admin/Seg_Data_Server
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

first commit

Browse Source
This commit is contained in:
admin
2026-05-20 15:05:35 +08:00
commit ac09b26253
2048 changed files with 189478 additions and 0 deletions
Download Patch File Download Diff File Expand all files Collapse all files

54
Seg_All_In_One_MMSeg/configs/resnest/README.md Normal file
View File

@@ -0,0 +1,54 @@
# ResNeSt
> [ResNeSt: Split-Attention Networks](https://arxiv.org/abs/2004.08955)
## Introduction
<!-- [BACKBONE] -->
<a href="https://github.com/zhanghang1989/ResNeSt">Official Repo</a>
<a href="https://github.com/open-mmlab/mmsegmentation/blob/v0.17.0/mmseg/models/backbones/resnest.py#L271">Code Snippet</a>
## Abstract
<!-- [ABSTRACT] -->
It is well known that featuremap attention and multi-path representation are important for visual recognition. In this paper, we present a modularized architecture, which applies the channel-wise attention on different network branches to leverage their success in capturing cross-feature interactions and learning diverse representations. Our design results in a simple and unified computation block, which can be parameterized using only a few variables. Our model, named ResNeSt, outperforms EfficientNet in accuracy and latency trade-off on image classification. In addition, ResNeSt has achieved superior transfer learning results on several public benchmarks serving as the backbone, and has been adopted by the winning entries of COCO-LVIS challenge. The source code for complete system and pretrained models are publicly available.
<!-- [IMAGE] -->
<div align=center>
<img src="https://user-images.githubusercontent.com/24582831/142902526-3cf33345-7e40-47a6-985e-4381857e21df.png" width="60%"/>
</div>
## Results and models
### Cityscapes
| Method | Backbone | Crop Size | Lr schd | Mem (GB) | Inf time (fps) | Device | mIoU | mIoU(ms+flip) | config | download |
| ---------- | -------- | --------- | ------: | -------: | -------------- | ------ | ----: | ------------- | ---------------------------------------------------------------------------------------------------------------------------------------------- | -------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
| FCN | S-101-D8 | 512x1024 | 80000 | 11.4 | 2.39 | V100 | 77.56 | 78.98 | [config](https://github.com/open-mmlab/mmsegmentation/blob/main/configs/resnest/resnest_s101-d8_fcn_4xb2-80k_cityscapes-512x1024.py) | [model](https://download.openmmlab.com/mmsegmentation/v0.5/resnest/fcn_s101-d8_512x1024_80k_cityscapes/fcn_s101-d8_512x1024_80k_cityscapes_20200807_140631-f8d155b3.pth) \| [log](https://download.openmmlab.com/mmsegmentation/v0.5/resnest/fcn_s101-d8_512x1024_80k_cityscapes/fcn_s101-d8_512x1024_80k_cityscapes-20200807_140631.log.json) |
| PSPNet | S-101-D8 | 512x1024 | 80000 | 11.8 | 2.52 | V100 | 78.57 | 79.19 | [config](https://github.com/open-mmlab/mmsegmentation/blob/main/configs/resnest/resnest_s101-d8_pspnet_4xb2-80k_cityscapes512x1024.py) | [model](https://download.openmmlab.com/mmsegmentation/v0.5/resnest/pspnet_s101-d8_512x1024_80k_cityscapes/pspnet_s101-d8_512x1024_80k_cityscapes_20200807_140631-c75f3b99.pth) \| [log](https://download.openmmlab.com/mmsegmentation/v0.5/resnest/pspnet_s101-d8_512x1024_80k_cityscapes/pspnet_s101-d8_512x1024_80k_cityscapes-20200807_140631.log.json) |
| DeepLabV3 | S-101-D8 | 512x1024 | 80000 | 11.9 | 1.88 | V100 | 79.67 | 80.51 | [config](https://github.com/open-mmlab/mmsegmentation/blob/main/configs/resnest/resnest_s101-d8_deeplabv3_4xb2-80k_cityscapes-512x1024.py) | [model](https://download.openmmlab.com/mmsegmentation/v0.5/resnest/deeplabv3_s101-d8_512x1024_80k_cityscapes/deeplabv3_s101-d8_512x1024_80k_cityscapes_20200807_144429-b73c4270.pth) \| [log](https://download.openmmlab.com/mmsegmentation/v0.5/resnest/deeplabv3_s101-d8_512x1024_80k_cityscapes/deeplabv3_s101-d8_512x1024_80k_cityscapes-20200807_144429.log.json) |
| DeepLabV3+ | S-101-D8 | 512x1024 | 80000 | 13.2 | 2.36 | V100 | 79.62 | 80.27 | [config](https://github.com/open-mmlab/mmsegmentation/blob/main/configs/resnest/resnest_s101-d8_deeplabv3plus_4xb2-80k_cityscapes-512x1024.py) | [model](https://download.openmmlab.com/mmsegmentation/v0.5/resnest/deeplabv3plus_s101-d8_512x1024_80k_cityscapes/deeplabv3plus_s101-d8_512x1024_80k_cityscapes_20200807_144429-1239eb43.pth) \| [log](https://download.openmmlab.com/mmsegmentation/v0.5/resnest/deeplabv3plus_s101-d8_512x1024_80k_cityscapes/deeplabv3plus_s101-d8_512x1024_80k_cityscapes-20200807_144429.log.json) |
### ADE20K
| Method | Backbone | Crop Size | Lr schd | Mem (GB) | Inf time (fps) | Device | mIoU | mIoU(ms+flip) | config | download |
| ---------- | -------- | --------- | ------: | -------: | -------------- | ------ | ----: | ------------- | ------------------------------------------------------------------------------------------------------------------------------------------ | ---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
| FCN | S-101-D8 | 512x512 | 160000 | 14.2 | 12.86 | V100 | 45.62 | 46.16 | [config](https://github.com/open-mmlab/mmsegmentation/blob/main/configs/resnest/resnest_s101-d8_fcn_4xb4-160k_ade20k-512x512.py) | [model](https://download.openmmlab.com/mmsegmentation/v0.5/resnest/fcn_s101-d8_512x512_160k_ade20k/fcn_s101-d8_512x512_160k_ade20k_20200807_145416-d3160329.pth) \| [log](https://download.openmmlab.com/mmsegmentation/v0.5/resnest/fcn_s101-d8_512x512_160k_ade20k/fcn_s101-d8_512x512_160k_ade20k-20200807_145416.log.json) |
| PSPNet | S-101-D8 | 512x512 | 160000 | 14.2 | 13.02 | V100 | 45.44 | 46.28 | [config](https://github.com/open-mmlab/mmsegmentation/blob/main/configs/resnest/resnest_s101-d8_pspnet_4xb4-160k_ade20k-512x512.py) | [model](https://download.openmmlab.com/mmsegmentation/v0.5/resnest/pspnet_s101-d8_512x512_160k_ade20k/pspnet_s101-d8_512x512_160k_ade20k_20200807_145416-a6daa92a.pth) \| [log](https://download.openmmlab.com/mmsegmentation/v0.5/resnest/pspnet_s101-d8_512x512_160k_ade20k/pspnet_s101-d8_512x512_160k_ade20k-20200807_145416.log.json) |
| DeepLabV3 | S-101-D8 | 512x512 | 160000 | 14.6 | 9.28 | V100 | 45.71 | 46.59 | [config](https://github.com/open-mmlab/mmsegmentation/blob/main/configs/resnest/resnest_s101-d8_deeplabv3_4xb4-160k_ade20k-512x512.py) | [model](https://download.openmmlab.com/mmsegmentation/v0.5/resnest/deeplabv3_s101-d8_512x512_160k_ade20k/deeplabv3_s101-d8_512x512_160k_ade20k_20200807_144503-17ecabe5.pth) \| [log](https://download.openmmlab.com/mmsegmentation/v0.5/resnest/deeplabv3_s101-d8_512x512_160k_ade20k/deeplabv3_s101-d8_512x512_160k_ade20k-20200807_144503.log.json) |
| DeepLabV3+ | S-101-D8 | 512x512 | 160000 | 16.2 | 11.96 | V100 | 46.47 | 47.27 | [config](https://github.com/open-mmlab/mmsegmentation/blob/main/configs/resnest/resnest_s101-d8_deeplabv3plus_4xb4-160k_ade20k-512x512.py) | [model](https://download.openmmlab.com/mmsegmentation/v0.5/resnest/deeplabv3plus_s101-d8_512x512_160k_ade20k/deeplabv3plus_s101-d8_512x512_160k_ade20k_20200807_144503-27b26226.pth) \| [log](https://download.openmmlab.com/mmsegmentation/v0.5/resnest/deeplabv3plus_s101-d8_512x512_160k_ade20k/deeplabv3plus_s101-d8_512x512_160k_ade20k-20200807_144503.log.json) |
## Citation
```bibtex
@article{zhang2020resnest,
title={ResNeSt: Split-Attention Networks},
author={Zhang, Hang and Wu, Chongruo and Zhang, Zhongyue and Zhu, Yi and Zhang, Zhi and Lin, Haibin and Sun, Yue and He, Tong and Muller, Jonas and Manmatha, R. and Li, Mu and Smola, Alexander},
journal={arXiv preprint arXiv:2004.08955},
year={2020}
}
```

193
Seg_All_In_One_MMSeg/configs/resnest/metafile.yaml Normal file
View File

@@ -0,0 +1,193 @@
Models:
- Name: resnest_s101-d8_fcn_4xb2-80k_cityscapes-512x1024
In Collection: FCN
Results:
Task: Semantic Segmentation
Dataset: Cityscapes
Metrics:
mIoU: 77.56
mIoU(ms+flip): 78.98
Config: configs/resnest/resnest_s101-d8_fcn_4xb2-80k_cityscapes-512x1024.py
Metadata:
Training Data: Cityscapes
Batch Size: 8
Architecture:
- S-101-D8
- FCN
Training Resources: 4x V100 GPUS
Memory (GB): 11.4
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/resnest/fcn_s101-d8_512x1024_80k_cityscapes/fcn_s101-d8_512x1024_80k_cityscapes_20200807_140631-f8d155b3.pth
Training log: https://download.openmmlab.com/mmsegmentation/v0.5/resnest/fcn_s101-d8_512x1024_80k_cityscapes/fcn_s101-d8_512x1024_80k_cityscapes-20200807_140631.log.json
Paper:
Title: 'ResNeSt: Split-Attention Networks'
URL: https://arxiv.org/abs/2004.08955
Code: https://github.com/open-mmlab/mmsegmentation/blob/v0.17.0/mmseg/models/backbones/resnest.py#L271
Framework: PyTorch
- Name: resnest_s101-d8_pspnet_4xb2-80k_cityscapes512x1024
In Collection: PSPNet
Results:
Task: Semantic Segmentation
Dataset: Cityscapes
Metrics:
mIoU: 78.57
mIoU(ms+flip): 79.19
Config: configs/resnest/resnest_s101-d8_pspnet_4xb2-80k_cityscapes512x1024.py
Metadata:
Training Data: Cityscapes
Batch Size: 8
Architecture:
- S-101-D8
- PSPNet
Training Resources: 4x V100 GPUS
Memory (GB): 11.8
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/resnest/pspnet_s101-d8_512x1024_80k_cityscapes/pspnet_s101-d8_512x1024_80k_cityscapes_20200807_140631-c75f3b99.pth
Training log: https://download.openmmlab.com/mmsegmentation/v0.5/resnest/pspnet_s101-d8_512x1024_80k_cityscapes/pspnet_s101-d8_512x1024_80k_cityscapes-20200807_140631.log.json
Paper:
Title: 'ResNeSt: Split-Attention Networks'
URL: https://arxiv.org/abs/2004.08955
Code: https://github.com/open-mmlab/mmsegmentation/blob/v0.17.0/mmseg/models/backbones/resnest.py#L271
Framework: PyTorch
- Name: resnest_s101-d8_deeplabv3_4xb2-80k_cityscapes-512x1024
In Collection: DeepLabV3
Results:
Task: Semantic Segmentation
Dataset: Cityscapes
Metrics:
mIoU: 79.67
mIoU(ms+flip): 80.51
Config: configs/resnest/resnest_s101-d8_deeplabv3_4xb2-80k_cityscapes-512x1024.py
Metadata:
Training Data: Cityscapes
Batch Size: 8
Architecture:
- S-101-D8
- DeepLabV3
Training Resources: 4x V100 GPUS
Memory (GB): 11.9
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/resnest/deeplabv3_s101-d8_512x1024_80k_cityscapes/deeplabv3_s101-d8_512x1024_80k_cityscapes_20200807_144429-b73c4270.pth
Training log: https://download.openmmlab.com/mmsegmentation/v0.5/resnest/deeplabv3_s101-d8_512x1024_80k_cityscapes/deeplabv3_s101-d8_512x1024_80k_cityscapes-20200807_144429.log.json
Paper:
Title: 'ResNeSt: Split-Attention Networks'
URL: https://arxiv.org/abs/2004.08955
Code: https://github.com/open-mmlab/mmsegmentation/blob/v0.17.0/mmseg/models/backbones/resnest.py#L271
Framework: PyTorch
- Name: resnest_s101-d8_deeplabv3plus_4xb2-80k_cityscapes-512x1024
In Collection: DeepLabV3+
Results:
Task: Semantic Segmentation
Dataset: Cityscapes
Metrics:
mIoU: 79.62
mIoU(ms+flip): 80.27
Config: configs/resnest/resnest_s101-d8_deeplabv3plus_4xb2-80k_cityscapes-512x1024.py
Metadata:
Training Data: Cityscapes
Batch Size: 8
Architecture:
- S-101-D8
- DeepLabV3+
Training Resources: 4x V100 GPUS
Memory (GB): 13.2
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/resnest/deeplabv3plus_s101-d8_512x1024_80k_cityscapes/deeplabv3plus_s101-d8_512x1024_80k_cityscapes_20200807_144429-1239eb43.pth
Training log: https://download.openmmlab.com/mmsegmentation/v0.5/resnest/deeplabv3plus_s101-d8_512x1024_80k_cityscapes/deeplabv3plus_s101-d8_512x1024_80k_cityscapes-20200807_144429.log.json
Paper:
Title: 'ResNeSt: Split-Attention Networks'
URL: https://arxiv.org/abs/2004.08955
Code: https://github.com/open-mmlab/mmsegmentation/blob/v0.17.0/mmseg/models/backbones/resnest.py#L271
Framework: PyTorch
- Name: resnest_s101-d8_fcn_4xb4-160k_ade20k-512x512
In Collection: FCN
Results:
Task: Semantic Segmentation
Dataset: ADE20K
Metrics:
mIoU: 45.62
mIoU(ms+flip): 46.16
Config: configs/resnest/resnest_s101-d8_fcn_4xb4-160k_ade20k-512x512.py
Metadata:
Training Data: ADE20K
Batch Size: 16
Architecture:
- S-101-D8
- FCN
Training Resources: 4x V100 GPUS
Memory (GB): 14.2
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/resnest/fcn_s101-d8_512x512_160k_ade20k/fcn_s101-d8_512x512_160k_ade20k_20200807_145416-d3160329.pth
Training log: https://download.openmmlab.com/mmsegmentation/v0.5/resnest/fcn_s101-d8_512x512_160k_ade20k/fcn_s101-d8_512x512_160k_ade20k-20200807_145416.log.json
Paper:
Title: 'ResNeSt: Split-Attention Networks'
URL: https://arxiv.org/abs/2004.08955
Code: https://github.com/open-mmlab/mmsegmentation/blob/v0.17.0/mmseg/models/backbones/resnest.py#L271
Framework: PyTorch
- Name: resnest_s101-d8_pspnet_4xb4-160k_ade20k-512x512
In Collection: PSPNet
Results:
Task: Semantic Segmentation
Dataset: ADE20K
Metrics:
mIoU: 45.44
mIoU(ms+flip): 46.28
Config: configs/resnest/resnest_s101-d8_pspnet_4xb4-160k_ade20k-512x512.py
Metadata:
Training Data: ADE20K
Batch Size: 16
Architecture:
- S-101-D8
- PSPNet
Training Resources: 4x V100 GPUS
Memory (GB): 14.2
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/resnest/pspnet_s101-d8_512x512_160k_ade20k/pspnet_s101-d8_512x512_160k_ade20k_20200807_145416-a6daa92a.pth
Training log: https://download.openmmlab.com/mmsegmentation/v0.5/resnest/pspnet_s101-d8_512x512_160k_ade20k/pspnet_s101-d8_512x512_160k_ade20k-20200807_145416.log.json
Paper:
Title: 'ResNeSt: Split-Attention Networks'
URL: https://arxiv.org/abs/2004.08955
Code: https://github.com/open-mmlab/mmsegmentation/blob/v0.17.0/mmseg/models/backbones/resnest.py#L271
Framework: PyTorch
- Name: resnest_s101-d8_deeplabv3_4xb4-160k_ade20k-512x512
In Collection: DeepLabV3
Results:
Task: Semantic Segmentation
Dataset: ADE20K
Metrics:
mIoU: 45.71
mIoU(ms+flip): 46.59
Config: configs/resnest/resnest_s101-d8_deeplabv3_4xb4-160k_ade20k-512x512.py
Metadata:
Training Data: ADE20K
Batch Size: 16
Architecture:
- S-101-D8
- DeepLabV3
Training Resources: 4x V100 GPUS
Memory (GB): 14.6
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/resnest/deeplabv3_s101-d8_512x512_160k_ade20k/deeplabv3_s101-d8_512x512_160k_ade20k_20200807_144503-17ecabe5.pth
Training log: https://download.openmmlab.com/mmsegmentation/v0.5/resnest/deeplabv3_s101-d8_512x512_160k_ade20k/deeplabv3_s101-d8_512x512_160k_ade20k-20200807_144503.log.json
Paper:
Title: 'ResNeSt: Split-Attention Networks'
URL: https://arxiv.org/abs/2004.08955
Code: https://github.com/open-mmlab/mmsegmentation/blob/v0.17.0/mmseg/models/backbones/resnest.py#L271
Framework: PyTorch
- Name: resnest_s101-d8_deeplabv3plus_4xb4-160k_ade20k-512x512
In Collection: DeepLabV3+
Results:
Task: Semantic Segmentation
Dataset: ADE20K
Metrics:
mIoU: 46.47
mIoU(ms+flip): 47.27
Config: configs/resnest/resnest_s101-d8_deeplabv3plus_4xb4-160k_ade20k-512x512.py
Metadata:
Training Data: ADE20K
Batch Size: 16
Architecture:
- S-101-D8
- DeepLabV3+
Training Resources: 4x V100 GPUS
Memory (GB): 16.2
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/resnest/deeplabv3plus_s101-d8_512x512_160k_ade20k/deeplabv3plus_s101-d8_512x512_160k_ade20k_20200807_144503-27b26226.pth
Training log: https://download.openmmlab.com/mmsegmentation/v0.5/resnest/deeplabv3plus_s101-d8_512x512_160k_ade20k/deeplabv3plus_s101-d8_512x512_160k_ade20k-20200807_144503.log.json
Paper:
Title: 'ResNeSt: Split-Attention Networks'
URL: https://arxiv.org/abs/2004.08955
Code: https://github.com/open-mmlab/mmsegmentation/blob/v0.17.0/mmseg/models/backbones/resnest.py#L271
Framework: PyTorch

9
Seg_All_In_One_MMSeg/configs/resnest/resnest_s101-d8_deeplabv3_4xb2-80k_cityscapes-512x1024.py Normal file
View File

@@ -0,0 +1,9 @@
_base_ = '../deeplabv3/deeplabv3_r101-d8_4xb2-80k_cityscapes-512x1024.py'
model = dict(
pretrained='open-mmlab://resnest101',
backbone=dict(
type='ResNeSt',
stem_channels=128,
radix=2,
reduction_factor=4,
avg_down_stride=True))

9
Seg_All_In_One_MMSeg/configs/resnest/resnest_s101-d8_deeplabv3_4xb4-160k_ade20k-512x512.py Normal file
View File

@@ -0,0 +1,9 @@
_base_ = '../deeplabv3/deeplabv3_r101-d8_4xb4-160k_ade20k-512x512.py'
model = dict(
pretrained='open-mmlab://resnest101',
backbone=dict(
type='ResNeSt',
stem_channels=128,
radix=2,
reduction_factor=4,
avg_down_stride=True))

9
Seg_All_In_One_MMSeg/configs/resnest/resnest_s101-d8_deeplabv3plus_4xb2-80k_cityscapes-512x1024.py Normal file
View File

@@ -0,0 +1,9 @@
_base_ = '../deeplabv3plus/deeplabv3plus_r101-d8_4xb2-80k_cityscapes-512x1024.py' # noqa
model = dict(
pretrained='open-mmlab://resnest101',
backbone=dict(
type='ResNeSt',
stem_channels=128,
radix=2,
reduction_factor=4,
avg_down_stride=True))

9
Seg_All_In_One_MMSeg/configs/resnest/resnest_s101-d8_deeplabv3plus_4xb4-160k_ade20k-512x512.py Normal file
View File

@@ -0,0 +1,9 @@
_base_ = '../deeplabv3plus/deeplabv3plus_r101-d8_4xb4-160k_ade20k-512x512.py'
model = dict(
pretrained='open-mmlab://resnest101',
backbone=dict(
type='ResNeSt',
stem_channels=128,
radix=2,
reduction_factor=4,
avg_down_stride=True))

9
Seg_All_In_One_MMSeg/configs/resnest/resnest_s101-d8_fcn_4xb2-80k_cityscapes-512x1024.py Normal file
View File

@@ -0,0 +1,9 @@
_base_ = '../fcn/fcn_r101-d8_4xb2-80k_cityscapes-512x1024.py'
model = dict(
pretrained='open-mmlab://resnest101',
backbone=dict(
type='ResNeSt',
stem_channels=128,
radix=2,
reduction_factor=4,
avg_down_stride=True))

9
Seg_All_In_One_MMSeg/configs/resnest/resnest_s101-d8_fcn_4xb4-160k_ade20k-512x512.py Normal file
View File

@@ -0,0 +1,9 @@
_base_ = '../fcn/fcn_r101-d8_4xb4-160k_ade20k-512x512.py'
model = dict(
pretrained='open-mmlab://resnest101',
backbone=dict(
type='ResNeSt',
stem_channels=128,
radix=2,
reduction_factor=4,
avg_down_stride=True))

9
Seg_All_In_One_MMSeg/configs/resnest/resnest_s101-d8_pspnet_4xb2-80k_cityscapes512x1024.py Normal file
View File

@@ -0,0 +1,9 @@
_base_ = '../pspnet/pspnet_r101-d8_4xb2-80k_cityscapes-512x1024.py'
model = dict(
pretrained='open-mmlab://resnest101',
backbone=dict(
type='ResNeSt',
stem_channels=128,
radix=2,
reduction_factor=4,
avg_down_stride=True))

9
Seg_All_In_One_MMSeg/configs/resnest/resnest_s101-d8_pspnet_4xb4-160k_ade20k-512x512.py Normal file
View File

@@ -0,0 +1,9 @@
_base_ = '../pspnet/pspnet_r101-d8_4xb4-160k_ade20k-512x512.py'
model = dict(
pretrained='open-mmlab://resnest101',
backbone=dict(
type='ResNeSt',
stem_channels=128,
radix=2,
reduction_factor=4,
avg_down_stride=True))