcibiplusplus_aspp_hrnetv2w48_cityscapes.py) | [model](https://github.com/SegmentationBLWX/modelstore/releases/download/ssseg_mcibiplusplus/mcibiplusplus_aspp_hrnetv2w48_cityscapes.pth) | [log](https://github.com/SegmentationBLWX/modelstore/releases/download/ssseg_mcibiplusplus/mcibiplusplus_aspp_hrnetv2w48_cityscapes.log) |
#### COCOStuff-10k
| Segmentor | Pretrain | Backbone | Crop Size | Schedule | Train/Eval Set | mIoU/mIoU (ms+flip) | Download |
| :-: | :-: | :-: | :-: | :-: | :-: | :-: | :-: |
| UperNet | ImageNet-1k-224x224 | R-101-D8 | 512x512 | LR/POLICY/BS/EPOCH: 0.001/poly/16/110 | train/test | 40.41%/41.84% | [cfg](https://raw.githubusercontent.com/SegmentationBLWX/sssegmentation/main/ssseg/configs/mcibiplusplus/mcibiplusplus_upernet_resnet101os8_cocostuff10k.py) | [model](https://github.com/SegmentationBLWX/modelstore/releases/download/ssseg_mcibiplusplus/mcibiplusplus_upernet_resnet101os8_cocostuff10k.pth) | [log](https://github.com/SegmentationBLWX/modelstore/releases/download/ssseg_mcibiplusplus/mcibiplusplus_upernet_resnet101os8_cocostuff10k.log) |
| UperNet | ImageNet-1k-224x224 | S-101-D8 | 512x512 | LR/POLICY/BS/EPOCH: 0.001/poly/32/150 | train/test | 41.81%/42.71% | [cfg](https://raw.githubusercontent.com/SegmentationBLWX/sssegmentation/main/ssseg/configs/mcibiplusplus/mcibiplusplus_upernet_resnest101os8_cocostuff10k.py) | [model](https://github.com/SegmentationBLWX/modelstore/releases/download/ssseg_mcibiplusplus/mcibiplusplus_upernet_resnest101os8_cocostuff10k.pth) | [log](https://github.com/SegmentationBLWX/modelstore/releases/download/ssseg_mcibiplusplus/mcibiplusplus_upernet_resnest101os8_cocostuff10k.log) |
| UperNet | ImageNet-22k-384x384 | Swin-Large | 512x512 | LR/POLICY/BS/EPOCH: 0.00006/poly/16/150 | train/test | 49.11%/50.27% | [cfg](https://raw.githubusercontent.com/SegmentationBLWX/sssegmentation/main/ssseg/configs/mcibiplusplus/mcibiplusplus_upernet_swinlarge_cocostuff10k.py) | [model](https://github.com/SegmentationBLWX/modelstore/releases/download/ssseg_mcibiplusplus/mcibiplusplus_upernet_swinlarge_cocostuff10k.pth) | [log](https://github.com/SegmentationBLWX/modelstore/releases/download/ssseg_mcibiplusplus/mcibiplusplus_upernet_swinlarge_cocostuff10k.log) |
## More
You can also download the model weights from following sources:
- BaiduNetdisk: https://pan.baidu.com/s/1gD-NJJWOtaHCtB0qHE79rA with access code **s757**
In addition, in this repo, all of models above are evaluated on A100 rather than V100 mentioned in our original paper, thus the performance here will be slightly different from the reported results in the original paper.