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

62
Seg_All_In_One_MMSeg/configs/dnlnet/README.md Normal file
View File

@@ -0,0 +1,62 @@
# DNLNet
> [Disentangled Non-Local Neural Networks](https://arxiv.org/abs/2006.06668)
## Introduction
<!-- [ALGORITHM] -->
<a href="https://github.com/yinmh17/DNL-Semantic-Segmentation">Official Repo</a>
<a href="https://github.com/open-mmlab/mmsegmentation/blob/v0.17.0/mmseg/models/decode_heads/dnl_head.py#L88">Code Snippet</a>
## Abstract
<!-- [ABSTRACT] -->
The non-local block is a popular module for strengthening the context modeling ability of a regular convolutional neural network. This paper first studies the non-local block in depth, where we find that its attention computation can be split into two terms, a whitened pairwise term accounting for the relationship between two pixels and a unary term representing the saliency of every pixel. We also observe that the two terms trained alone tend to model different visual clues, e.g. the whitened pairwise term learns within-region relationships while the unary term learns salient boundaries. However, the two terms are tightly coupled in the non-local block, which hinders the learning of each. Based on these findings, we present the disentangled non-local block, where the two terms are decoupled to facilitate learning for both terms. We demonstrate the effectiveness of the decoupled design on various tasks, such as semantic segmentation on Cityscapes, ADE20K and PASCAL Context, object detection on COCO, and action recognition on Kinetics.
<!-- [IMAGE] -->
<div align=center>
<img src="https://user-images.githubusercontent.com/24582831/142900944-b8d93301-d2ce-488e-a461-b0813f96be49.png" width="70%"/>
</div>
## Results and models (in progress)
### Cityscapes
| Method | Backbone | Crop Size | Lr schd | Mem (GB) | Inf time (fps) | Device | mIoU | mIoU(ms+flip) | config | download |
| ------ | -------- | --------- | ------: | -------: | -------------- | ------ | ----: | ------------- | --------------------------------------------------------------------------------------------------------------------------- | -------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
| DNLNet | R-50-D8 | 512x1024 | 40000 | 7.3 | 2.56 | V100 | 78.61 | - | [config](https://github.com/open-mmlab/mmsegmentation/blob/main/configs/dnlnet/dnl_r50-d8_4xb2-40k_cityscapes-512x1024.py) | [model](https://download.openmmlab.com/mmsegmentation/v0.5/dnlnet/dnl_r50-d8_512x1024_40k_cityscapes/dnl_r50-d8_512x1024_40k_cityscapes_20200904_233629-53d4ea93.pth) \| [log](https://download.openmmlab.com/mmsegmentation/v0.5/dnlnet/dnl_r50-d8_512x1024_40k_cityscapes/dnl_r50-d8_512x1024_40k_cityscapes-20200904_233629.log.json) |
| DNLNet | R-101-D8 | 512x1024 | 40000 | 10.9 | 1.96 | V100 | 78.31 | - | [config](https://github.com/open-mmlab/mmsegmentation/blob/main/configs/dnlnet/dnl_r101-d8_4xb2-40k_cityscapes-512x1024.py) | [model](https://download.openmmlab.com/mmsegmentation/v0.5/dnlnet/dnl_r101-d8_512x1024_40k_cityscapes/dnl_r101-d8_512x1024_40k_cityscapes_20200904_233629-9928ffef.pth) \| [log](https://download.openmmlab.com/mmsegmentation/v0.5/dnlnet/dnl_r101-d8_512x1024_40k_cityscapes/dnl_r101-d8_512x1024_40k_cityscapes-20200904_233629.log.json) |
| DNLNet | R-50-D8 | 769x769 | 40000 | 9.2 | 1.50 | V100 | 78.44 | 80.27 | [config](https://github.com/open-mmlab/mmsegmentation/blob/main/configs/dnlnet/dnl_r50-d8_4xb2-40k_cityscapes-769x769.py) | [model](https://download.openmmlab.com/mmsegmentation/v0.5/dnlnet/dnl_r50-d8_769x769_40k_cityscapes/dnl_r50-d8_769x769_40k_cityscapes_20200820_232206-0f283785.pth) \| [log](https://download.openmmlab.com/mmsegmentation/v0.5/dnlnet/dnl_r50-d8_769x769_40k_cityscapes/dnl_r50-d8_769x769_40k_cityscapes-20200820_232206.log.json) |
| DNLNet | R-101-D8 | 769x769 | 40000 | 12.6 | 1.02 | V100 | 76.39 | 77.77 | [config](https://github.com/open-mmlab/mmsegmentation/blob/main/configs/dnlnet/dnl_r101-d8_4xb2-40k_cityscapes-769x769.py) | [model](https://download.openmmlab.com/mmsegmentation/v0.5/dnlnet/dnl_r101-d8_769x769_40k_cityscapes/dnl_r101-d8_769x769_40k_cityscapes_20200820_171256-76c596df.pth) \| [log](https://download.openmmlab.com/mmsegmentation/v0.5/dnlnet/dnl_r101-d8_769x769_40k_cityscapes/dnl_r101-d8_769x769_40k_cityscapes-20200820_171256.log.json) |
| DNLNet | R-50-D8 | 512x1024 | 80000 | - | - | V100 | 79.33 | - | [config](https://github.com/open-mmlab/mmsegmentation/blob/main/configs/dnlnet/dnl_r50-d8_4xb2-80k_cityscapes-512x1024.py) | [model](https://download.openmmlab.com/mmsegmentation/v0.5/dnlnet/dnl_r50-d8_512x1024_80k_cityscapes/dnl_r50-d8_512x1024_80k_cityscapes_20200904_233629-58b2f778.pth) \| [log](https://download.openmmlab.com/mmsegmentation/v0.5/dnlnet/dnl_r50-d8_512x1024_80k_cityscapes/dnl_r50-d8_512x1024_80k_cityscapes-20200904_233629.log.json) |
| DNLNet | R-101-D8 | 512x1024 | 80000 | - | - | V100 | 80.41 | - | [config](https://github.com/open-mmlab/mmsegmentation/blob/main/configs/dnlnet/dnl_r101-d8_4xb2-80k_cityscapes-512x1024.py) | [model](https://download.openmmlab.com/mmsegmentation/v0.5/dnlnet/dnl_r101-d8_512x1024_80k_cityscapes/dnl_r101-d8_512x1024_80k_cityscapes_20200904_233629-758e2dd4.pth) \| [log](https://download.openmmlab.com/mmsegmentation/v0.5/dnlnet/dnl_r101-d8_512x1024_80k_cityscapes/dnl_r101-d8_512x1024_80k_cityscapes-20200904_233629.log.json) |
| DNLNet | R-50-D8 | 769x769 | 80000 | - | - | V100 | 79.36 | 80.70 | [config](https://github.com/open-mmlab/mmsegmentation/blob/main/configs/dnlnet/dnl_r50-d8_4xb2-80k_cityscapes-769x769.py) | [model](https://download.openmmlab.com/mmsegmentation/v0.5/dnlnet/dnl_r50-d8_769x769_80k_cityscapes/dnl_r50-d8_769x769_80k_cityscapes_20200820_011925-366bc4c7.pth) \| [log](https://download.openmmlab.com/mmsegmentation/v0.5/dnlnet/dnl_r50-d8_769x769_80k_cityscapes/dnl_r50-d8_769x769_80k_cityscapes-20200820_011925.log.json) |
| DNLNet | R-101-D8 | 769x769 | 80000 | - | - | V100 | 79.41 | 80.68 | [config](https://github.com/open-mmlab/mmsegmentation/blob/main/configs/dnlnet/dnl_r101-d8_4xb2-80k_cityscapes-769x769.py) | [model](https://download.openmmlab.com/mmsegmentation/v0.5/dnlnet/dnl_r101-d8_769x769_80k_cityscapes/dnl_r101-d8_769x769_80k_cityscapes_20200821_051111-95ff84ab.pth) \| [log](https://download.openmmlab.com/mmsegmentation/v0.5/dnlnet/dnl_r101-d8_769x769_80k_cityscapes/dnl_r101-d8_769x769_80k_cityscapes-20200821_051111.log.json) |
### ADE20K
| Method | Backbone | Crop Size | Lr schd | Mem (GB) | Inf time (fps) | Device | mIoU | mIoU(ms+flip) | config | download |
| ------ | -------- | --------- | ------: | -------: | -------------- | ------ | ----: | ------------- | ----------------------------------------------------------------------------------------------------------------------- | ---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
| DNLNet | R-50-D8 | 512x512 | 80000 | 8.8 | 20.66 | V100 | 41.76 | 42.99 | [config](https://github.com/open-mmlab/mmsegmentation/blob/main/configs/dnlnet/dnl_r50-d8_4xb4-80k_ade20k-512x512.py) | [model](https://download.openmmlab.com/mmsegmentation/v0.5/dnlnet/dnl_r50-d8_512x512_80k_ade20k/dnl_r50-d8_512x512_80k_ade20k_20200826_183354-1cf6e0c1.pth) \| [log](https://download.openmmlab.com/mmsegmentation/v0.5/dnlnet/dnl_r50-d8_512x512_80k_ade20k/dnl_r50-d8_512x512_80k_ade20k-20200826_183354.log.json) |
| DNLNet | R-101-D8 | 512x512 | 80000 | 12.8 | 12.54 | V100 | 43.76 | 44.91 | [config](https://github.com/open-mmlab/mmsegmentation/blob/main/configs/dnlnet/dnl_r101-d8_4xb4-80k_ade20k-512x512.py) | [model](https://download.openmmlab.com/mmsegmentation/v0.5/dnlnet/dnl_r101-d8_512x512_80k_ade20k/dnl_r101-d8_512x512_80k_ade20k_20200826_183354-d820d6ea.pth) \| [log](https://download.openmmlab.com/mmsegmentation/v0.5/dnlnet/dnl_r101-d8_512x512_80k_ade20k/dnl_r101-d8_512x512_80k_ade20k-20200826_183354.log.json) |
| DNLNet | R-50-D8 | 512x512 | 160000 | - | - | V100 | 41.87 | 43.01 | [config](https://github.com/open-mmlab/mmsegmentation/blob/main/configs/dnlnet/dnl_r50-d8_4xb4-160k_ade20k-512x512.py) | [model](https://download.openmmlab.com/mmsegmentation/v0.5/dnlnet/dnl_r50-d8_512x512_160k_ade20k/dnl_r50-d8_512x512_160k_ade20k_20200826_183350-37837798.pth) \| [log](https://download.openmmlab.com/mmsegmentation/v0.5/dnlnet/dnl_r50-d8_512x512_160k_ade20k/dnl_r50-d8_512x512_160k_ade20k-20200826_183350.log.json) |
| DNLNet | R-101-D8 | 512x512 | 160000 | - | - | V100 | 44.25 | 45.78 | [config](https://github.com/open-mmlab/mmsegmentation/blob/main/configs/dnlnet/dnl_r101-d8_4xb4-160k_ade20k-512x512.py) | [model](https://download.openmmlab.com/mmsegmentation/v0.5/dnlnet/dnl_r101-d8_512x512_160k_ade20k/dnl_r101-d8_512x512_160k_ade20k_20200826_183350-ed522c61.pth) \| [log](https://download.openmmlab.com/mmsegmentation/v0.5/dnlnet/dnl_r101-d8_512x512_160k_ade20k/dnl_r101-d8_512x512_160k_ade20k-20200826_183350.log.json) |
## Notes
This example is to reproduce ["Disentangled Non-Local Neural Networks"](https://arxiv.org/abs/2006.06668) for semantic segmentation. It is still in progress.
## Citation
```bibtex
@misc{yin2020disentangled,
title={Disentangled Non-Local Neural Networks},
author={Minghao Yin and Zhuliang Yao and Yue Cao and Xiu Li and Zheng Zhang and Stephen Lin and Han Hu},
year={2020},
booktitle={ECCV}
}
```

2
Seg_All_In_One_MMSeg/configs/dnlnet/dnl_r101-d8_4xb2-40k_cityscapes-512x1024.py Normal file
View File

@@ -0,0 +1,2 @@
_base_ = './dnl_r50-d8_4xb2-40k_cityscapes-512x1024.py'
model = dict(pretrained='open-mmlab://resnet101_v1c', backbone=dict(depth=101))

2
Seg_All_In_One_MMSeg/configs/dnlnet/dnl_r101-d8_4xb2-40k_cityscapes-769x769.py Normal file
View File

@@ -0,0 +1,2 @@
_base_ = './dnl_r50-d8_4xb2-40k_cityscapes-769x769.py'
model = dict(pretrained='open-mmlab://resnet101_v1c', backbone=dict(depth=101))

2
Seg_All_In_One_MMSeg/configs/dnlnet/dnl_r101-d8_4xb2-80k_cityscapes-512x1024.py Normal file
View File

@@ -0,0 +1,2 @@
_base_ = './dnl_r50-d8_4xb2-80k_cityscapes-512x1024.py'
model = dict(pretrained='open-mmlab://resnet101_v1c', backbone=dict(depth=101))

2
Seg_All_In_One_MMSeg/configs/dnlnet/dnl_r101-d8_4xb2-80k_cityscapes-769x769.py Normal file
View File

@@ -0,0 +1,2 @@
_base_ = './dnl_r50-d8_4xb2-80k_cityscapes-769x769.py'
model = dict(pretrained='open-mmlab://resnet101_v1c', backbone=dict(depth=101))

2
Seg_All_In_One_MMSeg/configs/dnlnet/dnl_r101-d8_4xb4-160k_ade20k-512x512.py Normal file
View File

@@ -0,0 +1,2 @@
_base_ = './dnl_r50-d8_4xb4-160k_ade20k-512x512.py'
model = dict(pretrained='open-mmlab://resnet101_v1c', backbone=dict(depth=101))

2
Seg_All_In_One_MMSeg/configs/dnlnet/dnl_r101-d8_4xb4-80k_ade20k-512x512.py Normal file
View File

@@ -0,0 +1,2 @@
_base_ = './dnl_r50-d8_4xb4-80k_ade20k-512x512.py'
model = dict(pretrained='open-mmlab://resnet101_v1c', backbone=dict(depth=101))

7
Seg_All_In_One_MMSeg/configs/dnlnet/dnl_r50-d8_4xb2-40k_cityscapes-512x1024.py Normal file
View File

@@ -0,0 +1,7 @@
_base_ = [
'../_base_/models/dnl_r50-d8.py', '../_base_/datasets/cityscapes.py',
'../_base_/default_runtime.py', '../_base_/schedules/schedule_40k.py'
]
crop_size = (512, 1024)
data_preprocessor = dict(size=crop_size)
model = dict(data_preprocessor=data_preprocessor)

12
Seg_All_In_One_MMSeg/configs/dnlnet/dnl_r50-d8_4xb2-40k_cityscapes-769x769.py Normal file
View File

@@ -0,0 +1,12 @@
_base_ = [
'../_base_/models/dnl_r50-d8.py',
'../_base_/datasets/cityscapes_769x769.py', '../_base_/default_runtime.py',
'../_base_/schedules/schedule_40k.py'
]
crop_size = (769, 769)
data_preprocessor = dict(size=crop_size)
model = dict(
data_preprocessor=data_preprocessor,
decode_head=dict(align_corners=True),
auxiliary_head=dict(align_corners=True),
test_cfg=dict(mode='slide', crop_size=(769, 769), stride=(513, 513)))

7
Seg_All_In_One_MMSeg/configs/dnlnet/dnl_r50-d8_4xb2-80k_cityscapes-512x1024.py Normal file
View File

@@ -0,0 +1,7 @@
_base_ = [
'../_base_/models/dnl_r50-d8.py', '../_base_/datasets/cityscapes.py',
'../_base_/default_runtime.py', '../_base_/schedules/schedule_80k.py'
]
crop_size = (512, 1024)
data_preprocessor = dict(size=crop_size)
model = dict(data_preprocessor=data_preprocessor)

16
Seg_All_In_One_MMSeg/configs/dnlnet/dnl_r50-d8_4xb2-80k_cityscapes-769x769.py Normal file
View File

@@ -0,0 +1,16 @@
_base_ = [
'../_base_/models/dnl_r50-d8.py',
'../_base_/datasets/cityscapes_769x769.py', '../_base_/default_runtime.py',
'../_base_/schedules/schedule_80k.py'
]
crop_size = (769, 769)
data_preprocessor = dict(size=crop_size)
model = dict(
data_preprocessor=data_preprocessor,
decode_head=dict(align_corners=True),
auxiliary_head=dict(align_corners=True),
test_cfg=dict(mode='slide', crop_size=(769, 769), stride=(513, 513)))
optim_wrapper = dict(
paramwise_cfg=dict(
custom_keys=dict(theta=dict(wd_mult=0.), phi=dict(wd_mult=0.))))

10
Seg_All_In_One_MMSeg/configs/dnlnet/dnl_r50-d8_4xb4-160k_ade20k-512x512.py Normal file
View File

@@ -0,0 +1,10 @@
_base_ = [
'../_base_/models/dnl_r50-d8.py', '../_base_/datasets/ade20k.py',
'../_base_/default_runtime.py', '../_base_/schedules/schedule_160k.py'
]
crop_size = (512, 512)
data_preprocessor = dict(size=crop_size)
model = dict(
data_preprocessor=data_preprocessor,
decode_head=dict(num_classes=150),
auxiliary_head=dict(num_classes=150))

10
Seg_All_In_One_MMSeg/configs/dnlnet/dnl_r50-d8_4xb4-80k_ade20k-512x512.py Normal file
View File

@@ -0,0 +1,10 @@
_base_ = [
'../_base_/models/dnl_r50-d8.py', '../_base_/datasets/ade20k.py',
'../_base_/default_runtime.py', '../_base_/schedules/schedule_80k.py'
]
crop_size = (512, 512)
data_preprocessor = dict(size=crop_size)
model = dict(
data_preprocessor=data_preprocessor,
decode_head=dict(num_classes=150),
auxiliary_head=dict(num_classes=150))

292
Seg_All_In_One_MMSeg/configs/dnlnet/metafile.yaml Normal file
View File

@@ -0,0 +1,292 @@
Collections:
- Name: DNLNet
License: Apache License 2.0
Metadata:
Training Data:
- Cityscapes
- ADE20K
Paper:
Title: Disentangled Non-Local Neural Networks
URL: https://arxiv.org/abs/2006.06668
README: configs/dnlnet/README.md
Frameworks:
- PyTorch
Models:
- Name: dnl_r50-d8_4xb2-40k_cityscapes-512x1024
In Collection: DNLNet
Results:
Task: Semantic Segmentation
Dataset: Cityscapes
Metrics:
mIoU: 78.61
Config: configs/dnlnet/dnl_r50-d8_4xb2-40k_cityscapes-512x1024.py
Metadata:
Training Data: Cityscapes
Batch Size: 8
Architecture:
- R-50-D8
- DNLNet
Training Resources: 4x V100 GPUS
Memory (GB): 7.3
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/dnlnet/dnl_r50-d8_512x1024_40k_cityscapes/dnl_r50-d8_512x1024_40k_cityscapes_20200904_233629-53d4ea93.pth
Training log: https://download.openmmlab.com/mmsegmentation/v0.5/dnlnet/dnl_r50-d8_512x1024_40k_cityscapes/dnl_r50-d8_512x1024_40k_cityscapes-20200904_233629.log.json
Paper:
Title: Disentangled Non-Local Neural Networks
URL: https://arxiv.org/abs/2006.06668
Code: https://github.com/open-mmlab/mmsegmentation/blob/v0.17.0/mmseg/models/decode_heads/dnl_head.py#L88
Framework: PyTorch
- Name: dnl_r101-d8_4xb2-40k_cityscapes-512x1024
In Collection: DNLNet
Results:
Task: Semantic Segmentation
Dataset: Cityscapes
Metrics:
mIoU: 78.31
Config: configs/dnlnet/dnl_r101-d8_4xb2-40k_cityscapes-512x1024.py
Metadata:
Training Data: Cityscapes
Batch Size: 8
Architecture:
- R-101-D8
- DNLNet
Training Resources: 4x V100 GPUS
Memory (GB): 10.9
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/dnlnet/dnl_r101-d8_512x1024_40k_cityscapes/dnl_r101-d8_512x1024_40k_cityscapes_20200904_233629-9928ffef.pth
Training log: https://download.openmmlab.com/mmsegmentation/v0.5/dnlnet/dnl_r101-d8_512x1024_40k_cityscapes/dnl_r101-d8_512x1024_40k_cityscapes-20200904_233629.log.json
Paper:
Title: Disentangled Non-Local Neural Networks
URL: https://arxiv.org/abs/2006.06668
Code: https://github.com/open-mmlab/mmsegmentation/blob/v0.17.0/mmseg/models/decode_heads/dnl_head.py#L88
Framework: PyTorch
- Name: dnl_r50-d8_4xb2-40k_cityscapes-769x769
In Collection: DNLNet
Results:
Task: Semantic Segmentation
Dataset: Cityscapes
Metrics:
mIoU: 78.44
mIoU(ms+flip): 80.27
Config: configs/dnlnet/dnl_r50-d8_4xb2-40k_cityscapes-769x769.py
Metadata:
Training Data: Cityscapes
Batch Size: 8
Architecture:
- R-50-D8
- DNLNet
Training Resources: 4x V100 GPUS
Memory (GB): 9.2
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/dnlnet/dnl_r50-d8_769x769_40k_cityscapes/dnl_r50-d8_769x769_40k_cityscapes_20200820_232206-0f283785.pth
Training log: https://download.openmmlab.com/mmsegmentation/v0.5/dnlnet/dnl_r50-d8_769x769_40k_cityscapes/dnl_r50-d8_769x769_40k_cityscapes-20200820_232206.log.json
Paper:
Title: Disentangled Non-Local Neural Networks
URL: https://arxiv.org/abs/2006.06668
Code: https://github.com/open-mmlab/mmsegmentation/blob/v0.17.0/mmseg/models/decode_heads/dnl_head.py#L88
Framework: PyTorch
- Name: dnl_r101-d8_4xb2-40k_cityscapes-769x769
In Collection: DNLNet
Results:
Task: Semantic Segmentation
Dataset: Cityscapes
Metrics:
mIoU: 76.39
mIoU(ms+flip): 77.77
Config: configs/dnlnet/dnl_r101-d8_4xb2-40k_cityscapes-769x769.py
Metadata:
Training Data: Cityscapes
Batch Size: 8
Architecture:
- R-101-D8
- DNLNet
Training Resources: 4x V100 GPUS
Memory (GB): 12.6
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/dnlnet/dnl_r101-d8_769x769_40k_cityscapes/dnl_r101-d8_769x769_40k_cityscapes_20200820_171256-76c596df.pth
Training log: https://download.openmmlab.com/mmsegmentation/v0.5/dnlnet/dnl_r101-d8_769x769_40k_cityscapes/dnl_r101-d8_769x769_40k_cityscapes-20200820_171256.log.json
Paper:
Title: Disentangled Non-Local Neural Networks
URL: https://arxiv.org/abs/2006.06668
Code: https://github.com/open-mmlab/mmsegmentation/blob/v0.17.0/mmseg/models/decode_heads/dnl_head.py#L88
Framework: PyTorch
- Name: dnl_r50-d8_4xb2-80k_cityscapes-512x1024
In Collection: DNLNet
Results:
Task: Semantic Segmentation
Dataset: Cityscapes
Metrics:
mIoU: 79.33
Config: configs/dnlnet/dnl_r50-d8_4xb2-80k_cityscapes-512x1024.py
Metadata:
Training Data: Cityscapes
Batch Size: 8
Architecture:
- R-50-D8
- DNLNet
Training Resources: 4x V100 GPUS
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/dnlnet/dnl_r50-d8_512x1024_80k_cityscapes/dnl_r50-d8_512x1024_80k_cityscapes_20200904_233629-58b2f778.pth
Training log: https://download.openmmlab.com/mmsegmentation/v0.5/dnlnet/dnl_r50-d8_512x1024_80k_cityscapes/dnl_r50-d8_512x1024_80k_cityscapes-20200904_233629.log.json
Paper:
Title: Disentangled Non-Local Neural Networks
URL: https://arxiv.org/abs/2006.06668
Code: https://github.com/open-mmlab/mmsegmentation/blob/v0.17.0/mmseg/models/decode_heads/dnl_head.py#L88
Framework: PyTorch
- Name: dnl_r101-d8_4xb2-80k_cityscapes-512x1024
In Collection: DNLNet
Results:
Task: Semantic Segmentation
Dataset: Cityscapes
Metrics:
mIoU: 80.41
Config: configs/dnlnet/dnl_r101-d8_4xb2-80k_cityscapes-512x1024.py
Metadata:
Training Data: Cityscapes
Batch Size: 8
Architecture:
- R-101-D8
- DNLNet
Training Resources: 4x V100 GPUS
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/dnlnet/dnl_r101-d8_512x1024_80k_cityscapes/dnl_r101-d8_512x1024_80k_cityscapes_20200904_233629-758e2dd4.pth
Training log: https://download.openmmlab.com/mmsegmentation/v0.5/dnlnet/dnl_r101-d8_512x1024_80k_cityscapes/dnl_r101-d8_512x1024_80k_cityscapes-20200904_233629.log.json
Paper:
Title: Disentangled Non-Local Neural Networks
URL: https://arxiv.org/abs/2006.06668
Code: https://github.com/open-mmlab/mmsegmentation/blob/v0.17.0/mmseg/models/decode_heads/dnl_head.py#L88
Framework: PyTorch
- Name: dnl_r50-d8_4xb2-80k_cityscapes-769x769
In Collection: DNLNet
Results:
Task: Semantic Segmentation
Dataset: Cityscapes
Metrics:
mIoU: 79.36
mIoU(ms+flip): 80.7
Config: configs/dnlnet/dnl_r50-d8_4xb2-80k_cityscapes-769x769.py
Metadata:
Training Data: Cityscapes
Batch Size: 8
Architecture:
- R-50-D8
- DNLNet
Training Resources: 4x V100 GPUS
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/dnlnet/dnl_r50-d8_769x769_80k_cityscapes/dnl_r50-d8_769x769_80k_cityscapes_20200820_011925-366bc4c7.pth
Training log: https://download.openmmlab.com/mmsegmentation/v0.5/dnlnet/dnl_r50-d8_769x769_80k_cityscapes/dnl_r50-d8_769x769_80k_cityscapes-20200820_011925.log.json
Paper:
Title: Disentangled Non-Local Neural Networks
URL: https://arxiv.org/abs/2006.06668
Code: https://github.com/open-mmlab/mmsegmentation/blob/v0.17.0/mmseg/models/decode_heads/dnl_head.py#L88
Framework: PyTorch
- Name: dnl_r101-d8_4xb2-80k_cityscapes-769x769
In Collection: DNLNet
Results:
Task: Semantic Segmentation
Dataset: Cityscapes
Metrics:
mIoU: 79.41
mIoU(ms+flip): 80.68
Config: configs/dnlnet/dnl_r101-d8_4xb2-80k_cityscapes-769x769.py
Metadata:
Training Data: Cityscapes
Batch Size: 8
Architecture:
- R-101-D8
- DNLNet
Training Resources: 4x V100 GPUS
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/dnlnet/dnl_r101-d8_769x769_80k_cityscapes/dnl_r101-d8_769x769_80k_cityscapes_20200821_051111-95ff84ab.pth
Training log: https://download.openmmlab.com/mmsegmentation/v0.5/dnlnet/dnl_r101-d8_769x769_80k_cityscapes/dnl_r101-d8_769x769_80k_cityscapes-20200821_051111.log.json
Paper:
Title: Disentangled Non-Local Neural Networks
URL: https://arxiv.org/abs/2006.06668
Code: https://github.com/open-mmlab/mmsegmentation/blob/v0.17.0/mmseg/models/decode_heads/dnl_head.py#L88
Framework: PyTorch
- Name: dnl_r50-d8_4xb4-80k_ade20k-512x512
In Collection: DNLNet
Results:
Task: Semantic Segmentation
Dataset: ADE20K
Metrics:
mIoU: 41.76
mIoU(ms+flip): 42.99
Config: configs/dnlnet/dnl_r50-d8_4xb4-80k_ade20k-512x512.py
Metadata:
Training Data: ADE20K
Batch Size: 16
Architecture:
- R-50-D8
- DNLNet
Training Resources: 4x V100 GPUS
Memory (GB): 8.8
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/dnlnet/dnl_r50-d8_512x512_80k_ade20k/dnl_r50-d8_512x512_80k_ade20k_20200826_183354-1cf6e0c1.pth
Training log: https://download.openmmlab.com/mmsegmentation/v0.5/dnlnet/dnl_r50-d8_512x512_80k_ade20k/dnl_r50-d8_512x512_80k_ade20k-20200826_183354.log.json
Paper:
Title: Disentangled Non-Local Neural Networks
URL: https://arxiv.org/abs/2006.06668
Code: https://github.com/open-mmlab/mmsegmentation/blob/v0.17.0/mmseg/models/decode_heads/dnl_head.py#L88
Framework: PyTorch
- Name: dnl_r101-d8_4xb4-80k_ade20k-512x512
In Collection: DNLNet
Results:
Task: Semantic Segmentation
Dataset: ADE20K
Metrics:
mIoU: 43.76
mIoU(ms+flip): 44.91
Config: configs/dnlnet/dnl_r101-d8_4xb4-80k_ade20k-512x512.py
Metadata:
Training Data: ADE20K
Batch Size: 16
Architecture:
- R-101-D8
- DNLNet
Training Resources: 4x V100 GPUS
Memory (GB): 12.8
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/dnlnet/dnl_r101-d8_512x512_80k_ade20k/dnl_r101-d8_512x512_80k_ade20k_20200826_183354-d820d6ea.pth
Training log: https://download.openmmlab.com/mmsegmentation/v0.5/dnlnet/dnl_r101-d8_512x512_80k_ade20k/dnl_r101-d8_512x512_80k_ade20k-20200826_183354.log.json
Paper:
Title: Disentangled Non-Local Neural Networks
URL: https://arxiv.org/abs/2006.06668
Code: https://github.com/open-mmlab/mmsegmentation/blob/v0.17.0/mmseg/models/decode_heads/dnl_head.py#L88
Framework: PyTorch
- Name: dnl_r50-d8_4xb4-160k_ade20k-512x512
In Collection: DNLNet
Results:
Task: Semantic Segmentation
Dataset: ADE20K
Metrics:
mIoU: 41.87
mIoU(ms+flip): 43.01
Config: configs/dnlnet/dnl_r50-d8_4xb4-160k_ade20k-512x512.py
Metadata:
Training Data: ADE20K
Batch Size: 16
Architecture:
- R-50-D8
- DNLNet
Training Resources: 4x V100 GPUS
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/dnlnet/dnl_r50-d8_512x512_160k_ade20k/dnl_r50-d8_512x512_160k_ade20k_20200826_183350-37837798.pth
Training log: https://download.openmmlab.com/mmsegmentation/v0.5/dnlnet/dnl_r50-d8_512x512_160k_ade20k/dnl_r50-d8_512x512_160k_ade20k-20200826_183350.log.json
Paper:
Title: Disentangled Non-Local Neural Networks
URL: https://arxiv.org/abs/2006.06668
Code: https://github.com/open-mmlab/mmsegmentation/blob/v0.17.0/mmseg/models/decode_heads/dnl_head.py#L88
Framework: PyTorch
- Name: dnl_r101-d8_4xb4-160k_ade20k-512x512
In Collection: DNLNet
Results:
Task: Semantic Segmentation
Dataset: ADE20K
Metrics:
mIoU: 44.25
mIoU(ms+flip): 45.78
Config: configs/dnlnet/dnl_r101-d8_4xb4-160k_ade20k-512x512.py
Metadata:
Training Data: ADE20K
Batch Size: 16
Architecture:
- R-101-D8
- DNLNet
Training Resources: 4x V100 GPUS
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/dnlnet/dnl_r101-d8_512x512_160k_ade20k/dnl_r101-d8_512x512_160k_ade20k_20200826_183350-ed522c61.pth
Training log: https://download.openmmlab.com/mmsegmentation/v0.5/dnlnet/dnl_r101-d8_512x512_160k_ade20k/dnl_r101-d8_512x512_160k_ade20k-20200826_183350.log.json
Paper:
Title: Disentangled Non-Local Neural Networks
URL: https://arxiv.org/abs/2006.06668
Code: https://github.com/open-mmlab/mmsegmentation/blob/v0.17.0/mmseg/models/decode_heads/dnl_head.py#L88
Framework: PyTorch

105
Seg_All_In_One_MMSeg/configs/dnlnet/my_dnlnet_r50_Pre_g1-40k_check_4000_my_dataset_model-512x512-testslide.py Normal file
View File

@@ -0,0 +1,105 @@
_base_ = [
'../_base_/models/dnl_r50-d8.py',
'../_base_/datasets/my_dataset_model.py',
'../_base_/default_runtime.py',
'../_base_/schedules/schedule_40k_check_4000.py',
]
norm_cfg = dict(
type='BN',
)
crop_size = (512, 512)
data_preprocessor = dict(
size=(512, 512),
mean=[
94.94709810464303,
61.72942233949928,
75.93763705236906,
],
std=[
44.005506081132594,
42.69595666984776,
44.99354156225523,
],
bgr_to_rgb=False,
)
model = dict(
pretrained='./My_Local_Model/open_mmlab/resnet50_v1c.pth',
backbone=dict(
depth=50,
),
data_preprocessor=dict(
size=(512, 512),
mean=[
94.94709810464303,
61.72942233949928,
75.93763705236906,
],
std=[
44.005506081132594,
42.69595666984776,
44.99354156225523,
],
bgr_to_rgb=False,
),
decode_head=dict(
num_classes=36,
loss_decode=dict(
type='DiceLoss',
use_sigmoid=False,
loss_weight=1.0,
),
align_corners=True,
),
auxiliary_head=dict(
num_classes=36,
loss_decode=dict(
type='DiceLoss',
use_sigmoid=False,
loss_weight=0.4,
),
align_corners=True,
),
)
test_cfg = dict(
mode='slide',
crop_size=(512, 512),
stride=(341, 341),
)
optim_wrapper = dict(
type='OptimWrapper',
_delete_=True,
optimizer=dict(
type='AdamW',
lr=0.0001,
weight_decay=0.0005,
),
clip_grad=dict(
max_norm=1,
norm_type=2,
),
)
param_scheduler = [
dict(
type='LinearLR',
start_factor=1e-06,
by_epoch=False,
begin=0,
end=1500,
),
dict(
type='PolyLR',
power=0.9,
begin=1500,
end=40000,
eta_min=1e-05,
by_epoch=False,
),
]