charlie880624/STDC
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

[Enhancement] Change readme style and Update metafiles. (#895)

Browse Source 
* [Enhancement] Change readme style and prepare for metafiles update.

* Update apcnet github repo url.

* add code snippet.

* split code snippet & official repo.

* update md2yml hook.

* Update metafiles.

* Add converted from attribute.

* process conflict.

* Put defualt variable value.

* update bisenet v2 metafile.

* checkout to ubuntu environment.

* pop empty attribute & make task attribute list.

* update readme style

* update readme style

* update metafiles

Co-authored-by: Junjun2016 <hejunjun@sjtu.edu.cn>
This commit is contained in:
sennnnn 2021-09-28 16:25:37 +08:00 committed by GitHub
parent 36eb2d8beb
commit 2aa632ebe7
No known key found for this signature in database
GPG Key ID: 4AEE18F83AFDEB23
72 changed files with 3262 additions and 2573 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

112
.dev/md2yml.py
View File

@ -9,25 +9,28 @@
import glob
import os
import os.path as osp
import re
import sys
import mmcv
from lxml import etree
MMSEG_ROOT = osp.dirname(osp.dirname((osp.dirname(__file__))))
def dump_yaml_and_check_difference(obj, filename):
def dump_yaml_and_check_difference(obj, filename, sort_keys=False):
"""Dump object to a yaml file, and check if the file content is different
from the original.
Args:
obj (any): The python object to be dumped.
filename (str): YAML filename to dump the object to.
sort_keys (str); Sort key by dictionary order.
Returns:
Bool: If the target YAML file is different from the original.
"""
str_dump = mmcv.dump(obj, None, file_format='yaml', sort_keys=True)
str_dump = mmcv.dump(obj, None, file_format='yaml', sort_keys=sort_keys)
if osp.isfile(filename):
file_exists = True
with open(filename, 'r', encoding='utf-8') as f:
@ -54,12 +57,29 @@ def parse_md(md_file):
Returns:
Bool: If the target YAML file is different from the original.
"""
collection_name = osp.dirname(md_file).split('/')[-1]
collection_name = osp.split(osp.dirname(md_file))[1]
configs = os.listdir(osp.dirname(md_file))
collection = dict(Name=collection_name, Metadata={'Training Data': []})
collection = dict(
Name=collection_name,
Metadata={'Training Data': []},
Paper={
'URL': '',
'Title': ''
},
README=md_file,
Code={
'URL': '',
'Version': ''
})
collection.update({'Converted From': {'Weights': '', 'Code': ''}})
models = []
datasets = []
paper_url = None
paper_title = None
code_url = None
code_version = None
repo_url = None
with open(md_file, 'r') as md:
lines = md.readlines()
@ -70,7 +90,36 @@ def parse_md(md_file):
if len(line) == 0:
i += 1
continue
if line[:3] == '###':
if line[:2] == '# ':
paper_title = line.replace('# ', '')
i += 1
elif line[:3] == '<a ':
content = etree.HTML(line)
node = content.xpath('//a')[0]
if node.text == 'Code Snippet':
code_url = node.get('href', None)
assert code_url is not None, (
f'{collection_name} hasn\'t code snippet url.')
# version extraction
filter_str = r'blob/(.*)/mm'
pattern = re.compile(filter_str)
code_version = pattern.findall(code_url)
assert len(code_version) == 1, (
f'false regular expression ({filter_str}) use.')
code_version = code_version[0]
elif node.text == 'Official Repo':
repo_url = node.get('href', None)
assert repo_url is not None, (
f'{collection_name} hasn\'t official repo url.')
i += 1
elif line[:9] == '<summary ':
content = etree.HTML(line)
nodes = content.xpath('//a')
assert len(nodes) == 1, (
'summary tag should only have single a tag.')
paper_url = nodes[0].get('href', None)
i += 1
elif line[:4] == '### ':
datasets.append(line[4:])
current_dataset = line[4:]
i += 2
@ -113,22 +162,28 @@ def parse_md(md_file):
crop_size = els[crop_size_id].split('x')
assert len(crop_size) == 2
model = {
'Name': model_name,
'In Collection': collection_name,
'Name':
model_name,
'In Collection':
collection_name,
'Metadata': {
'backbone': els[backbone_id],
'crop size': f'({crop_size[0]},{crop_size[1]})',
'lr schd': int(els[lr_schd_id]),
},
'Results': {
'Task': 'Semantic Segmentation',
'Dataset': current_dataset,
'Metrics': {
'mIoU': float(els[ss_id]),
'Results': [
{
'Task': 'Semantic Segmentation',
'Dataset': current_dataset,
'Metrics': {
'mIoU': float(els[ss_id]),
},
},
},
'Config': config,
'Weights': weight,
],
'Config':
config,
'Weights':
weight,
}
if fps != -1:
try:
@ -152,15 +207,38 @@ def parse_md(md_file):
}]
if mem != -1:
model['Metadata']['memory (GB)'] = float(mem)
# Only have semantic segmentation now
if ms_id and els[ms_id] != '-' and els[ms_id] != '':
model['Results']['Metrics']['mIoU(ms+flip)'] = float(
els[ms_id])
model['Results'][0]['Metrics'][
'mIoU(ms+flip)'] = float(els[ms_id])
models.append(model)
j += 1
i = j
else:
i += 1
flag = (code_url is not None) and (paper_url is not None) and (repo_url
is not None)
assert flag, f'{collection_name} readme error'
collection['Metadata']['Training Data'] = datasets
collection['Code']['URL'] = code_url
collection['Code']['Version'] = code_version
collection['Paper']['URL'] = paper_url
collection['Paper']['Title'] = paper_title
collection['Converted From']['Code'] = repo_url
# ['Converted From']['Weights] miss
# remove empty attribute
check_key_list = ['Code', 'Paper', 'Converted From']
for check_key in check_key_list:
key_list = list(collection[check_key].keys())
for key in key_list:
if check_key not in collection:
break
if collection[check_key][key] == '':
if len(collection[check_key].keys()) == 1:
collection.pop(check_key)
else:
collection[check_key].pop(key)
result = {'Collections': [collection], 'Models': models}
yml_file = f'{md_file[:-9]}{collection_name}.yml'
return dump_yaml_and_check_difference(result, yml_file)

2
.pre-commit-config.yaml
View File

@ -45,7 +45,7 @@ repos:
name: update-model-index
description: Collect model information and update model-index.yml
entry: .dev/md2yml.py
additional_dependencies: [mmcv]
additional_dependencies: [mmcv, lxml]
language: python
files: ^configs/.*\.md$
require_serial: true

3
README.md
View File

@ -66,6 +66,7 @@ Supported backbones:
- [x] [MobileNetV3 (ICCV'2019)](configs/mobilenet_v3)
- [x] [Vision Transformer (ICLR'2021)](configs/vit)
- [x] [Swin Transformer (ArXiv'2021)](configs/swin)
- [x] [BiSeNetV2 (IJCV'2021)](configs/bisenetv2)
Supported methods:
@ -94,7 +95,7 @@ Supported methods:
- [x] [PointRend (CVPR'2020)](configs/point_rend)
- [x] [CGNet (TIP'2020)](configs/cgnet)
- [x] [SETR (CVPR'2021)](configs/setr)
- [x] [BiSeNetV2 (IJCV'2021)](configs/bisenetv2)
- [x] [DPT (ArXiv'2021)](configs/dpt)
- [x] [SegFormer (ArXiv'2021)](configs/segformer)
Supported datasets:

3
README_zh-CN.md
View File

@ -65,6 +65,7 @@ MMSegmentation 是一个基于 PyTorch 的语义分割开源工具箱。它是 O
- [x] [MobileNetV3 (ICCV'2019)](configs/mobilenet_v3)
- [x] [Vision Transformer (ICLR'2021)](configs/vit)
- [x] [Swin Transformer (ArXiv'2021)](configs/swin)
- [x] [BiSeNetV2 (IJCV'2021)](configs/bisenetv2)
已支持的算法：
@ -93,7 +94,7 @@ MMSegmentation 是一个基于 PyTorch 的语义分割开源工具箱。它是 O
- [x] [PointRend (CVPR'2020)](configs/point_rend)
- [x] [CGNet (TIP'2020)](configs/cgnet)
- [x] [SETR (CVPR'2021)](configs/setr)
- [x] [BiSeNetV2 (IJCV'2021)](configs/bisenetv2)
- [x] [DPT (ArXiv'2021)](configs/dpt)
- [x] [SegFormer (ArXiv'2021)](configs/segformer)
已支持的数据集：

9
configs/ann/README.md
View File

@ -4,6 +4,13 @@
<!-- [ALGORITHM] -->
<a href="https://github.com/MendelXu/ANN">Official Repo</a>
<a href="https://github.com/open-mmlab/mmsegmentation/blob/v0.17.0/mmseg/models/decode_heads/ann_head.py#L185">Code Snippet</a>
<details>
<summary align="right"><a href="https://arxiv.org/abs/1908.07678">ANN (ICCV'2019)</a></summary>
```latex
@inproceedings{zhu2019asymmetric,
title={Asymmetric non-local neural networks for semantic segmentation},
@ -14,6 +21,8 @@
}
```
</details>
## Results and models
### Cityscapes

173
configs/ann/ann.yml
View File

@ -1,296 +1,305 @@
Collections:
- Metadata:
- Name: ann
Metadata:
Training Data:
- Cityscapes
- ADE20K
- Pascal VOC 2012 + Aug
Name: ann
Paper:
URL: https://arxiv.org/abs/1908.07678
Title: Asymmetric Non-local Neural Networks for Semantic Segmentation
README: configs/ann/README.md
Code:
URL: https://github.com/open-mmlab/mmsegmentation/blob/v0.17.0/mmseg/models/decode_heads/ann_head.py#L185
Version: v0.17.0
Converted From:
Code: https://github.com/MendelXu/ANN
Models:
- Config: configs/ann/ann_r50-d8_512x1024_40k_cityscapes.py
- Name: ann_r50-d8_512x1024_40k_cityscapes
In Collection: ann
Metadata:
backbone: R-50-D8
crop size: (512,1024)
lr schd: 40000
inference time (ms/im):
- backend: PyTorch
batch size: 1
- value: 269.54
hardware: V100
backend: PyTorch
batch size: 1
mode: FP32
resolution: (512,1024)
value: 269.54
lr schd: 40000
memory (GB): 6.0
Name: ann_r50-d8_512x1024_40k_cityscapes
Results:
- Task: Semantic Segmentation
Dataset: Cityscapes
Metrics:
mIoU: 77.4
mIoU(ms+flip): 78.57
Task: Semantic Segmentation
Config: configs/ann/ann_r50-d8_512x1024_40k_cityscapes.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/ann/ann_r50-d8_512x1024_40k_cityscapes/ann_r50-d8_512x1024_40k_cityscapes_20200605_095211-049fc292.pth
- Config: configs/ann/ann_r101-d8_512x1024_40k_cityscapes.py
- Name: ann_r101-d8_512x1024_40k_cityscapes
In Collection: ann
Metadata:
backbone: R-101-D8
crop size: (512,1024)
lr schd: 40000
inference time (ms/im):
- backend: PyTorch
batch size: 1
- value: 392.16
hardware: V100
backend: PyTorch
batch size: 1
mode: FP32
resolution: (512,1024)
value: 392.16
lr schd: 40000
memory (GB): 9.5
Name: ann_r101-d8_512x1024_40k_cityscapes
Results:
- Task: Semantic Segmentation
Dataset: Cityscapes
Metrics:
mIoU: 76.55
mIoU(ms+flip): 78.85
Task: Semantic Segmentation
Config: configs/ann/ann_r101-d8_512x1024_40k_cityscapes.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/ann/ann_r101-d8_512x1024_40k_cityscapes/ann_r101-d8_512x1024_40k_cityscapes_20200605_095243-adf6eece.pth
- Config: configs/ann/ann_r50-d8_769x769_40k_cityscapes.py
- Name: ann_r50-d8_769x769_40k_cityscapes
In Collection: ann
Metadata:
backbone: R-50-D8
crop size: (769,769)
lr schd: 40000
inference time (ms/im):
- backend: PyTorch
batch size: 1
- value: 588.24
hardware: V100
backend: PyTorch
batch size: 1
mode: FP32
resolution: (769,769)
value: 588.24
lr schd: 40000
memory (GB): 6.8
Name: ann_r50-d8_769x769_40k_cityscapes
Results:
- Task: Semantic Segmentation
Dataset: Cityscapes
Metrics:
mIoU: 78.89
mIoU(ms+flip): 80.46
Task: Semantic Segmentation
Config: configs/ann/ann_r50-d8_769x769_40k_cityscapes.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/ann/ann_r50-d8_769x769_40k_cityscapes/ann_r50-d8_769x769_40k_cityscapes_20200530_025712-2b46b04d.pth
- Config: configs/ann/ann_r101-d8_769x769_40k_cityscapes.py
- Name: ann_r101-d8_769x769_40k_cityscapes
In Collection: ann
Metadata:
backbone: R-101-D8
crop size: (769,769)
lr schd: 40000
inference time (ms/im):
- backend: PyTorch
batch size: 1
- value: 869.57
hardware: V100
backend: PyTorch
batch size: 1
mode: FP32
resolution: (769,769)
value: 869.57
lr schd: 40000
memory (GB): 10.7
Name: ann_r101-d8_769x769_40k_cityscapes
Results:
- Task: Semantic Segmentation
Dataset: Cityscapes
Metrics:
mIoU: 79.32
mIoU(ms+flip): 80.94
Task: Semantic Segmentation
Config: configs/ann/ann_r101-d8_769x769_40k_cityscapes.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/ann/ann_r101-d8_769x769_40k_cityscapes/ann_r101-d8_769x769_40k_cityscapes_20200530_025720-059bff28.pth
- Config: configs/ann/ann_r50-d8_512x1024_80k_cityscapes.py
- Name: ann_r50-d8_512x1024_80k_cityscapes
In Collection: ann
Metadata:
backbone: R-50-D8
crop size: (512,1024)
lr schd: 80000
Name: ann_r50-d8_512x1024_80k_cityscapes
Results:
- Task: Semantic Segmentation
Dataset: Cityscapes
Metrics:
mIoU: 77.34
mIoU(ms+flip): 78.65
Task: Semantic Segmentation
Config: configs/ann/ann_r50-d8_512x1024_80k_cityscapes.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/ann/ann_r50-d8_512x1024_80k_cityscapes/ann_r50-d8_512x1024_80k_cityscapes_20200607_101911-5a9ad545.pth
- Config: configs/ann/ann_r101-d8_512x1024_80k_cityscapes.py
- Name: ann_r101-d8_512x1024_80k_cityscapes
In Collection: ann
Metadata:
backbone: R-101-D8
crop size: (512,1024)
lr schd: 80000
Name: ann_r101-d8_512x1024_80k_cityscapes
Results:
- Task: Semantic Segmentation
Dataset: Cityscapes
Metrics:
mIoU: 77.14
mIoU(ms+flip): 78.81
Task: Semantic Segmentation
Config: configs/ann/ann_r101-d8_512x1024_80k_cityscapes.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/ann/ann_r101-d8_512x1024_80k_cityscapes/ann_r101-d8_512x1024_80k_cityscapes_20200607_013728-aceccc6e.pth
- Config: configs/ann/ann_r50-d8_769x769_80k_cityscapes.py
- Name: ann_r50-d8_769x769_80k_cityscapes
In Collection: ann
Metadata:
backbone: R-50-D8
crop size: (769,769)
lr schd: 80000
Name: ann_r50-d8_769x769_80k_cityscapes
Results:
- Task: Semantic Segmentation
Dataset: Cityscapes
Metrics:
mIoU: 78.88
mIoU(ms+flip): 80.57
Task: Semantic Segmentation
Config: configs/ann/ann_r50-d8_769x769_80k_cityscapes.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/ann/ann_r50-d8_769x769_80k_cityscapes/ann_r50-d8_769x769_80k_cityscapes_20200607_044426-cc7ff323.pth
- Config: configs/ann/ann_r101-d8_769x769_80k_cityscapes.py
- Name: ann_r101-d8_769x769_80k_cityscapes
In Collection: ann
Metadata:
backbone: R-101-D8
crop size: (769,769)
lr schd: 80000
Name: ann_r101-d8_769x769_80k_cityscapes
Results:
- Task: Semantic Segmentation
Dataset: Cityscapes
Metrics:
mIoU: 78.8
mIoU(ms+flip): 80.34
Task: Semantic Segmentation
Config: configs/ann/ann_r101-d8_769x769_80k_cityscapes.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/ann/ann_r101-d8_769x769_80k_cityscapes/ann_r101-d8_769x769_80k_cityscapes_20200607_013713-a9d4be8d.pth
- Config: configs/ann/ann_r50-d8_512x512_80k_ade20k.py
- Name: ann_r50-d8_512x512_80k_ade20k
In Collection: ann
Metadata:
backbone: R-50-D8
crop size: (512,512)
lr schd: 80000
inference time (ms/im):
- backend: PyTorch
batch size: 1
- value: 47.6
hardware: V100
backend: PyTorch
batch size: 1
mode: FP32
resolution: (512,512)
value: 47.6
lr schd: 80000
memory (GB): 9.1
Name: ann_r50-d8_512x512_80k_ade20k
Results:
- Task: Semantic Segmentation
Dataset: ADE20K
Metrics:
mIoU: 41.01
mIoU(ms+flip): 42.3
Task: Semantic Segmentation
Config: configs/ann/ann_r50-d8_512x512_80k_ade20k.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/ann/ann_r50-d8_512x512_80k_ade20k/ann_r50-d8_512x512_80k_ade20k_20200615_014818-26f75e11.pth
- Config: configs/ann/ann_r101-d8_512x512_80k_ade20k.py
- Name: ann_r101-d8_512x512_80k_ade20k
In Collection: ann
Metadata:
backbone: R-101-D8
crop size: (512,512)
lr schd: 80000
inference time (ms/im):
- backend: PyTorch
batch size: 1
- value: 70.82
hardware: V100
backend: PyTorch
batch size: 1
mode: FP32
resolution: (512,512)
value: 70.82
lr schd: 80000
memory (GB): 12.5
Name: ann_r101-d8_512x512_80k_ade20k
Results:
- Task: Semantic Segmentation
Dataset: ADE20K
Metrics:
mIoU: 42.94
mIoU(ms+flip): 44.18
Task: Semantic Segmentation
Config: configs/ann/ann_r101-d8_512x512_80k_ade20k.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/ann/ann_r101-d8_512x512_80k_ade20k/ann_r101-d8_512x512_80k_ade20k_20200615_014818-c0153543.pth
- Config: configs/ann/ann_r50-d8_512x512_160k_ade20k.py
- Name: ann_r50-d8_512x512_160k_ade20k
In Collection: ann
Metadata:
backbone: R-50-D8
crop size: (512,512)
lr schd: 160000
Name: ann_r50-d8_512x512_160k_ade20k
Results:
- Task: Semantic Segmentation
Dataset: ADE20K
Metrics:
mIoU: 41.74
mIoU(ms+flip): 42.62
Task: Semantic Segmentation
Config: configs/ann/ann_r50-d8_512x512_160k_ade20k.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/ann/ann_r50-d8_512x512_160k_ade20k/ann_r50-d8_512x512_160k_ade20k_20200615_231733-892247bc.pth
- Config: configs/ann/ann_r101-d8_512x512_160k_ade20k.py
- Name: ann_r101-d8_512x512_160k_ade20k
In Collection: ann
Metadata:
backbone: R-101-D8
crop size: (512,512)
lr schd: 160000
Name: ann_r101-d8_512x512_160k_ade20k
Results:
- Task: Semantic Segmentation
Dataset: ADE20K
Metrics:
mIoU: 42.94
mIoU(ms+flip): 44.06
Task: Semantic Segmentation
Config: configs/ann/ann_r101-d8_512x512_160k_ade20k.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/ann/ann_r101-d8_512x512_160k_ade20k/ann_r101-d8_512x512_160k_ade20k_20200615_231733-955eb1ec.pth
- Config: configs/ann/ann_r50-d8_512x512_20k_voc12aug.py
- Name: ann_r50-d8_512x512_20k_voc12aug
In Collection: ann
Metadata:
backbone: R-50-D8
crop size: (512,512)
lr schd: 20000
inference time (ms/im):
- backend: PyTorch
batch size: 1
- value: 47.8
hardware: V100
backend: PyTorch
batch size: 1
mode: FP32
resolution: (512,512)
value: 47.8
lr schd: 20000
memory (GB): 6.0
Name: ann_r50-d8_512x512_20k_voc12aug
Results:
- Task: Semantic Segmentation
Dataset: Pascal VOC 2012 + Aug
Metrics:
mIoU: 74.86
mIoU(ms+flip): 76.13
Task: Semantic Segmentation
Config: configs/ann/ann_r50-d8_512x512_20k_voc12aug.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/ann/ann_r50-d8_512x512_20k_voc12aug/ann_r50-d8_512x512_20k_voc12aug_20200617_222246-dfcb1c62.pth
- Config: configs/ann/ann_r101-d8_512x512_20k_voc12aug.py
- Name: ann_r101-d8_512x512_20k_voc12aug
In Collection: ann
Metadata:
backbone: R-101-D8
crop size: (512,512)
lr schd: 20000
inference time (ms/im):
- backend: PyTorch
batch size: 1
- value: 71.74
hardware: V100
backend: PyTorch
batch size: 1
mode: FP32
resolution: (512,512)
value: 71.74
lr schd: 20000
memory (GB): 9.5
Name: ann_r101-d8_512x512_20k_voc12aug
Results:
- Task: Semantic Segmentation
Dataset: Pascal VOC 2012 + Aug
Metrics:
mIoU: 77.47
mIoU(ms+flip): 78.7
Task: Semantic Segmentation
Config: configs/ann/ann_r101-d8_512x512_20k_voc12aug.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/ann/ann_r101-d8_512x512_20k_voc12aug/ann_r101-d8_512x512_20k_voc12aug_20200617_222246-2fad0042.pth
- Config: configs/ann/ann_r50-d8_512x512_40k_voc12aug.py
- Name: ann_r50-d8_512x512_40k_voc12aug
In Collection: ann
Metadata:
backbone: R-50-D8
crop size: (512,512)
lr schd: 40000
Name: ann_r50-d8_512x512_40k_voc12aug
Results:
- Task: Semantic Segmentation
Dataset: Pascal VOC 2012 + Aug
Metrics:
mIoU: 76.56
mIoU(ms+flip): 77.51
Task: Semantic Segmentation
Config: configs/ann/ann_r50-d8_512x512_40k_voc12aug.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/ann/ann_r50-d8_512x512_40k_voc12aug/ann_r50-d8_512x512_40k_voc12aug_20200613_231314-b5dac322.pth
- Config: configs/ann/ann_r101-d8_512x512_40k_voc12aug.py
- Name: ann_r101-d8_512x512_40k_voc12aug
In Collection: ann
Metadata:
backbone: R-101-D8
crop size: (512,512)
lr schd: 40000
Name: ann_r101-d8_512x512_40k_voc12aug
Results:
- Task: Semantic Segmentation
Dataset: Pascal VOC 2012 + Aug
Metrics:
mIoU: 76.7
mIoU(ms+flip): 78.06
Task: Semantic Segmentation
Config: configs/ann/ann_r101-d8_512x512_40k_voc12aug.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/ann/ann_r101-d8_512x512_40k_voc12aug/ann_r101-d8_512x512_40k_voc12aug_20200613_231314-bd205bbe.pth

9
configs/apcnet/README.md
View File

@ -4,6 +4,13 @@
<!-- [ALGORITHM] -->
<a href="https://github.com/Junjun2016/APCNet">Official Repo</a>
<a href="https://github.com/open-mmlab/mmsegmentation/blob/v0.17.0/mmseg/models/decode_heads/apc_head.py#L111">Code Snippet</a>
<details>
<summary align="right"><a href="https://openaccess.thecvf.com/content_CVPR_2019/html/He_Adaptive_Pyramid_Context_Network_for_Semantic_Segmentation_CVPR_2019_paper.html">APCNet (CVPR'2019)</a></summary>
```latex
@InProceedings{He_2019_CVPR,
author = {He, Junjun and Deng, Zhongying and Zhou, Lei and Wang, Yali and Qiao, Yu},
@ -14,6 +21,8 @@ year = {2019}
}
```
</details>
## Results and models
### Cityscapes

133
configs/apcnet/apcnet.yml
View File

@ -1,223 +1,232 @@
Collections:
- Metadata:
- Name: apcnet
Metadata:
Training Data:
- Cityscapes
- ADE20K
Name: apcnet
Paper:
URL: https://openaccess.thecvf.com/content_CVPR_2019/html/He_Adaptive_Pyramid_Context_Network_for_Semantic_Segmentation_CVPR_2019_paper.html
Title: Adaptive Pyramid Context Network for Semantic Segmentation
README: configs/apcnet/README.md
Code:
URL: https://github.com/open-mmlab/mmsegmentation/blob/v0.17.0/mmseg/models/decode_heads/apc_head.py#L111
Version: v0.17.0
Converted From:
Code: https://github.com/Junjun2016/APCNet
Models:
- Config: configs/apcnet/apcnet_r50-d8_512x1024_40k_cityscapes.py
- Name: apcnet_r50-d8_512x1024_40k_cityscapes
In Collection: apcnet
Metadata:
backbone: R-50-D8
crop size: (512,1024)
lr schd: 40000
inference time (ms/im):
- backend: PyTorch
batch size: 1
- value: 280.11
hardware: V100
backend: PyTorch
batch size: 1
mode: FP32
resolution: (512,1024)
value: 280.11
lr schd: 40000
memory (GB): 7.7
Name: apcnet_r50-d8_512x1024_40k_cityscapes
Results:
- Task: Semantic Segmentation
Dataset: Cityscapes
Metrics:
mIoU: 78.02
mIoU(ms+flip): 79.26
Task: Semantic Segmentation
Config: configs/apcnet/apcnet_r50-d8_512x1024_40k_cityscapes.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/apcnet/apcnet_r50-d8_512x1024_40k_cityscapes/apcnet_r50-d8_512x1024_40k_cityscapes_20201214_115717-5e88fa33.pth
- Config: configs/apcnet/apcnet_r101-d8_512x1024_40k_cityscapes.py
- Name: apcnet_r101-d8_512x1024_40k_cityscapes
In Collection: apcnet
Metadata:
backbone: R-101-D8
crop size: (512,1024)
lr schd: 40000
inference time (ms/im):
- backend: PyTorch
batch size: 1
- value: 465.12
hardware: V100
backend: PyTorch
batch size: 1
mode: FP32
resolution: (512,1024)
value: 465.12
lr schd: 40000
memory (GB): 11.2
Name: apcnet_r101-d8_512x1024_40k_cityscapes
Results:
- Task: Semantic Segmentation
Dataset: Cityscapes
Metrics:
mIoU: 79.08
mIoU(ms+flip): 80.34
Task: Semantic Segmentation
Config: configs/apcnet/apcnet_r101-d8_512x1024_40k_cityscapes.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/apcnet/apcnet_r101-d8_512x1024_40k_cityscapes/apcnet_r101-d8_512x1024_40k_cityscapes_20201214_115716-abc9d111.pth
- Config: configs/apcnet/apcnet_r50-d8_769x769_40k_cityscapes.py
- Name: apcnet_r50-d8_769x769_40k_cityscapes
In Collection: apcnet
Metadata:
backbone: R-50-D8
crop size: (769,769)
lr schd: 40000
inference time (ms/im):
- backend: PyTorch
batch size: 1
- value: 657.89
hardware: V100
backend: PyTorch
batch size: 1
mode: FP32
resolution: (769,769)
value: 657.89
lr schd: 40000
memory (GB): 8.7
Name: apcnet_r50-d8_769x769_40k_cityscapes
Results:
- Task: Semantic Segmentation
Dataset: Cityscapes
Metrics:
mIoU: 77.89
mIoU(ms+flip): 79.75
Task: Semantic Segmentation
Config: configs/apcnet/apcnet_r50-d8_769x769_40k_cityscapes.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/apcnet/apcnet_r50-d8_769x769_40k_cityscapes/apcnet_r50-d8_769x769_40k_cityscapes_20201214_115717-2a2628d7.pth
- Config: configs/apcnet/apcnet_r101-d8_769x769_40k_cityscapes.py
- Name: apcnet_r101-d8_769x769_40k_cityscapes
In Collection: apcnet
Metadata:
backbone: R-101-D8
crop size: (769,769)
lr schd: 40000
inference time (ms/im):
- backend: PyTorch
batch size: 1
- value: 970.87
hardware: V100
backend: PyTorch
batch size: 1
mode: FP32
resolution: (769,769)
value: 970.87
lr schd: 40000
memory (GB): 12.7
Name: apcnet_r101-d8_769x769_40k_cityscapes
Results:
- Task: Semantic Segmentation
Dataset: Cityscapes
Metrics:
mIoU: 77.96
mIoU(ms+flip): 79.24
Task: Semantic Segmentation
Config: configs/apcnet/apcnet_r101-d8_769x769_40k_cityscapes.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/apcnet/apcnet_r101-d8_769x769_40k_cityscapes/apcnet_r101-d8_769x769_40k_cityscapes_20201214_115718-b650de90.pth
- Config: configs/apcnet/apcnet_r50-d8_512x1024_80k_cityscapes.py
- Name: apcnet_r50-d8_512x1024_80k_cityscapes
In Collection: apcnet
Metadata:
backbone: R-50-D8
crop size: (512,1024)
lr schd: 80000
Name: apcnet_r50-d8_512x1024_80k_cityscapes
Results:
- Task: Semantic Segmentation
Dataset: Cityscapes
Metrics:
mIoU: 78.96
mIoU(ms+flip): 79.94
Task: Semantic Segmentation
Config: configs/apcnet/apcnet_r50-d8_512x1024_80k_cityscapes.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/apcnet/apcnet_r50-d8_512x1024_80k_cityscapes/apcnet_r50-d8_512x1024_80k_cityscapes_20201214_115716-987f51e3.pth
- Config: configs/apcnet/apcnet_r101-d8_512x1024_80k_cityscapes.py
- Name: apcnet_r101-d8_512x1024_80k_cityscapes
In Collection: apcnet
Metadata:
backbone: R-101-D8
crop size: (512,1024)
lr schd: 80000
Name: apcnet_r101-d8_512x1024_80k_cityscapes
Results:
- Task: Semantic Segmentation
Dataset: Cityscapes
Metrics:
mIoU: 79.64
mIoU(ms+flip): 80.61
Task: Semantic Segmentation
Config: configs/apcnet/apcnet_r101-d8_512x1024_80k_cityscapes.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/apcnet/apcnet_r101-d8_512x1024_80k_cityscapes/apcnet_r101-d8_512x1024_80k_cityscapes_20201214_115705-b1ff208a.pth
- Config: configs/apcnet/apcnet_r50-d8_769x769_80k_cityscapes.py
- Name: apcnet_r50-d8_769x769_80k_cityscapes
In Collection: apcnet
Metadata:
backbone: R-50-D8
crop size: (769,769)
lr schd: 80000
Name: apcnet_r50-d8_769x769_80k_cityscapes
Results:
- Task: Semantic Segmentation
Dataset: Cityscapes
Metrics:
mIoU: 78.79
mIoU(ms+flip): 80.35
Task: Semantic Segmentation
Config: configs/apcnet/apcnet_r50-d8_769x769_80k_cityscapes.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/apcnet/apcnet_r50-d8_769x769_80k_cityscapes/apcnet_r50-d8_769x769_80k_cityscapes_20201214_115718-7ea9fa12.pth
- Config: configs/apcnet/apcnet_r101-d8_769x769_80k_cityscapes.py
- Name: apcnet_r101-d8_769x769_80k_cityscapes
In Collection: apcnet
Metadata:
backbone: R-101-D8
crop size: (769,769)
lr schd: 80000
Name: apcnet_r101-d8_769x769_80k_cityscapes
Results:
- Task: Semantic Segmentation
Dataset: Cityscapes
Metrics:
mIoU: 78.45
mIoU(ms+flip): 79.91
Task: Semantic Segmentation
Config: configs/apcnet/apcnet_r101-d8_769x769_80k_cityscapes.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/apcnet/apcnet_r101-d8_769x769_80k_cityscapes/apcnet_r101-d8_769x769_80k_cityscapes_20201214_115716-a7fbc2ab.pth
- Config: configs/apcnet/apcnet_r50-d8_512x512_80k_ade20k.py
- Name: apcnet_r50-d8_512x512_80k_ade20k
In Collection: apcnet
Metadata:
backbone: R-50-D8
crop size: (512,512)
lr schd: 80000
inference time (ms/im):
- backend: PyTorch
batch size: 1
- value: 50.99
hardware: V100
backend: PyTorch
batch size: 1
mode: FP32
resolution: (512,512)
value: 50.99
lr schd: 80000
memory (GB): 10.1
Name: apcnet_r50-d8_512x512_80k_ade20k
Results:
- Task: Semantic Segmentation
Dataset: ADE20K
Metrics:
mIoU: 42.2
mIoU(ms+flip): 43.3
Task: Semantic Segmentation
Config: configs/apcnet/apcnet_r50-d8_512x512_80k_ade20k.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/apcnet/apcnet_r50-d8_512x512_80k_ade20k/apcnet_r50-d8_512x512_80k_ade20k_20201214_115705-a8626293.pth
- Config: configs/apcnet/apcnet_r101-d8_512x512_80k_ade20k.py
- Name: apcnet_r101-d8_512x512_80k_ade20k
In Collection: apcnet
Metadata:
backbone: R-101-D8
crop size: (512,512)
lr schd: 80000
inference time (ms/im):
- backend: PyTorch
batch size: 1
- value: 76.34
hardware: V100
backend: PyTorch
batch size: 1
mode: FP32
resolution: (512,512)
value: 76.34
lr schd: 80000
memory (GB): 13.6
Name: apcnet_r101-d8_512x512_80k_ade20k
Results:
- Task: Semantic Segmentation
Dataset: ADE20K
Metrics:
mIoU: 45.54
mIoU(ms+flip): 46.65
Task: Semantic Segmentation
Config: configs/apcnet/apcnet_r101-d8_512x512_80k_ade20k.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/apcnet/apcnet_r101-d8_512x512_80k_ade20k/apcnet_r101-d8_512x512_80k_ade20k_20201214_115704-c656c3fb.pth
- Config: configs/apcnet/apcnet_r50-d8_512x512_160k_ade20k.py
- Name: apcnet_r50-d8_512x512_160k_ade20k
In Collection: apcnet
Metadata:
backbone: R-50-D8
crop size: (512,512)
lr schd: 160000
Name: apcnet_r50-d8_512x512_160k_ade20k
Results:
- Task: Semantic Segmentation
Dataset: ADE20K
Metrics:
mIoU: 43.4
mIoU(ms+flip): 43.94
Task: Semantic Segmentation
Config: configs/apcnet/apcnet_r50-d8_512x512_160k_ade20k.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/apcnet/apcnet_r50-d8_512x512_160k_ade20k/apcnet_r50-d8_512x512_160k_ade20k_20201214_115706-25fb92c2.pth
- Config: configs/apcnet/apcnet_r101-d8_512x512_160k_ade20k.py
- Name: apcnet_r101-d8_512x512_160k_ade20k
In Collection: apcnet
Metadata:
backbone: R-101-D8
crop size: (512,512)
lr schd: 160000
Name: apcnet_r101-d8_512x512_160k_ade20k
Results:
- Task: Semantic Segmentation
Dataset: ADE20K
Metrics:
mIoU: 45.41
mIoU(ms+flip): 46.63
Task: Semantic Segmentation
Config: configs/apcnet/apcnet_r101-d8_512x512_160k_ade20k.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/apcnet/apcnet_r101-d8_512x512_160k_ade20k/apcnet_r101-d8_512x512_160k_ade20k_20201214_115705-73f9a8d7.pth

9
configs/bisenetv2/README.md
View File

@ -4,6 +4,13 @@
<!-- [ALGORITHM] -->
<a href="">Official Repo</a>
<a href="https://github.com/open-mmlab/mmsegmentation/blob/v0.18.0/mmseg/models/backbones/bisenetv2.py#L545">Code Snippet</a>
<details>
<summary align="right"><a href="https://arxiv.org/abs/2004.02147">BiSeNetV2 (IJCV'2021)</a></summary>
```latex
@article{yu2021bisenet,
title={Bisenet v2: Bilateral network with guided aggregation for real-time semantic segmentation},
@ -15,6 +22,8 @@
}
```
</details>
## Results and models
### Cityscapes

52
configs/bisenetv2/bisenetv2.yml
View File

@ -1,80 +1,88 @@
Collections:
- Metadata:
- Name: bisenetv2
Metadata:
Training Data:
- Cityscapes
Name: bisenetv2
Paper:
URL: https://arxiv.org/abs/2004.02147
Title: 'Bisenet v2: Bilateral Network with Guided Aggregation for Real-time Semantic
Segmentation'
README: configs/bisenetv2/README.md
Code:
URL: https://github.com/open-mmlab/mmsegmentation/blob/v0.18.0/mmseg/models/backbones/bisenetv2.py#L545
Version: v0.18.0
Models:
- Config: configs/bisenetv2/bisenetv2_fcn_4x4_1024x1024_160k_cityscapes.py
- Name: bisenetv2_fcn_4x4_1024x1024_160k_cityscapes
In Collection: bisenetv2
Metadata:
backbone: BiSeNetV2
crop size: (1024,1024)
lr schd: 160000
inference time (ms/im):
- backend: PyTorch
batch size: 1
- value: 31.48
hardware: V100
backend: PyTorch
batch size: 1
mode: FP32
resolution: (1024,1024)
value: 31.48
lr schd: 160000
memory (GB): 7.64
Name: bisenetv2_fcn_4x4_1024x1024_160k_cityscapes
Results:
- Task: Semantic Segmentation
Dataset: Cityscapes
Metrics:
mIoU: 73.21
mIoU(ms+flip): 75.74
Task: Semantic Segmentation
Config: configs/bisenetv2/bisenetv2_fcn_4x4_1024x1024_160k_cityscapes.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/bisenetv2/bisenetv2_fcn_4x4_1024x1024_160k_cityscapes/bisenetv2_fcn_4x4_1024x1024_160k_cityscapes_20210902_015551-bcf10f09.pth
- Config: configs/bisenetv2/bisenetv2_fcn_ohem_4x4_1024x1024_160k_cityscapes.py
- Name: bisenetv2_fcn_ohem_4x4_1024x1024_160k_cityscapes
In Collection: bisenetv2
Metadata:
backbone: BiSeNetV2
crop size: (1024,1024)
lr schd: 160000
memory (GB): 7.64
Name: bisenetv2_fcn_ohem_4x4_1024x1024_160k_cityscapes
Results:
- Task: Semantic Segmentation
Dataset: Cityscapes
Metrics:
mIoU: 73.57
mIoU(ms+flip): 75.8
Task: Semantic Segmentation
Config: configs/bisenetv2/bisenetv2_fcn_ohem_4x4_1024x1024_160k_cityscapes.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/bisenetv2/bisenetv2_fcn_ohem_4x4_1024x1024_160k_cityscapes/bisenetv2_fcn_ohem_4x4_1024x1024_160k_cityscapes_20210902_112947-5f8103b4.pth
- Config: configs/bisenetv2/bisenetv2_fcn_4x8_1024x1024_160k_cityscapes.py
- Name: bisenetv2_fcn_4x8_1024x1024_160k_cityscapes
In Collection: bisenetv2
Metadata:
backbone: BiSeNetV2
crop size: (1024,1024)
lr schd: 160000
memory (GB): 15.05
Name: bisenetv2_fcn_4x8_1024x1024_160k_cityscapes
Results:
- Task: Semantic Segmentation
Dataset: Cityscapes
Metrics:
mIoU: 75.76
mIoU(ms+flip): 77.79
Task: Semantic Segmentation
Config: configs/bisenetv2/bisenetv2_fcn_4x8_1024x1024_160k_cityscapes.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/bisenetv2/bisenetv2_fcn_4x8_1024x1024_160k_cityscapes/bisenetv2_fcn_4x8_1024x1024_160k_cityscapes_20210903_000032-e1a2eed6.pth
- Config: configs/bisenetv2/bisenetv2_fcn_fp16_4x4_1024x1024_160k_cityscapes.py
- Name: bisenetv2_fcn_fp16_4x4_1024x1024_160k_cityscapes
In Collection: bisenetv2
Metadata:
backbone: BiSeNetV2
crop size: (1024,1024)
lr schd: 160000
inference time (ms/im):
- backend: PyTorch
batch size: 1
- value: 27.29
hardware: V100
backend: PyTorch
batch size: 1
mode: FP32
resolution: (1024,1024)
value: 27.29
lr schd: 160000
memory (GB): 5.77
Name: bisenetv2_fcn_fp16_4x4_1024x1024_160k_cityscapes
Results:
- Task: Semantic Segmentation
Dataset: Cityscapes
Metrics:
mIoU: 73.07
mIoU(ms+flip): 75.13
Task: Semantic Segmentation
Config: configs/bisenetv2/bisenetv2_fcn_fp16_4x4_1024x1024_160k_cityscapes.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/bisenetv2/bisenetv2_fcn_fp16_4x4_1024x1024_160k_cityscapes/bisenetv2_fcn_fp16_4x4_1024x1024_160k_cityscapes_20210902_045942-b979777b.pth

9
configs/ccnet/README.md
View File

@ -4,6 +4,13 @@
<!-- [ALGORITHM] -->
<a href="https://github.com/speedinghzl/CCNet">Official Repo</a>
<a href="https://github.com/open-mmlab/mmsegmentation/blob/v0.17.0/mmseg/models/decode_heads/apc_head.py#L111">Code Snippet</a>
<details>
<summary align="right"><a href="https://arxiv.org/abs/1811.11721">CCNet (ICCV'2019)</a></summary>
```latex
@article{huang2018ccnet,
title={CCNet: Criss-Cross Attention for Semantic Segmentation},
@ -13,6 +20,8 @@
}
```
</details>
## Results and models
### Cityscapes

173
configs/ccnet/ccnet.yml
View File

@ -1,296 +1,305 @@
Collections:
- Metadata:
- Name: ccnet
Metadata:
Training Data:
- Cityscapes
- ADE20K
- Pascal VOC 2012 + Aug
Name: ccnet
Paper:
URL: https://arxiv.org/abs/1811.11721
Title: 'CCNet: Criss-Cross Attention for Semantic Segmentation'
README: configs/ccnet/README.md
Code:
URL: https://github.com/open-mmlab/mmsegmentation/blob/v0.17.0/mmseg/models/decode_heads/apc_head.py#L111
Version: v0.17.0
Converted From:
Code: https://github.com/speedinghzl/CCNet
Models:
- Config: configs/ccnet/ccnet_r50-d8_512x1024_40k_cityscapes.py
- Name: ccnet_r50-d8_512x1024_40k_cityscapes
In Collection: ccnet
Metadata:
backbone: R-50-D8
crop size: (512,1024)
lr schd: 40000
inference time (ms/im):
- backend: PyTorch
batch size: 1
- value: 301.2
hardware: V100
backend: PyTorch
batch size: 1
mode: FP32
resolution: (512,1024)
value: 301.2
lr schd: 40000
memory (GB): 6.0
Name: ccnet_r50-d8_512x1024_40k_cityscapes
Results:
- Task: Semantic Segmentation
Dataset: Cityscapes
Metrics:
mIoU: 77.76
mIoU(ms+flip): 78.87
Task: Semantic Segmentation
Config: configs/ccnet/ccnet_r50-d8_512x1024_40k_cityscapes.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/ccnet/ccnet_r50-d8_512x1024_40k_cityscapes/ccnet_r50-d8_512x1024_40k_cityscapes_20200616_142517-4123f401.pth
- Config: configs/ccnet/ccnet_r101-d8_512x1024_40k_cityscapes.py
- Name: ccnet_r101-d8_512x1024_40k_cityscapes
In Collection: ccnet
Metadata:
backbone: R-101-D8
crop size: (512,1024)
lr schd: 40000
inference time (ms/im):
- backend: PyTorch
batch size: 1
- value: 432.9
hardware: V100
backend: PyTorch
batch size: 1
mode: FP32
resolution: (512,1024)
value: 432.9
lr schd: 40000
memory (GB): 9.5
Name: ccnet_r101-d8_512x1024_40k_cityscapes
Results:
- Task: Semantic Segmentation
Dataset: Cityscapes
Metrics:
mIoU: 76.35
mIoU(ms+flip): 78.19
Task: Semantic Segmentation
Config: configs/ccnet/ccnet_r101-d8_512x1024_40k_cityscapes.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/ccnet/ccnet_r101-d8_512x1024_40k_cityscapes/ccnet_r101-d8_512x1024_40k_cityscapes_20200616_142540-a3b84ba6.pth
- Config: configs/ccnet/ccnet_r50-d8_769x769_40k_cityscapes.py
- Name: ccnet_r50-d8_769x769_40k_cityscapes
In Collection: ccnet
Metadata:
backbone: R-50-D8
crop size: (769,769)
lr schd: 40000
inference time (ms/im):
- backend: PyTorch
batch size: 1
- value: 699.3
hardware: V100
backend: PyTorch
batch size: 1
mode: FP32
resolution: (769,769)
value: 699.3
lr schd: 40000
memory (GB): 6.8
Name: ccnet_r50-d8_769x769_40k_cityscapes
Results:
- Task: Semantic Segmentation
Dataset: Cityscapes
Metrics:
mIoU: 78.46
mIoU(ms+flip): 79.93
Task: Semantic Segmentation
Config: configs/ccnet/ccnet_r50-d8_769x769_40k_cityscapes.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/ccnet/ccnet_r50-d8_769x769_40k_cityscapes/ccnet_r50-d8_769x769_40k_cityscapes_20200616_145125-76d11884.pth
- Config: configs/ccnet/ccnet_r101-d8_769x769_40k_cityscapes.py
- Name: ccnet_r101-d8_769x769_40k_cityscapes
In Collection: ccnet
Metadata:
backbone: R-101-D8
crop size: (769,769)
lr schd: 40000
inference time (ms/im):
- backend: PyTorch
batch size: 1
- value: 990.1
hardware: V100
backend: PyTorch
batch size: 1
mode: FP32
resolution: (769,769)
value: 990.1
lr schd: 40000
memory (GB): 10.7
Name: ccnet_r101-d8_769x769_40k_cityscapes
Results:
- Task: Semantic Segmentation
Dataset: Cityscapes
Metrics:
mIoU: 76.94
mIoU(ms+flip): 78.62
Task: Semantic Segmentation
Config: configs/ccnet/ccnet_r101-d8_769x769_40k_cityscapes.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/ccnet/ccnet_r101-d8_769x769_40k_cityscapes/ccnet_r101-d8_769x769_40k_cityscapes_20200617_101428-4f57c8d0.pth
- Config: configs/ccnet/ccnet_r50-d8_512x1024_80k_cityscapes.py
- Name: ccnet_r50-d8_512x1024_80k_cityscapes
In Collection: ccnet
Metadata:
backbone: R-50-D8
crop size: (512,1024)
lr schd: 80000
Name: ccnet_r50-d8_512x1024_80k_cityscapes
Results:
- Task: Semantic Segmentation
Dataset: Cityscapes
Metrics:
mIoU: 79.03
mIoU(ms+flip): 80.16
Task: Semantic Segmentation
Config: configs/ccnet/ccnet_r50-d8_512x1024_80k_cityscapes.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/ccnet/ccnet_r50-d8_512x1024_80k_cityscapes/ccnet_r50-d8_512x1024_80k_cityscapes_20200617_010421-869a3423.pth
- Config: configs/ccnet/ccnet_r101-d8_512x1024_80k_cityscapes.py
- Name: ccnet_r101-d8_512x1024_80k_cityscapes
In Collection: ccnet
Metadata:
backbone: R-101-D8
crop size: (512,1024)
lr schd: 80000
Name: ccnet_r101-d8_512x1024_80k_cityscapes
Results:
- Task: Semantic Segmentation
Dataset: Cityscapes
Metrics:
mIoU: 78.87
mIoU(ms+flip): 79.9
Task: Semantic Segmentation
Config: configs/ccnet/ccnet_r101-d8_512x1024_80k_cityscapes.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/ccnet/ccnet_r101-d8_512x1024_80k_cityscapes/ccnet_r101-d8_512x1024_80k_cityscapes_20200617_203935-ffae8917.pth
- Config: configs/ccnet/ccnet_r50-d8_769x769_80k_cityscapes.py
- Name: ccnet_r50-d8_769x769_80k_cityscapes
In Collection: ccnet
Metadata:
backbone: R-50-D8
crop size: (769,769)
lr schd: 80000
Name: ccnet_r50-d8_769x769_80k_cityscapes
Results:
- Task: Semantic Segmentation
Dataset: Cityscapes
Metrics:
mIoU: 79.29
mIoU(ms+flip): 81.08
Task: Semantic Segmentation
Config: configs/ccnet/ccnet_r50-d8_769x769_80k_cityscapes.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/ccnet/ccnet_r50-d8_769x769_80k_cityscapes/ccnet_r50-d8_769x769_80k_cityscapes_20200617_010421-73eed8ca.pth
- Config: configs/ccnet/ccnet_r101-d8_769x769_80k_cityscapes.py
- Name: ccnet_r101-d8_769x769_80k_cityscapes
In Collection: ccnet
Metadata:
backbone: R-101-D8
crop size: (769,769)
lr schd: 80000
Name: ccnet_r101-d8_769x769_80k_cityscapes
Results:
- Task: Semantic Segmentation
Dataset: Cityscapes
Metrics:
mIoU: 79.45
mIoU(ms+flip): 80.66
Task: Semantic Segmentation
Config: configs/ccnet/ccnet_r101-d8_769x769_80k_cityscapes.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/ccnet/ccnet_r101-d8_769x769_80k_cityscapes/ccnet_r101-d8_769x769_80k_cityscapes_20200618_011502-ad3cd481.pth
- Config: configs/ccnet/ccnet_r50-d8_512x512_80k_ade20k.py
- Name: ccnet_r50-d8_512x512_80k_ade20k
In Collection: ccnet
Metadata:
backbone: R-50-D8
crop size: (512,512)
lr schd: 80000
inference time (ms/im):
- backend: PyTorch
batch size: 1
- value: 47.87
hardware: V100
backend: PyTorch
batch size: 1
mode: FP32
resolution: (512,512)
value: 47.87
lr schd: 80000
memory (GB): 8.8
Name: ccnet_r50-d8_512x512_80k_ade20k
Results:
- Task: Semantic Segmentation
Dataset: ADE20K
Metrics:
mIoU: 41.78
mIoU(ms+flip): 42.98
Task: Semantic Segmentation
Config: configs/ccnet/ccnet_r50-d8_512x512_80k_ade20k.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/ccnet/ccnet_r50-d8_512x512_80k_ade20k/ccnet_r50-d8_512x512_80k_ade20k_20200615_014848-aa37f61e.pth
- Config: configs/ccnet/ccnet_r101-d8_512x512_80k_ade20k.py
- Name: ccnet_r101-d8_512x512_80k_ade20k
In Collection: ccnet
Metadata:
backbone: R-101-D8
crop size: (512,512)
lr schd: 80000
inference time (ms/im):
- backend: PyTorch
batch size: 1
- value: 70.87
hardware: V100
backend: PyTorch
batch size: 1
mode: FP32
resolution: (512,512)
value: 70.87
lr schd: 80000
memory (GB): 12.2
Name: ccnet_r101-d8_512x512_80k_ade20k
Results:
- Task: Semantic Segmentation
Dataset: ADE20K
Metrics:
mIoU: 43.97
mIoU(ms+flip): 45.13
Task: Semantic Segmentation
Config: configs/ccnet/ccnet_r101-d8_512x512_80k_ade20k.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/ccnet/ccnet_r101-d8_512x512_80k_ade20k/ccnet_r101-d8_512x512_80k_ade20k_20200615_014848-1f4929a3.pth
- Config: configs/ccnet/ccnet_r50-d8_512x512_160k_ade20k.py
- Name: ccnet_r50-d8_512x512_160k_ade20k
In Collection: ccnet
Metadata:
backbone: R-50-D8
crop size: (512,512)
lr schd: 160000
Name: ccnet_r50-d8_512x512_160k_ade20k
Results:
- Task: Semantic Segmentation
Dataset: ADE20K
Metrics:
mIoU: 42.08
mIoU(ms+flip): 43.13
Task: Semantic Segmentation
Config: configs/ccnet/ccnet_r50-d8_512x512_160k_ade20k.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/ccnet/ccnet_r50-d8_512x512_160k_ade20k/ccnet_r50-d8_512x512_160k_ade20k_20200616_084435-7c97193b.pth
- Config: configs/ccnet/ccnet_r101-d8_512x512_160k_ade20k.py
- Name: ccnet_r101-d8_512x512_160k_ade20k
In Collection: ccnet
Metadata:
backbone: R-101-D8
crop size: (512,512)
lr schd: 160000
Name: ccnet_r101-d8_512x512_160k_ade20k
Results:
- Task: Semantic Segmentation
Dataset: ADE20K
Metrics:
mIoU: 43.71
mIoU(ms+flip): 45.04
Task: Semantic Segmentation
Config: configs/ccnet/ccnet_r101-d8_512x512_160k_ade20k.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/ccnet/ccnet_r101-d8_512x512_160k_ade20k/ccnet_r101-d8_512x512_160k_ade20k_20200616_000644-e849e007.pth
- Config: configs/ccnet/ccnet_r50-d8_512x512_20k_voc12aug.py
- Name: ccnet_r50-d8_512x512_20k_voc12aug
In Collection: ccnet
Metadata:
backbone: R-50-D8
crop size: (512,512)
lr schd: 20000
inference time (ms/im):
- backend: PyTorch
batch size: 1
- value: 48.9
hardware: V100
backend: PyTorch
batch size: 1
mode: FP32
resolution: (512,512)
value: 48.9
lr schd: 20000
memory (GB): 6.0
Name: ccnet_r50-d8_512x512_20k_voc12aug
Results:
- Task: Semantic Segmentation
Dataset: Pascal VOC 2012 + Aug
Metrics:
mIoU: 76.17
mIoU(ms+flip): 77.51
Task: Semantic Segmentation
Config: configs/ccnet/ccnet_r50-d8_512x512_20k_voc12aug.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/ccnet/ccnet_r50-d8_512x512_20k_voc12aug/ccnet_r50-d8_512x512_20k_voc12aug_20200617_193212-fad81784.pth
- Config: configs/ccnet/ccnet_r101-d8_512x512_20k_voc12aug.py
- Name: ccnet_r101-d8_512x512_20k_voc12aug
In Collection: ccnet
Metadata:
backbone: R-101-D8
crop size: (512,512)
lr schd: 20000
inference time (ms/im):
- backend: PyTorch
batch size: 1
- value: 73.31
hardware: V100
backend: PyTorch
batch size: 1
mode: FP32
resolution: (512,512)
value: 73.31
lr schd: 20000
memory (GB): 9.5
Name: ccnet_r101-d8_512x512_20k_voc12aug
Results:
- Task: Semantic Segmentation
Dataset: Pascal VOC 2012 + Aug
Metrics:
mIoU: 77.27
mIoU(ms+flip): 79.02
Task: Semantic Segmentation
Config: configs/ccnet/ccnet_r101-d8_512x512_20k_voc12aug.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/ccnet/ccnet_r101-d8_512x512_20k_voc12aug/ccnet_r101-d8_512x512_20k_voc12aug_20200617_193212-0007b61d.pth
- Config: configs/ccnet/ccnet_r50-d8_512x512_40k_voc12aug.py
- Name: ccnet_r50-d8_512x512_40k_voc12aug
In Collection: ccnet
Metadata:
backbone: R-50-D8
crop size: (512,512)
lr schd: 40000
Name: ccnet_r50-d8_512x512_40k_voc12aug
Results:
- Task: Semantic Segmentation
Dataset: Pascal VOC 2012 + Aug
Metrics:
mIoU: 75.96
mIoU(ms+flip): 77.04
Task: Semantic Segmentation
Config: configs/ccnet/ccnet_r50-d8_512x512_40k_voc12aug.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/ccnet/ccnet_r50-d8_512x512_40k_voc12aug/ccnet_r50-d8_512x512_40k_voc12aug_20200613_232127-c2a15f02.pth
- Config: configs/ccnet/ccnet_r101-d8_512x512_40k_voc12aug.py
- Name: ccnet_r101-d8_512x512_40k_voc12aug
In Collection: ccnet
Metadata:
backbone: R-101-D8
crop size: (512,512)
lr schd: 40000
Name: ccnet_r101-d8_512x512_40k_voc12aug
Results:
- Task: Semantic Segmentation
Dataset: Pascal VOC 2012 + Aug
Metrics:
mIoU: 77.87
mIoU(ms+flip): 78.9
Task: Semantic Segmentation
Config: configs/ccnet/ccnet_r101-d8_512x512_40k_voc12aug.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/ccnet/ccnet_r101-d8_512x512_40k_voc12aug/ccnet_r101-d8_512x512_40k_voc12aug_20200613_232127-c30da577.pth

9
configs/cgnet/README.md
View File

@ -4,6 +4,13 @@
<!-- [ALGORITHM] -->
<a href="https://github.com/wutianyiRosun/CGNet">Official Repo</a>
<a href="https://github.com/open-mmlab/mmsegmentation/blob/v0.17.0/mmseg/models/backbones/cgnet.py#L187">Code Snippet</a>
<details>
<summary align="right"><a href="https://arxiv.org/pdf/1811.08201.pdf">CGNet (TIP'2020)</a></summary>
```latext
@article{wu2020cgnet,
title={Cgnet: A light-weight context guided network for semantic segmentation},
@ -16,6 +23,8 @@
}
```
</details>
## Results and models
### Cityscapes

41
configs/cgnet/cgnet.yml
View File

@ -1,50 +1,59 @@
Collections:
- Metadata:
- Name: cgnet
Metadata:
Training Data:
- Cityscapes
Name: cgnet
Paper:
URL: https://arxiv.org/pdf/1811.08201.pdf
Title: 'CGNet: A Light-weight Context Guided Network for Semantic Segmentation'
README: configs/cgnet/README.md
Code:
URL: https://github.com/open-mmlab/mmsegmentation/blob/v0.17.0/mmseg/models/backbones/cgnet.py#L187
Version: v0.17.0
Converted From:
Code: https://github.com/wutianyiRosun/CGNet
Models:
- Config: configs/cgnet/cgnet_680x680_60k_cityscapes.py
- Name: cgnet_680x680_60k_cityscapes
In Collection: cgnet
Metadata:
backbone: M3N21
crop size: (680,680)
lr schd: 60000
inference time (ms/im):
- backend: PyTorch
batch size: 1
- value: 32.78
hardware: V100
backend: PyTorch
batch size: 1
mode: FP32
resolution: (680,680)
value: 32.78
lr schd: 60000
memory (GB): 7.5
Name: cgnet_680x680_60k_cityscapes
Results:
- Task: Semantic Segmentation
Dataset: Cityscapes
Metrics:
mIoU: 65.63
mIoU(ms+flip): 68.04
Task: Semantic Segmentation
Config: configs/cgnet/cgnet_680x680_60k_cityscapes.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/cgnet/cgnet_680x680_60k_cityscapes/cgnet_680x680_60k_cityscapes_20201101_110253-4c0b2f2d.pth
- Config: configs/cgnet/cgnet_512x1024_60k_cityscapes.py
- Name: cgnet_512x1024_60k_cityscapes
In Collection: cgnet
Metadata:
backbone: M3N21
crop size: (512,1024)
lr schd: 60000
inference time (ms/im):
- backend: PyTorch
batch size: 1
- value: 32.11
hardware: V100
backend: PyTorch
batch size: 1
mode: FP32
resolution: (512,1024)
value: 32.11
lr schd: 60000
memory (GB): 8.3
Name: cgnet_512x1024_60k_cityscapes
Results:
- Task: Semantic Segmentation
Dataset: Cityscapes
Metrics:
mIoU: 68.27
mIoU(ms+flip): 70.33
Task: Semantic Segmentation
Config: configs/cgnet/cgnet_512x1024_60k_cityscapes.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/cgnet/cgnet_512x1024_60k_cityscapes/cgnet_512x1024_60k_cityscapes_20201101_110254-124ea03b.pth

9
configs/danet/README.md
View File

@ -4,6 +4,13 @@
<!-- [ALGORITHM] -->
<a href="https://github.com/junfu1115/DANet/">Official Repo</a>
<a href="https://github.com/open-mmlab/mmsegmentation/blob/v0.17.0/mmseg/models/decode_heads/da_head.py#L76">Code Snippet</a>
<details>
<summary align="right"><a href="https://arxiv.org/abs/1809.02983">DANet (CVPR'2019)</a></summary>
```latex
@article{fu2018dual,
title={Dual Attention Network for Scene Segmentation},
@ -13,6 +20,8 @@
}
```
</details>
## Results and models
### Cityscapes

173
configs/danet/danet.yml
View File

@ -1,292 +1,301 @@
Collections:
- Metadata:
- Name: danet
Metadata:
Training Data:
- Cityscapes
- ADE20K
- Pascal VOC 2012 + Aug
Name: danet
Paper:
URL: https://arxiv.org/abs/1809.02983
Title: Dual Attention Network for Scene Segmentation
README: configs/danet/README.md
Code:
URL: https://github.com/open-mmlab/mmsegmentation/blob/v0.17.0/mmseg/models/decode_heads/da_head.py#L76
Version: v0.17.0
Converted From:
Code: https://github.com/junfu1115/DANet/
Models:
- Config: configs/danet/danet_r50-d8_512x1024_40k_cityscapes.py
- Name: danet_r50-d8_512x1024_40k_cityscapes
In Collection: danet
Metadata:
backbone: R-50-D8
crop size: (512,1024)
lr schd: 40000
inference time (ms/im):
- backend: PyTorch
batch size: 1
- value: 375.94
hardware: V100
backend: PyTorch
batch size: 1
mode: FP32
resolution: (512,1024)
value: 375.94
lr schd: 40000
memory (GB): 7.4
Name: danet_r50-d8_512x1024_40k_cityscapes
Results:
- Task: Semantic Segmentation
Dataset: Cityscapes
Metrics:
mIoU: 78.74
Task: Semantic Segmentation
Config: configs/danet/danet_r50-d8_512x1024_40k_cityscapes.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/danet/danet_r50-d8_512x1024_40k_cityscapes/danet_r50-d8_512x1024_40k_cityscapes_20200605_191324-c0dbfa5f.pth
- Config: configs/danet/danet_r101-d8_512x1024_40k_cityscapes.py
- Name: danet_r101-d8_512x1024_40k_cityscapes
In Collection: danet
Metadata:
backbone: R-101-D8
crop size: (512,1024)
lr schd: 40000
inference time (ms/im):
- backend: PyTorch
batch size: 1
- value: 502.51
hardware: V100
backend: PyTorch
batch size: 1
mode: FP32
resolution: (512,1024)
value: 502.51
lr schd: 40000
memory (GB): 10.9
Name: danet_r101-d8_512x1024_40k_cityscapes
Results:
- Task: Semantic Segmentation
Dataset: Cityscapes
Metrics:
mIoU: 80.52
Task: Semantic Segmentation
Config: configs/danet/danet_r101-d8_512x1024_40k_cityscapes.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/danet/danet_r101-d8_512x1024_40k_cityscapes/danet_r101-d8_512x1024_40k_cityscapes_20200605_200831-c57a7157.pth
- Config: configs/danet/danet_r50-d8_769x769_40k_cityscapes.py
- Name: danet_r50-d8_769x769_40k_cityscapes
In Collection: danet
Metadata:
backbone: R-50-D8
crop size: (769,769)
lr schd: 40000
inference time (ms/im):
- backend: PyTorch
batch size: 1
- value: 641.03
hardware: V100
backend: PyTorch
batch size: 1
mode: FP32
resolution: (769,769)
value: 641.03
lr schd: 40000
memory (GB): 8.8
Name: danet_r50-d8_769x769_40k_cityscapes
Results:
- Task: Semantic Segmentation
Dataset: Cityscapes
Metrics:
mIoU: 78.88
mIoU(ms+flip): 80.62
Task: Semantic Segmentation
Config: configs/danet/danet_r50-d8_769x769_40k_cityscapes.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/danet/danet_r50-d8_769x769_40k_cityscapes/danet_r50-d8_769x769_40k_cityscapes_20200530_025703-76681c60.pth
- Config: configs/danet/danet_r101-d8_769x769_40k_cityscapes.py
- Name: danet_r101-d8_769x769_40k_cityscapes
In Collection: danet
Metadata:
backbone: R-101-D8
crop size: (769,769)
lr schd: 40000
inference time (ms/im):
- backend: PyTorch
batch size: 1
- value: 934.58
hardware: V100
backend: PyTorch
batch size: 1
mode: FP32
resolution: (769,769)
value: 934.58
lr schd: 40000
memory (GB): 12.8
Name: danet_r101-d8_769x769_40k_cityscapes
Results:
- Task: Semantic Segmentation
Dataset: Cityscapes
Metrics:
mIoU: 79.88
mIoU(ms+flip): 81.47
Task: Semantic Segmentation
Config: configs/danet/danet_r101-d8_769x769_40k_cityscapes.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/danet/danet_r101-d8_769x769_40k_cityscapes/danet_r101-d8_769x769_40k_cityscapes_20200530_025717-dcb7fd4e.pth
- Config: configs/danet/danet_r50-d8_512x1024_80k_cityscapes.py
- Name: danet_r50-d8_512x1024_80k_cityscapes
In Collection: danet
Metadata:
backbone: R-50-D8
crop size: (512,1024)
lr schd: 80000
Name: danet_r50-d8_512x1024_80k_cityscapes
Results:
- Task: Semantic Segmentation
Dataset: Cityscapes
Metrics:
mIoU: 79.34
Task: Semantic Segmentation
Config: configs/danet/danet_r50-d8_512x1024_80k_cityscapes.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/danet/danet_r50-d8_512x1024_80k_cityscapes/danet_r50-d8_512x1024_80k_cityscapes_20200607_133029-2bfa2293.pth
- Config: configs/danet/danet_r101-d8_512x1024_80k_cityscapes.py
- Name: danet_r101-d8_512x1024_80k_cityscapes
In Collection: danet
Metadata:
backbone: R-101-D8
crop size: (512,1024)
lr schd: 80000
Name: danet_r101-d8_512x1024_80k_cityscapes
Results:
- Task: Semantic Segmentation
Dataset: Cityscapes
Metrics:
mIoU: 80.41
Task: Semantic Segmentation
Config: configs/danet/danet_r101-d8_512x1024_80k_cityscapes.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/danet/danet_r101-d8_512x1024_80k_cityscapes/danet_r101-d8_512x1024_80k_cityscapes_20200607_132918-955e6350.pth
- Config: configs/danet/danet_r50-d8_769x769_80k_cityscapes.py
- Name: danet_r50-d8_769x769_80k_cityscapes
In Collection: danet
Metadata:
backbone: R-50-D8
crop size: (769,769)
lr schd: 80000
Name: danet_r50-d8_769x769_80k_cityscapes
Results:
- Task: Semantic Segmentation
Dataset: Cityscapes
Metrics:
mIoU: 79.27
mIoU(ms+flip): 80.96
Task: Semantic Segmentation
Config: configs/danet/danet_r50-d8_769x769_80k_cityscapes.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/danet/danet_r50-d8_769x769_80k_cityscapes/danet_r50-d8_769x769_80k_cityscapes_20200607_132954-495689b4.pth
- Config: configs/danet/danet_r101-d8_769x769_80k_cityscapes.py
- Name: danet_r101-d8_769x769_80k_cityscapes
In Collection: danet
Metadata:
backbone: R-101-D8
crop size: (769,769)
lr schd: 80000
Name: danet_r101-d8_769x769_80k_cityscapes
Results:
- Task: Semantic Segmentation
Dataset: Cityscapes
Metrics:
mIoU: 80.47
mIoU(ms+flip): 82.02
Task: Semantic Segmentation
Config: configs/danet/danet_r101-d8_769x769_80k_cityscapes.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/danet/danet_r101-d8_769x769_80k_cityscapes/danet_r101-d8_769x769_80k_cityscapes_20200607_132918-f3a929e7.pth
- Config: configs/danet/danet_r50-d8_512x512_80k_ade20k.py
- Name: danet_r50-d8_512x512_80k_ade20k
In Collection: danet
Metadata:
backbone: R-50-D8
crop size: (512,512)
lr schd: 80000
inference time (ms/im):
- backend: PyTorch
batch size: 1
- value: 47.17
hardware: V100
backend: PyTorch
batch size: 1
mode: FP32
resolution: (512,512)
value: 47.17
lr schd: 80000
memory (GB): 11.5
Name: danet_r50-d8_512x512_80k_ade20k
Results:
- Task: Semantic Segmentation
Dataset: ADE20K
Metrics:
mIoU: 41.66
mIoU(ms+flip): 42.9
Task: Semantic Segmentation
Config: configs/danet/danet_r50-d8_512x512_80k_ade20k.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/danet/danet_r50-d8_512x512_80k_ade20k/danet_r50-d8_512x512_80k_ade20k_20200615_015125-edb18e08.pth
- Config: configs/danet/danet_r101-d8_512x512_80k_ade20k.py
- Name: danet_r101-d8_512x512_80k_ade20k
In Collection: danet
Metadata:
backbone: R-101-D8
crop size: (512,512)
lr schd: 80000
inference time (ms/im):
- backend: PyTorch
batch size: 1
- value: 70.52
hardware: V100
backend: PyTorch
batch size: 1
mode: FP32
resolution: (512,512)
value: 70.52
lr schd: 80000
memory (GB): 15.0
Name: danet_r101-d8_512x512_80k_ade20k
Results:
- Task: Semantic Segmentation
Dataset: ADE20K
Metrics:
mIoU: 43.64
mIoU(ms+flip): 45.19
Task: Semantic Segmentation
Config: configs/danet/danet_r101-d8_512x512_80k_ade20k.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/danet/danet_r101-d8_512x512_80k_ade20k/danet_r101-d8_512x512_80k_ade20k_20200615_015126-d0357c73.pth
- Config: configs/danet/danet_r50-d8_512x512_160k_ade20k.py
- Name: danet_r50-d8_512x512_160k_ade20k
In Collection: danet
Metadata:
backbone: R-50-D8
crop size: (512,512)
lr schd: 160000
Name: danet_r50-d8_512x512_160k_ade20k
Results:
- Task: Semantic Segmentation
Dataset: ADE20K
Metrics:
mIoU: 42.45
mIoU(ms+flip): 43.25
Task: Semantic Segmentation
Config: configs/danet/danet_r50-d8_512x512_160k_ade20k.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/danet/danet_r50-d8_512x512_160k_ade20k/danet_r50-d8_512x512_160k_ade20k_20200616_082340-9cb35dcd.pth
- Config: configs/danet/danet_r101-d8_512x512_160k_ade20k.py
- Name: danet_r101-d8_512x512_160k_ade20k
In Collection: danet
Metadata:
backbone: R-101-D8
crop size: (512,512)
lr schd: 160000
Name: danet_r101-d8_512x512_160k_ade20k
Results:
- Task: Semantic Segmentation
Dataset: ADE20K
Metrics:
mIoU: 44.17
mIoU(ms+flip): 45.02
Task: Semantic Segmentation
Config: configs/danet/danet_r101-d8_512x512_160k_ade20k.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/danet/danet_r101-d8_512x512_160k_ade20k/danet_r101-d8_512x512_160k_ade20k_20200616_082348-23bf12f9.pth
- Config: configs/danet/danet_r50-d8_512x512_20k_voc12aug.py
- Name: danet_r50-d8_512x512_20k_voc12aug
In Collection: danet
Metadata:
backbone: R-50-D8
crop size: (512,512)
lr schd: 20000
inference time (ms/im):
- backend: PyTorch
batch size: 1
- value: 47.76
hardware: V100
backend: PyTorch
batch size: 1
mode: FP32
resolution: (512,512)
value: 47.76
lr schd: 20000
memory (GB): 6.5
Name: danet_r50-d8_512x512_20k_voc12aug
Results:
- Task: Semantic Segmentation
Dataset: Pascal VOC 2012 + Aug
Metrics:
mIoU: 74.45
mIoU(ms+flip): 75.69
Task: Semantic Segmentation
Config: configs/danet/danet_r50-d8_512x512_20k_voc12aug.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/danet/danet_r50-d8_512x512_20k_voc12aug/danet_r50-d8_512x512_20k_voc12aug_20200618_070026-9e9e3ab3.pth
- Config: configs/danet/danet_r101-d8_512x512_20k_voc12aug.py
- Name: danet_r101-d8_512x512_20k_voc12aug
In Collection: danet
Metadata:
backbone: R-101-D8
crop size: (512,512)
lr schd: 20000
inference time (ms/im):
- backend: PyTorch
batch size: 1
- value: 72.67
hardware: V100
backend: PyTorch
batch size: 1
mode: FP32
resolution: (512,512)
value: 72.67
lr schd: 20000
memory (GB): 9.9
Name: danet_r101-d8_512x512_20k_voc12aug
Results:
- Task: Semantic Segmentation
Dataset: Pascal VOC 2012 + Aug
Metrics:
mIoU: 76.02
mIoU(ms+flip): 77.23
Task: Semantic Segmentation
Config: configs/danet/danet_r101-d8_512x512_20k_voc12aug.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/danet/danet_r101-d8_512x512_20k_voc12aug/danet_r101-d8_512x512_20k_voc12aug_20200618_070026-d48d23b2.pth
- Config: configs/danet/danet_r50-d8_512x512_40k_voc12aug.py
- Name: danet_r50-d8_512x512_40k_voc12aug
In Collection: danet
Metadata:
backbone: R-50-D8
crop size: (512,512)
lr schd: 40000
Name: danet_r50-d8_512x512_40k_voc12aug
Results:
- Task: Semantic Segmentation
Dataset: Pascal VOC 2012 + Aug
Metrics:
mIoU: 76.37
mIoU(ms+flip): 77.29
Task: Semantic Segmentation
Config: configs/danet/danet_r50-d8_512x512_40k_voc12aug.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/danet/danet_r50-d8_512x512_40k_voc12aug/danet_r50-d8_512x512_40k_voc12aug_20200613_235526-426e3a64.pth
- Config: configs/danet/danet_r101-d8_512x512_40k_voc12aug.py
- Name: danet_r101-d8_512x512_40k_voc12aug
In Collection: danet
Metadata:
backbone: R-101-D8
crop size: (512,512)
lr schd: 40000
Name: danet_r101-d8_512x512_40k_voc12aug
Results:
- Task: Semantic Segmentation
Dataset: Pascal VOC 2012 + Aug
Metrics:
mIoU: 76.51
mIoU(ms+flip): 77.32
Task: Semantic Segmentation
Config: configs/danet/danet_r101-d8_512x512_40k_voc12aug.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/danet/danet_r101-d8_512x512_40k_voc12aug/danet_r101-d8_512x512_40k_voc12aug_20200613_223031-788e232a.pth

9
configs/deeplabv3/README.md
View File

@ -4,6 +4,13 @@
<!-- [ALGORITHM] -->
<a href="https://github.com/tensorflow/models/tree/master/research/deeplab">Official Repo</a>
<a href="https://github.com/open-mmlab/mmsegmentation/blob/v0.17.0/mmseg/models/decode_heads/aspp_head.py#L54">Code Snippet</a>
<details>
<summary align="right"><a href="https://arxiv.org/abs/1706.05587">DeepLabV3 (ArXiv'2017)</a></summary>
```latext
@article{chen2017rethinking,
title={Rethinking atrous convolution for semantic image segmentation},
@ -13,6 +20,8 @@
}
```
</details>
## Results and models
:::{note}

415
configs/deeplabv3/deeplabv3.yml
View File

@ -1,5 +1,6 @@
Collections:
- Metadata:
- Name: deeplabv3
Metadata:
Training Data:
- Cityscapes
- ADE20K
@ -8,719 +9,727 @@ Collections:
- Pascal Context 59
- COCO-Stuff 10k
- COCO-Stuff 164k
Name: deeplabv3
Paper:
URL: https://arxiv.org/abs/1706.05587
Title: Rethinking atrous convolution for semantic image segmentation
README: configs/deeplabv3/README.md
Code:
URL: https://github.com/open-mmlab/mmsegmentation/blob/v0.17.0/mmseg/models/decode_heads/aspp_head.py#L54
Version: v0.17.0
Converted From:
Code: https://github.com/tensorflow/models/tree/master/research/deeplab
Models:
- Config: configs/deeplabv3/deeplabv3_r50-d8_512x1024_40k_cityscapes.py
- Name: deeplabv3_r50-d8_512x1024_40k_cityscapes
In Collection: deeplabv3
Metadata:
backbone: R-50-D8
crop size: (512,1024)
lr schd: 40000
inference time (ms/im):
- backend: PyTorch
batch size: 1
- value: 389.11
hardware: V100
backend: PyTorch
batch size: 1
mode: FP32
resolution: (512,1024)
value: 389.11
lr schd: 40000
memory (GB): 6.1
Name: deeplabv3_r50-d8_512x1024_40k_cityscapes
Results:
- Task: Semantic Segmentation
Dataset: Cityscapes
Metrics:
mIoU: 79.09
mIoU(ms+flip): 80.45
Task: Semantic Segmentation
Config: configs/deeplabv3/deeplabv3_r50-d8_512x1024_40k_cityscapes.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/deeplabv3/deeplabv3_r50-d8_512x1024_40k_cityscapes/deeplabv3_r50-d8_512x1024_40k_cityscapes_20200605_022449-acadc2f8.pth
- Config: configs/deeplabv3/deeplabv3_r101-d8_512x1024_40k_cityscapes.py
- Name: deeplabv3_r101-d8_512x1024_40k_cityscapes
In Collection: deeplabv3
Metadata:
backbone: R-101-D8
crop size: (512,1024)
lr schd: 40000
inference time (ms/im):
- backend: PyTorch
batch size: 1
- value: 520.83
hardware: V100
backend: PyTorch
batch size: 1
mode: FP32
resolution: (512,1024)
value: 520.83
lr schd: 40000
memory (GB): 9.6
Name: deeplabv3_r101-d8_512x1024_40k_cityscapes
Results:
- Task: Semantic Segmentation
Dataset: Cityscapes
Metrics:
mIoU: 77.12
mIoU(ms+flip): 79.61
Task: Semantic Segmentation
Config: configs/deeplabv3/deeplabv3_r101-d8_512x1024_40k_cityscapes.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/deeplabv3/deeplabv3_r101-d8_512x1024_40k_cityscapes/deeplabv3_r101-d8_512x1024_40k_cityscapes_20200605_012241-7fd3f799.pth
- Config: configs/deeplabv3/deeplabv3_r50-d8_769x769_40k_cityscapes.py
- Name: deeplabv3_r50-d8_769x769_40k_cityscapes
In Collection: deeplabv3
Metadata:
backbone: R-50-D8
crop size: (769,769)
lr schd: 40000
inference time (ms/im):
- backend: PyTorch
batch size: 1
- value: 900.9
hardware: V100
backend: PyTorch
batch size: 1
mode: FP32
resolution: (769,769)
value: 900.9
lr schd: 40000
memory (GB): 6.9
Name: deeplabv3_r50-d8_769x769_40k_cityscapes
Results:
- Task: Semantic Segmentation
Dataset: Cityscapes
Metrics:
mIoU: 78.58
mIoU(ms+flip): 79.89
Task: Semantic Segmentation
Config: configs/deeplabv3/deeplabv3_r50-d8_769x769_40k_cityscapes.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/deeplabv3/deeplabv3_r50-d8_769x769_40k_cityscapes/deeplabv3_r50-d8_769x769_40k_cityscapes_20200606_113723-7eda553c.pth
- Config: configs/deeplabv3/deeplabv3_r101-d8_769x769_40k_cityscapes.py
- Name: deeplabv3_r101-d8_769x769_40k_cityscapes
In Collection: deeplabv3
Metadata:
backbone: R-101-D8
crop size: (769,769)
lr schd: 40000
inference time (ms/im):
- backend: PyTorch
batch size: 1
- value: 1204.82
hardware: V100
backend: PyTorch
batch size: 1
mode: FP32
resolution: (769,769)
value: 1204.82
lr schd: 40000
memory (GB): 10.9
Name: deeplabv3_r101-d8_769x769_40k_cityscapes
Results:
- Task: Semantic Segmentation
Dataset: Cityscapes
Metrics:
mIoU: 79.27
mIoU(ms+flip): 80.11
Task: Semantic Segmentation
Config: configs/deeplabv3/deeplabv3_r101-d8_769x769_40k_cityscapes.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/deeplabv3/deeplabv3_r101-d8_769x769_40k_cityscapes/deeplabv3_r101-d8_769x769_40k_cityscapes_20200606_113809-c64f889f.pth
- Config: configs/deeplabv3/deeplabv3_r18-d8_512x1024_80k_cityscapes.py
- Name: deeplabv3_r18-d8_512x1024_80k_cityscapes
In Collection: deeplabv3
Metadata:
backbone: R-18-D8
crop size: (512,1024)
lr schd: 80000
inference time (ms/im):
- backend: PyTorch
batch size: 1
- value: 72.57
hardware: V100
backend: PyTorch
batch size: 1
mode: FP32
resolution: (512,1024)
value: 72.57
lr schd: 80000
memory (GB): 1.7
Name: deeplabv3_r18-d8_512x1024_80k_cityscapes
Results:
- Task: Semantic Segmentation
Dataset: Cityscapes
Metrics:
mIoU: 76.7
mIoU(ms+flip): 78.27
Task: Semantic Segmentation
Config: configs/deeplabv3/deeplabv3_r18-d8_512x1024_80k_cityscapes.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/deeplabv3/deeplabv3_r18-d8_512x1024_80k_cityscapes/deeplabv3_r18-d8_512x1024_80k_cityscapes_20201225_021506-23dffbe2.pth
- Config: configs/deeplabv3/deeplabv3_r50-d8_512x1024_80k_cityscapes.py
- Name: deeplabv3_r50-d8_512x1024_80k_cityscapes
In Collection: deeplabv3
Metadata:
backbone: R-50-D8
crop size: (512,1024)
lr schd: 80000
Name: deeplabv3_r50-d8_512x1024_80k_cityscapes
Results:
- Task: Semantic Segmentation
Dataset: Cityscapes
Metrics:
mIoU: 79.32
mIoU(ms+flip): 80.57
Task: Semantic Segmentation
Config: configs/deeplabv3/deeplabv3_r50-d8_512x1024_80k_cityscapes.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/deeplabv3/deeplabv3_r50-d8_512x1024_80k_cityscapes/deeplabv3_r50-d8_512x1024_80k_cityscapes_20200606_113404-b92cfdd4.pth
- Config: configs/deeplabv3/deeplabv3_r101-d8_512x1024_80k_cityscapes.py
- Name: deeplabv3_r101-d8_512x1024_80k_cityscapes
In Collection: deeplabv3
Metadata:
backbone: R-101-D8
crop size: (512,1024)
lr schd: 80000
Name: deeplabv3_r101-d8_512x1024_80k_cityscapes
Results:
- Task: Semantic Segmentation
Dataset: Cityscapes
Metrics:
mIoU: 80.2
mIoU(ms+flip): 81.21
Task: Semantic Segmentation
Config: configs/deeplabv3/deeplabv3_r101-d8_512x1024_80k_cityscapes.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/deeplabv3/deeplabv3_r101-d8_512x1024_80k_cityscapes/deeplabv3_r101-d8_512x1024_80k_cityscapes_20200606_113503-9e428899.pth
- Config: configs/deeplabv3/deeplabv3_r18-d8_769x769_80k_cityscapes.py
- Name: deeplabv3_r18-d8_769x769_80k_cityscapes
In Collection: deeplabv3
Metadata:
backbone: R-18-D8
crop size: (769,769)
lr schd: 80000
inference time (ms/im):
- backend: PyTorch
batch size: 1
- value: 180.18
hardware: V100
backend: PyTorch
batch size: 1
mode: FP32
resolution: (769,769)
value: 180.18
lr schd: 80000
memory (GB): 1.9
Name: deeplabv3_r18-d8_769x769_80k_cityscapes
Results:
- Task: Semantic Segmentation
Dataset: Cityscapes
Metrics:
mIoU: 76.6
mIoU(ms+flip): 78.26
Task: Semantic Segmentation
Config: configs/deeplabv3/deeplabv3_r18-d8_769x769_80k_cityscapes.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/deeplabv3/deeplabv3_r18-d8_769x769_80k_cityscapes/deeplabv3_r18-d8_769x769_80k_cityscapes_20201225_021506-6452126a.pth
- Config: configs/deeplabv3/deeplabv3_r50-d8_769x769_80k_cityscapes.py
- Name: deeplabv3_r50-d8_769x769_80k_cityscapes
In Collection: deeplabv3
Metadata:
backbone: R-50-D8
crop size: (769,769)
lr schd: 80000
Name: deeplabv3_r50-d8_769x769_80k_cityscapes
Results:
- Task: Semantic Segmentation
Dataset: Cityscapes
Metrics:
mIoU: 79.89
mIoU(ms+flip): 81.06
Task: Semantic Segmentation
Config: configs/deeplabv3/deeplabv3_r50-d8_769x769_80k_cityscapes.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/deeplabv3/deeplabv3_r50-d8_769x769_80k_cityscapes/deeplabv3_r50-d8_769x769_80k_cityscapes_20200606_221338-788d6228.pth
- Config: configs/deeplabv3/deeplabv3_r101-d8_769x769_80k_cityscapes.py
- Name: deeplabv3_r101-d8_769x769_80k_cityscapes
In Collection: deeplabv3
Metadata:
backbone: R-101-D8
crop size: (769,769)
lr schd: 80000
Name: deeplabv3_r101-d8_769x769_80k_cityscapes
Results:
- Task: Semantic Segmentation
Dataset: Cityscapes
Metrics:
mIoU: 79.67
mIoU(ms+flip): 80.81
Task: Semantic Segmentation
Config: configs/deeplabv3/deeplabv3_r101-d8_769x769_80k_cityscapes.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/deeplabv3/deeplabv3_r101-d8_769x769_80k_cityscapes/deeplabv3_r101-d8_769x769_80k_cityscapes_20200607_013353-60e95418.pth
- Config: configs/deeplabv3/deeplabv3_r101-d16-mg124_512x1024_80k_cityscapes.py
- Name: deeplabv3_r101-d16-mg124_512x1024_80k_cityscapes
In Collection: deeplabv3
Metadata:
backbone: R-101-D16-MG124
crop size: (512,1024)
lr schd: 80000
Name: deeplabv3_r101-d16-mg124_512x1024_80k_cityscapes
Results:
- Task: Semantic Segmentation
Dataset: Cityscapes
Metrics:
mIoU: 78.36
mIoU(ms+flip): 79.84
Task: Semantic Segmentation
Config: configs/deeplabv3/deeplabv3_r101-d16-mg124_512x1024_80k_cityscapes.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/deeplabv3/deeplabv3_r101-d16-mg124_512x1024_80k_cityscapes/deeplabv3_r101-d16-mg124_512x1024_80k_cityscapes_20200908_005644-57bb8425.pth
- Config: configs/deeplabv3/deeplabv3_r18b-d8_512x1024_80k_cityscapes.py
- Name: deeplabv3_r18b-d8_512x1024_80k_cityscapes
In Collection: deeplabv3
Metadata:
backbone: R-18b-D8
crop size: (512,1024)
lr schd: 80000
inference time (ms/im):
- backend: PyTorch
batch size: 1
- value: 71.79
hardware: V100
backend: PyTorch
batch size: 1
mode: FP32
resolution: (512,1024)
value: 71.79
lr schd: 80000
memory (GB): 1.6
Name: deeplabv3_r18b-d8_512x1024_80k_cityscapes
Results:
- Task: Semantic Segmentation
Dataset: Cityscapes
Metrics:
mIoU: 76.26
mIoU(ms+flip): 77.88
Task: Semantic Segmentation
Config: configs/deeplabv3/deeplabv3_r18b-d8_512x1024_80k_cityscapes.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/deeplabv3/deeplabv3_r18b-d8_512x1024_80k_cityscapes/deeplabv3_r18b-d8_512x1024_80k_cityscapes_20201225_094144-46040cef.pth
- Config: configs/deeplabv3/deeplabv3_r50b-d8_512x1024_80k_cityscapes.py
- Name: deeplabv3_r50b-d8_512x1024_80k_cityscapes
In Collection: deeplabv3
Metadata:
backbone: R-50b-D8
crop size: (512,1024)
lr schd: 80000
inference time (ms/im):
- backend: PyTorch
batch size: 1
- value: 364.96
hardware: V100
backend: PyTorch
batch size: 1
mode: FP32
resolution: (512,1024)
value: 364.96
lr schd: 80000
memory (GB): 6.0
Name: deeplabv3_r50b-d8_512x1024_80k_cityscapes
Results:
- Task: Semantic Segmentation
Dataset: Cityscapes
Metrics:
mIoU: 79.63
mIoU(ms+flip): 80.98
Task: Semantic Segmentation
Config: configs/deeplabv3/deeplabv3_r50b-d8_512x1024_80k_cityscapes.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/deeplabv3/deeplabv3_r50b-d8_512x1024_80k_cityscapes/deeplabv3_r50b-d8_512x1024_80k_cityscapes_20201225_155148-ec368954.pth
- Config: configs/deeplabv3/deeplabv3_r101b-d8_512x1024_80k_cityscapes.py
- Name: deeplabv3_r101b-d8_512x1024_80k_cityscapes
In Collection: deeplabv3
Metadata:
backbone: R-101b-D8
crop size: (512,1024)
lr schd: 80000
inference time (ms/im):
- backend: PyTorch
batch size: 1
- value: 552.49
hardware: V100
backend: PyTorch
batch size: 1
mode: FP32
resolution: (512,1024)
value: 552.49
lr schd: 80000
memory (GB): 9.5
Name: deeplabv3_r101b-d8_512x1024_80k_cityscapes
Results:
- Task: Semantic Segmentation
Dataset: Cityscapes
Metrics:
mIoU: 80.01
mIoU(ms+flip): 81.21
Task: Semantic Segmentation
Config: configs/deeplabv3/deeplabv3_r101b-d8_512x1024_80k_cityscapes.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/deeplabv3/deeplabv3_r101b-d8_512x1024_80k_cityscapes/deeplabv3_r101b-d8_512x1024_80k_cityscapes_20201226_171821-8fd49503.pth
- Config: configs/deeplabv3/deeplabv3_r18b-d8_769x769_80k_cityscapes.py
- Name: deeplabv3_r18b-d8_769x769_80k_cityscapes
In Collection: deeplabv3
Metadata:
backbone: R-18b-D8
crop size: (769,769)
lr schd: 80000
inference time (ms/im):
- backend: PyTorch
batch size: 1
- value: 172.71
hardware: V100
backend: PyTorch
batch size: 1
mode: FP32
resolution: (769,769)
value: 172.71
lr schd: 80000
memory (GB): 1.8
Name: deeplabv3_r18b-d8_769x769_80k_cityscapes
Results:
- Task: Semantic Segmentation
Dataset: Cityscapes
Metrics:
mIoU: 76.63
mIoU(ms+flip): 77.51
Task: Semantic Segmentation
Config: configs/deeplabv3/deeplabv3_r18b-d8_769x769_80k_cityscapes.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/deeplabv3/deeplabv3_r18b-d8_769x769_80k_cityscapes/deeplabv3_r18b-d8_769x769_80k_cityscapes_20201225_094144-fdc985d9.pth
- Config: configs/deeplabv3/deeplabv3_r50b-d8_769x769_80k_cityscapes.py
- Name: deeplabv3_r50b-d8_769x769_80k_cityscapes
In Collection: deeplabv3
Metadata:
backbone: R-50b-D8
crop size: (769,769)
lr schd: 80000
inference time (ms/im):
- backend: PyTorch
batch size: 1
- value: 862.07
hardware: V100
backend: PyTorch
batch size: 1
mode: FP32
resolution: (769,769)
value: 862.07
lr schd: 80000
memory (GB): 6.8
Name: deeplabv3_r50b-d8_769x769_80k_cityscapes
Results:
- Task: Semantic Segmentation
Dataset: Cityscapes
Metrics:
mIoU: 78.8
mIoU(ms+flip): 80.27
Task: Semantic Segmentation
Config: configs/deeplabv3/deeplabv3_r50b-d8_769x769_80k_cityscapes.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/deeplabv3/deeplabv3_r50b-d8_769x769_80k_cityscapes/deeplabv3_r50b-d8_769x769_80k_cityscapes_20201225_155404-87fb0cf4.pth
- Config: configs/deeplabv3/deeplabv3_r101b-d8_769x769_80k_cityscapes.py
- Name: deeplabv3_r101b-d8_769x769_80k_cityscapes
In Collection: deeplabv3
Metadata:
backbone: R-101b-D8
crop size: (769,769)
lr schd: 80000
inference time (ms/im):
- backend: PyTorch
batch size: 1
- value: 1219.51
hardware: V100
backend: PyTorch
batch size: 1
mode: FP32
resolution: (769,769)
value: 1219.51
lr schd: 80000
memory (GB): 10.7
Name: deeplabv3_r101b-d8_769x769_80k_cityscapes
Results:
- Task: Semantic Segmentation
Dataset: Cityscapes
Metrics:
mIoU: 79.41
mIoU(ms+flip): 80.73
Task: Semantic Segmentation
Config: configs/deeplabv3/deeplabv3_r101b-d8_769x769_80k_cityscapes.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/deeplabv3/deeplabv3_r101b-d8_769x769_80k_cityscapes/deeplabv3_r101b-d8_769x769_80k_cityscapes_20201226_190843-9142ee57.pth
- Config: configs/deeplabv3/deeplabv3_r50-d8_512x512_80k_ade20k.py
- Name: deeplabv3_r50-d8_512x512_80k_ade20k
In Collection: deeplabv3
Metadata:
backbone: R-50-D8
crop size: (512,512)
lr schd: 80000
inference time (ms/im):
- backend: PyTorch
batch size: 1
- value: 67.75
hardware: V100
backend: PyTorch
batch size: 1
mode: FP32
resolution: (512,512)
value: 67.75
lr schd: 80000
memory (GB): 8.9
Name: deeplabv3_r50-d8_512x512_80k_ade20k
Results:
- Task: Semantic Segmentation
Dataset: ADE20K
Metrics:
mIoU: 42.42
mIoU(ms+flip): 43.28
Task: Semantic Segmentation
Config: configs/deeplabv3/deeplabv3_r50-d8_512x512_80k_ade20k.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/deeplabv3/deeplabv3_r50-d8_512x512_80k_ade20k/deeplabv3_r50-d8_512x512_80k_ade20k_20200614_185028-0bb3f844.pth
- Config: configs/deeplabv3/deeplabv3_r101-d8_512x512_80k_ade20k.py
- Name: deeplabv3_r101-d8_512x512_80k_ade20k
In Collection: deeplabv3
Metadata:
backbone: R-101-D8
crop size: (512,512)
lr schd: 80000
inference time (ms/im):
- backend: PyTorch
batch size: 1
- value: 98.62
hardware: V100
backend: PyTorch
batch size: 1
mode: FP32
resolution: (512,512)
value: 98.62
lr schd: 80000
memory (GB): 12.4
Name: deeplabv3_r101-d8_512x512_80k_ade20k
Results:
- Task: Semantic Segmentation
Dataset: ADE20K
Metrics:
mIoU: 44.08
mIoU(ms+flip): 45.19
Task: Semantic Segmentation
Config: configs/deeplabv3/deeplabv3_r101-d8_512x512_80k_ade20k.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/deeplabv3/deeplabv3_r101-d8_512x512_80k_ade20k/deeplabv3_r101-d8_512x512_80k_ade20k_20200615_021256-d89c7fa4.pth
- Config: configs/deeplabv3/deeplabv3_r50-d8_512x512_160k_ade20k.py
- Name: deeplabv3_r50-d8_512x512_160k_ade20k
In Collection: deeplabv3
Metadata:
backbone: R-50-D8
crop size: (512,512)
lr schd: 160000
Name: deeplabv3_r50-d8_512x512_160k_ade20k
Results:
- Task: Semantic Segmentation
Dataset: ADE20K
Metrics:
mIoU: 42.66
mIoU(ms+flip): 44.09
Task: Semantic Segmentation
Config: configs/deeplabv3/deeplabv3_r50-d8_512x512_160k_ade20k.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/deeplabv3/deeplabv3_r50-d8_512x512_160k_ade20k/deeplabv3_r50-d8_512x512_160k_ade20k_20200615_123227-5d0ee427.pth
- Config: configs/deeplabv3/deeplabv3_r101-d8_512x512_160k_ade20k.py
- Name: deeplabv3_r101-d8_512x512_160k_ade20k
In Collection: deeplabv3
Metadata:
backbone: R-101-D8
crop size: (512,512)
lr schd: 160000
Name: deeplabv3_r101-d8_512x512_160k_ade20k
Results:
- Task: Semantic Segmentation
Dataset: ADE20K
Metrics:
mIoU: 45.0
mIoU(ms+flip): 46.66
Task: Semantic Segmentation
Config: configs/deeplabv3/deeplabv3_r101-d8_512x512_160k_ade20k.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/deeplabv3/deeplabv3_r101-d8_512x512_160k_ade20k/deeplabv3_r101-d8_512x512_160k_ade20k_20200615_105816-b1f72b3b.pth
- Config: configs/deeplabv3/deeplabv3_r50-d8_512x512_20k_voc12aug.py
- Name: deeplabv3_r50-d8_512x512_20k_voc12aug
In Collection: deeplabv3
Metadata:
backbone: R-50-D8
crop size: (512,512)
lr schd: 20000
inference time (ms/im):
- backend: PyTorch
batch size: 1
- value: 72.05
hardware: V100
backend: PyTorch
batch size: 1
mode: FP32
resolution: (512,512)
value: 72.05
lr schd: 20000
memory (GB): 6.1
Name: deeplabv3_r50-d8_512x512_20k_voc12aug
Results:
- Task: Semantic Segmentation
Dataset: Pascal VOC 2012 + Aug
Metrics:
mIoU: 76.17
mIoU(ms+flip): 77.42
Task: Semantic Segmentation
Config: configs/deeplabv3/deeplabv3_r50-d8_512x512_20k_voc12aug.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/deeplabv3/deeplabv3_r50-d8_512x512_20k_voc12aug/deeplabv3_r50-d8_512x512_20k_voc12aug_20200617_010906-596905ef.pth
- Config: configs/deeplabv3/deeplabv3_r101-d8_512x512_20k_voc12aug.py
- Name: deeplabv3_r101-d8_512x512_20k_voc12aug
In Collection: deeplabv3
Metadata:
backbone: R-101-D8
crop size: (512,512)
lr schd: 20000
inference time (ms/im):
- backend: PyTorch
batch size: 1
- value: 101.94
hardware: V100
backend: PyTorch
batch size: 1
mode: FP32
resolution: (512,512)
value: 101.94
lr schd: 20000
memory (GB): 9.6
Name: deeplabv3_r101-d8_512x512_20k_voc12aug
Results:
- Task: Semantic Segmentation
Dataset: Pascal VOC 2012 + Aug
Metrics:
mIoU: 78.7
mIoU(ms+flip): 79.95
Task: Semantic Segmentation
Config: configs/deeplabv3/deeplabv3_r101-d8_512x512_20k_voc12aug.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/deeplabv3/deeplabv3_r101-d8_512x512_20k_voc12aug/deeplabv3_r101-d8_512x512_20k_voc12aug_20200617_010932-8d13832f.pth
- Config: configs/deeplabv3/deeplabv3_r50-d8_512x512_40k_voc12aug.py
- Name: deeplabv3_r50-d8_512x512_40k_voc12aug
In Collection: deeplabv3
Metadata:
backbone: R-50-D8
crop size: (512,512)
lr schd: 40000
Name: deeplabv3_r50-d8_512x512_40k_voc12aug
Results:
- Task: Semantic Segmentation
Dataset: Pascal VOC 2012 + Aug
Metrics:
mIoU: 77.68
mIoU(ms+flip): 78.78
Task: Semantic Segmentation
Config: configs/deeplabv3/deeplabv3_r50-d8_512x512_40k_voc12aug.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/deeplabv3/deeplabv3_r50-d8_512x512_40k_voc12aug/deeplabv3_r50-d8_512x512_40k_voc12aug_20200613_161546-2ae96e7e.pth
- Config: configs/deeplabv3/deeplabv3_r101-d8_512x512_40k_voc12aug.py
- Name: deeplabv3_r101-d8_512x512_40k_voc12aug
In Collection: deeplabv3
Metadata:
backbone: R-101-D8
crop size: (512,512)
lr schd: 40000
Name: deeplabv3_r101-d8_512x512_40k_voc12aug
Results:
- Task: Semantic Segmentation
Dataset: Pascal VOC 2012 + Aug
Metrics:
mIoU: 77.92
mIoU(ms+flip): 79.18
Task: Semantic Segmentation
Config: configs/deeplabv3/deeplabv3_r101-d8_512x512_40k_voc12aug.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/deeplabv3/deeplabv3_r101-d8_512x512_40k_voc12aug/deeplabv3_r101-d8_512x512_40k_voc12aug_20200613_161432-0017d784.pth
- Config: configs/deeplabv3/deeplabv3_r101-d8_480x480_40k_pascal_context.py
- Name: deeplabv3_r101-d8_480x480_40k_pascal_context
In Collection: deeplabv3
Metadata:
backbone: R-101-D8
crop size: (480,480)
lr schd: 40000
inference time (ms/im):
- backend: PyTorch
batch size: 1
- value: 141.04
hardware: V100
backend: PyTorch
batch size: 1
mode: FP32
resolution: (480,480)
value: 141.04
lr schd: 40000
memory (GB): 9.2
Name: deeplabv3_r101-d8_480x480_40k_pascal_context
Results:
- Task: Semantic Segmentation
Dataset: Pascal Context
Metrics:
mIoU: 46.55
mIoU(ms+flip): 47.81
Task: Semantic Segmentation
Config: configs/deeplabv3/deeplabv3_r101-d8_480x480_40k_pascal_context.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/deeplabv3/deeplabv3_r101-d8_480x480_40k_pascal_context/deeplabv3_r101-d8_480x480_40k_pascal_context_20200911_204118-1aa27336.pth
- Config: configs/deeplabv3/deeplabv3_r101-d8_480x480_80k_pascal_context.py
- Name: deeplabv3_r101-d8_480x480_80k_pascal_context
In Collection: deeplabv3
Metadata:
backbone: R-101-D8
crop size: (480,480)
lr schd: 80000
Name: deeplabv3_r101-d8_480x480_80k_pascal_context
Results:
- Task: Semantic Segmentation
Dataset: Pascal Context
Metrics:
mIoU: 46.42
mIoU(ms+flip): 47.53
Task: Semantic Segmentation
Config: configs/deeplabv3/deeplabv3_r101-d8_480x480_80k_pascal_context.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/deeplabv3/deeplabv3_r101-d8_480x480_80k_pascal_context/deeplabv3_r101-d8_480x480_80k_pascal_context_20200911_170155-2a21fff3.pth
- Config: configs/deeplabv3/deeplabv3_r101-d8_480x480_40k_pascal_context_59.py
- Name: deeplabv3_r101-d8_480x480_40k_pascal_context_59
In Collection: deeplabv3
Metadata:
backbone: R-101-D8
crop size: (480,480)
lr schd: 40000
Name: deeplabv3_r101-d8_480x480_40k_pascal_context_59
Results:
- Task: Semantic Segmentation
Dataset: Pascal Context 59
Metrics:
mIoU: 52.61
mIoU(ms+flip): 54.28
Task: Semantic Segmentation
Config: configs/deeplabv3/deeplabv3_r101-d8_480x480_40k_pascal_context_59.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/deeplabv3/deeplabv3_r101-d8_480x480_40k_pascal_context_59/deeplabv3_r101-d8_480x480_40k_pascal_context_59_20210416_110332-cb08ea46.pth
- Config: configs/deeplabv3/deeplabv3_r101-d8_480x480_80k_pascal_context_59.py
- Name: deeplabv3_r101-d8_480x480_80k_pascal_context_59
In Collection: deeplabv3
Metadata:
backbone: R-101-D8
crop size: (480,480)
lr schd: 80000
Name: deeplabv3_r101-d8_480x480_80k_pascal_context_59
Results:
- Task: Semantic Segmentation
Dataset: Pascal Context 59
Metrics:
mIoU: 52.46
mIoU(ms+flip): 54.09
Task: Semantic Segmentation
Config: configs/deeplabv3/deeplabv3_r101-d8_480x480_80k_pascal_context_59.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/deeplabv3/deeplabv3_r101-d8_480x480_80k_pascal_context_59/deeplabv3_r101-d8_480x480_80k_pascal_context_59_20210416_113002-26303993.pth
- Config: configs/deeplabv3/deeplabv3_r50-d8_512x512_4x4_20k_coco-stuff10k.py
- Name: deeplabv3_r50-d8_512x512_4x4_20k_coco-stuff10k
In Collection: deeplabv3
Metadata:
backbone: R-50-D8
crop size: (512,512)
lr schd: 20000
inference time (ms/im):
- backend: PyTorch
batch size: 1
- value: 92.59
hardware: V100
backend: PyTorch
batch size: 1
mode: FP32
resolution: (512,512)
value: 92.59
lr schd: 20000
memory (GB): 9.6
Name: deeplabv3_r50-d8_512x512_4x4_20k_coco-stuff10k
Results:
- Task: Semantic Segmentation
Dataset: COCO-Stuff 10k
Metrics:
mIoU: 34.66
mIoU(ms+flip): 36.08
Task: Semantic Segmentation
Config: configs/deeplabv3/deeplabv3_r50-d8_512x512_4x4_20k_coco-stuff10k.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/deeplabv3/deeplabv3_r50-d8_512x512_4x4_20k_coco-stuff10k/deeplabv3_r50-d8_512x512_4x4_20k_coco-stuff10k_20210821_043025-b35f789d.pth
- Config: configs/deeplabv3/deeplabv3_r101-d8_512x512_4x4_20k_coco-stuff10k.py
- Name: deeplabv3_r101-d8_512x512_4x4_20k_coco-stuff10k
In Collection: deeplabv3
Metadata:
backbone: R-101-D8
crop size: (512,512)
lr schd: 20000
inference time (ms/im):
- backend: PyTorch
batch size: 1
- value: 114.94
hardware: V100
backend: PyTorch
batch size: 1
mode: FP32
resolution: (512,512)
value: 114.94
lr schd: 20000
memory (GB): 13.2
Name: deeplabv3_r101-d8_512x512_4x4_20k_coco-stuff10k
Results:
- Task: Semantic Segmentation
Dataset: COCO-Stuff 10k
Metrics:
mIoU: 37.3
mIoU(ms+flip): 38.42
Task: Semantic Segmentation
Config: configs/deeplabv3/deeplabv3_r101-d8_512x512_4x4_20k_coco-stuff10k.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/deeplabv3/deeplabv3_r101-d8_512x512_4x4_20k_coco-stuff10k/deeplabv3_r101-d8_512x512_4x4_20k_coco-stuff10k_20210821_043025-c49752cb.pth
- Config: configs/deeplabv3/deeplabv3_r50-d8_512x512_4x4_40k_coco-stuff10k.py
- Name: deeplabv3_r50-d8_512x512_4x4_40k_coco-stuff10k
In Collection: deeplabv3
Metadata:
backbone: R-50-D8
crop size: (512,512)
lr schd: 40000
Name: deeplabv3_r50-d8_512x512_4x4_40k_coco-stuff10k
Results:
- Task: Semantic Segmentation
Dataset: COCO-Stuff 10k
Metrics:
mIoU: 35.73
mIoU(ms+flip): 37.09
Task: Semantic Segmentation
Config: configs/deeplabv3/deeplabv3_r50-d8_512x512_4x4_40k_coco-stuff10k.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/deeplabv3/deeplabv3_r50-d8_512x512_4x4_40k_coco-stuff10k/deeplabv3_r50-d8_512x512_4x4_40k_coco-stuff10k_20210821_043305-dc76f3ff.pth
- Config: configs/deeplabv3/deeplabv3_r101-d8_512x512_4x4_40k_coco-stuff10k.py
- Name: deeplabv3_r101-d8_512x512_4x4_40k_coco-stuff10k
In Collection: deeplabv3
Metadata:
backbone: R-101-D8
crop size: (512,512)
lr schd: 40000
Name: deeplabv3_r101-d8_512x512_4x4_40k_coco-stuff10k
Results:
- Task: Semantic Segmentation
Dataset: COCO-Stuff 10k
Metrics:
mIoU: 37.81
mIoU(ms+flip): 38.8
Task: Semantic Segmentation
Config: configs/deeplabv3/deeplabv3_r101-d8_512x512_4x4_40k_coco-stuff10k.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/deeplabv3/deeplabv3_r101-d8_512x512_4x4_40k_coco-stuff10k/deeplabv3_r101-d8_512x512_4x4_40k_coco-stuff10k_20210821_043305-636cb433.pth
- Config: configs/deeplabv3/deeplabv3_r50-d8_512x512_4x4_80k_coco-stuff164k.py
- Name: deeplabv3_r50-d8_512x512_4x4_80k_coco-stuff164k
In Collection: deeplabv3
Metadata:
backbone: R-50-D8
crop size: (512,512)
lr schd: 80000
inference time (ms/im):
- backend: PyTorch
batch size: 1
- value: 92.59
hardware: V100
backend: PyTorch
batch size: 1
mode: FP32
resolution: (512,512)
value: 92.59
lr schd: 80000
memory (GB): 9.6
Name: deeplabv3_r50-d8_512x512_4x4_80k_coco-stuff164k
Results:
- Task: Semantic Segmentation
Dataset: COCO-Stuff 164k
Metrics:
mIoU: 39.38
mIoU(ms+flip): 40.03
Task: Semantic Segmentation
Config: configs/deeplabv3/deeplabv3_r50-d8_512x512_4x4_80k_coco-stuff164k.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/deeplabv3/deeplabv3_r50-d8_512x512_4x4_80k_coco-stuff164k/deeplabv3_r50-d8_512x512_4x4_80k_coco-stuff164k_20210709_163016-88675c24.pth
- Config: configs/deeplabv3/deeplabv3_r101-d8_512x512_4x4_80k_coco-stuff164k.py
- Name: deeplabv3_r101-d8_512x512_4x4_80k_coco-stuff164k
In Collection: deeplabv3
Metadata:
backbone: R-101-D8
crop size: (512,512)
lr schd: 80000
inference time (ms/im):
- backend: PyTorch
batch size: 1
- value: 114.94
hardware: V100
backend: PyTorch
batch size: 1
mode: FP32
resolution: (512,512)
value: 114.94
lr schd: 80000
memory (GB): 13.2
Name: deeplabv3_r101-d8_512x512_4x4_80k_coco-stuff164k
Results:
- Task: Semantic Segmentation
Dataset: COCO-Stuff 164k
Metrics:
mIoU: 40.87
mIoU(ms+flip): 41.5
Task: Semantic Segmentation
Config: configs/deeplabv3/deeplabv3_r101-d8_512x512_4x4_80k_coco-stuff164k.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/deeplabv3/deeplabv3_r101-d8_512x512_4x4_80k_coco-stuff164k/deeplabv3_r101-d8_512x512_4x4_80k_coco-stuff164k_20210709_201252-13600dc2.pth
- Config: configs/deeplabv3/deeplabv3_r50-d8_512x512_4x4_160k_coco-stuff164k.py
- Name: deeplabv3_r50-d8_512x512_4x4_160k_coco-stuff164k
In Collection: deeplabv3
Metadata:
backbone: R-50-D8
crop size: (512,512)
lr schd: 160000
Name: deeplabv3_r50-d8_512x512_4x4_160k_coco-stuff164k
Results:
- Task: Semantic Segmentation
Dataset: COCO-Stuff 164k
Metrics:
mIoU: 41.09
mIoU(ms+flip): 41.69
Task: Semantic Segmentation
Config: configs/deeplabv3/deeplabv3_r50-d8_512x512_4x4_160k_coco-stuff164k.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/deeplabv3/deeplabv3_r50-d8_512x512_4x4_160k_coco-stuff164k/deeplabv3_r50-d8_512x512_4x4_160k_coco-stuff164k_20210709_163016-49f2812b.pth
- Config: configs/deeplabv3/deeplabv3_r101-d8_512x512_4x4_160k_coco-stuff164k.py
- Name: deeplabv3_r101-d8_512x512_4x4_160k_coco-stuff164k
In Collection: deeplabv3
Metadata:
backbone: R-101-D8
crop size: (512,512)
lr schd: 160000
Name: deeplabv3_r101-d8_512x512_4x4_160k_coco-stuff164k
Results:
- Task: Semantic Segmentation
Dataset: COCO-Stuff 164k
Metrics:
mIoU: 41.82
mIoU(ms+flip): 42.49
Task: Semantic Segmentation
Config: configs/deeplabv3/deeplabv3_r101-d8_512x512_4x4_160k_coco-stuff164k.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/deeplabv3/deeplabv3_r101-d8_512x512_4x4_160k_coco-stuff164k/deeplabv3_r101-d8_512x512_4x4_160k_coco-stuff164k_20210709_155402-f035acfd.pth
- Config: configs/deeplabv3/deeplabv3_r50-d8_512x512_4x4_320k_coco-stuff164k.py
- Name: deeplabv3_r50-d8_512x512_4x4_320k_coco-stuff164k
In Collection: deeplabv3
Metadata:
backbone: R-50-D8
crop size: (512,512)
lr schd: 320000
Name: deeplabv3_r50-d8_512x512_4x4_320k_coco-stuff164k
Results:
- Task: Semantic Segmentation
Dataset: COCO-Stuff 164k
Metrics:
mIoU: 41.37
mIoU(ms+flip): 42.22
Task: Semantic Segmentation
Config: configs/deeplabv3/deeplabv3_r50-d8_512x512_4x4_320k_coco-stuff164k.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/deeplabv3/deeplabv3_r50-d8_512x512_4x4_320k_coco-stuff164k/deeplabv3_r50-d8_512x512_4x4_320k_coco-stuff164k_20210709_155403-51b21115.pth
- Config: configs/deeplabv3/deeplabv3_r101-d8_512x512_4x4_320k_coco-stuff164k.py
- Name: deeplabv3_r101-d8_512x512_4x4_320k_coco-stuff164k
In Collection: deeplabv3
Metadata:
backbone: R-101-D8
crop size: (512,512)
lr schd: 320000
Name: deeplabv3_r101-d8_512x512_4x4_320k_coco-stuff164k
Results:
- Task: Semantic Segmentation
Dataset: COCO-Stuff 164k
Metrics:
mIoU: 42.61
mIoU(ms+flip): 43.42
Task: Semantic Segmentation
Config: configs/deeplabv3/deeplabv3_r101-d8_512x512_4x4_320k_coco-stuff164k.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/deeplabv3/deeplabv3_r101-d8_512x512_4x4_320k_coco-stuff164k/deeplabv3_r101-d8_512x512_4x4_320k_coco-stuff164k_20210709_155402-3cbca14d.pth

9
configs/deeplabv3plus/README.md
View File

@ -4,6 +4,13 @@
<!-- [ALGORITHM] -->
<a href="https://github.com/tensorflow/models/tree/master/research/deeplab">Official Repo</a>
<a href="https://github.com/open-mmlab/mmsegmentation/blob/v0.17.0/mmseg/models/decode_heads/sep_aspp_head.py#L30">Code Snippet</a>
<details>
<summary align="right"><a href="https://arxiv.org/abs/1802.02611">DeepLabV3+ (CVPR'2018)</a></summary>
```latex
@inproceedings{deeplabv3plus2018,
title={Encoder-Decoder with Atrous Separable Convolution for Semantic Image Segmentation},
@ -13,6 +20,8 @@
}
```
</details>
## Results and models
:::{note}

356
configs/deeplabv3plus/deeplabv3plus.yml
View File

@ -1,574 +1,580 @@
Collections:
- Metadata:
- Name: deeplabv3plus
Metadata:
Training Data:
- Cityscapes
- ADE20K
- ' Pascal VOC 2012 + Aug'
- ' Pascal Context'
- ' Pascal Context 59'
Name: deeplabv3plus
Paper:
URL: https://arxiv.org/abs/1802.02611
Title: Encoder-Decoder with Atrous Separable Convolution for Semantic Image Segmentation
README: configs/deeplabv3plus/README.md
Code:
URL: https://github.com/open-mmlab/mmsegmentation/blob/v0.17.0/mmseg/models/decode_heads/sep_aspp_head.py#L30
Version: v0.17.0
Converted From:
Code: https://github.com/tensorflow/models/tree/master/research/deeplab
Models:
- Config: configs/deeplabv3plus/deeplabv3plus_r50-d8_512x1024_40k_cityscapes.py
- Name: deeplabv3plus_r50-d8_512x1024_40k_cityscapes
In Collection: deeplabv3plus
Metadata:
backbone: R-50-D8
crop size: (512,1024)
lr schd: 40000
inference time (ms/im):
- backend: PyTorch
batch size: 1
- value: 253.81
hardware: V100
backend: PyTorch
batch size: 1
mode: FP32
resolution: (512,1024)
value: 253.81
lr schd: 40000
memory (GB): 7.5
Name: deeplabv3plus_r50-d8_512x1024_40k_cityscapes
Results:
- Task: Semantic Segmentation
Dataset: Cityscapes
Metrics:
mIoU: 79.61
mIoU(ms+flip): 81.01
Task: Semantic Segmentation
Config: configs/deeplabv3plus/deeplabv3plus_r50-d8_512x1024_40k_cityscapes.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/deeplabv3plus/deeplabv3plus_r50-d8_512x1024_40k_cityscapes/deeplabv3plus_r50-d8_512x1024_40k_cityscapes_20200605_094610-d222ffcd.pth
- Config: configs/deeplabv3plus/deeplabv3plus_r101-d8_512x1024_40k_cityscapes.py
- Name: deeplabv3plus_r101-d8_512x1024_40k_cityscapes
In Collection: deeplabv3plus
Metadata:
backbone: R-101-D8
crop size: (512,1024)
lr schd: 40000
inference time (ms/im):
- backend: PyTorch
batch size: 1
- value: 384.62
hardware: V100
backend: PyTorch
batch size: 1
mode: FP32
resolution: (512,1024)
value: 384.62
lr schd: 40000
memory (GB): 11.0
Name: deeplabv3plus_r101-d8_512x1024_40k_cityscapes
Results:
- Task: Semantic Segmentation
Dataset: Cityscapes
Metrics:
mIoU: 80.21
mIoU(ms+flip): 81.82
Task: Semantic Segmentation
Config: configs/deeplabv3plus/deeplabv3plus_r101-d8_512x1024_40k_cityscapes.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/deeplabv3plus/deeplabv3plus_r101-d8_512x1024_40k_cityscapes/deeplabv3plus_r101-d8_512x1024_40k_cityscapes_20200605_094614-3769eecf.pth
- Config: configs/deeplabv3plus/deeplabv3plus_r50-d8_769x769_40k_cityscapes.py
- Name: deeplabv3plus_r50-d8_769x769_40k_cityscapes
In Collection: deeplabv3plus
Metadata:
backbone: R-50-D8
crop size: (769,769)
lr schd: 40000
inference time (ms/im):
- backend: PyTorch
batch size: 1
- value: 581.4
hardware: V100
backend: PyTorch
batch size: 1
mode: FP32
resolution: (769,769)
value: 581.4
lr schd: 40000
memory (GB): 8.5
Name: deeplabv3plus_r50-d8_769x769_40k_cityscapes
Results:
- Task: Semantic Segmentation
Dataset: Cityscapes
Metrics:
mIoU: 78.97
mIoU(ms+flip): 80.46
Task: Semantic Segmentation
Config: configs/deeplabv3plus/deeplabv3plus_r50-d8_769x769_40k_cityscapes.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/deeplabv3plus/deeplabv3plus_r50-d8_769x769_40k_cityscapes/deeplabv3plus_r50-d8_769x769_40k_cityscapes_20200606_114143-1dcb0e3c.pth
- Config: configs/deeplabv3plus/deeplabv3plus_r101-d8_769x769_40k_cityscapes.py
- Name: deeplabv3plus_r101-d8_769x769_40k_cityscapes
In Collection: deeplabv3plus
Metadata:
backbone: R-101-D8
crop size: (769,769)
lr schd: 40000
inference time (ms/im):
- backend: PyTorch
batch size: 1
- value: 869.57
hardware: V100
backend: PyTorch
batch size: 1
mode: FP32
resolution: (769,769)
value: 869.57
lr schd: 40000
memory (GB): 12.5
Name: deeplabv3plus_r101-d8_769x769_40k_cityscapes
Results:
- Task: Semantic Segmentation
Dataset: Cityscapes
Metrics:
mIoU: 79.46
mIoU(ms+flip): 80.5
Task: Semantic Segmentation
Config: configs/deeplabv3plus/deeplabv3plus_r101-d8_769x769_40k_cityscapes.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/deeplabv3plus/deeplabv3plus_r101-d8_769x769_40k_cityscapes/deeplabv3plus_r101-d8_769x769_40k_cityscapes_20200606_114304-ff414b9e.pth
- Config: configs/deeplabv3plus/deeplabv3plus_r18-d8_512x1024_80k_cityscapes.py
- Name: deeplabv3plus_r18-d8_512x1024_80k_cityscapes
In Collection: deeplabv3plus
Metadata:
backbone: R-18-D8
crop size: (512,1024)
lr schd: 80000
inference time (ms/im):
- backend: PyTorch
batch size: 1
- value: 70.08
hardware: V100
backend: PyTorch
batch size: 1
mode: FP32
resolution: (512,1024)
value: 70.08
lr schd: 80000
memory (GB): 2.2
Name: deeplabv3plus_r18-d8_512x1024_80k_cityscapes
Results:
- Task: Semantic Segmentation
Dataset: Cityscapes
Metrics:
mIoU: 76.89
mIoU(ms+flip): 78.76
Task: Semantic Segmentation
Config: configs/deeplabv3plus/deeplabv3plus_r18-d8_512x1024_80k_cityscapes.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/deeplabv3plus/deeplabv3plus_r18-d8_512x1024_80k_cityscapes/deeplabv3plus_r18-d8_512x1024_80k_cityscapes_20201226_080942-cff257fe.pth
- Config: configs/deeplabv3plus/deeplabv3plus_r50-d8_512x1024_80k_cityscapes.py
- Name: deeplabv3plus_r50-d8_512x1024_80k_cityscapes
In Collection: deeplabv3plus
Metadata:
backbone: R-50-D8
crop size: (512,1024)
lr schd: 80000
Name: deeplabv3plus_r50-d8_512x1024_80k_cityscapes
Results:
- Task: Semantic Segmentation
Dataset: Cityscapes
Metrics:
mIoU: 80.09
mIoU(ms+flip): 81.13
Task: Semantic Segmentation
Config: configs/deeplabv3plus/deeplabv3plus_r50-d8_512x1024_80k_cityscapes.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/deeplabv3plus/deeplabv3plus_r50-d8_512x1024_80k_cityscapes/deeplabv3plus_r50-d8_512x1024_80k_cityscapes_20200606_114049-f9fb496d.pth
- Config: configs/deeplabv3plus/deeplabv3plus_r101-d8_512x1024_80k_cityscapes.py
- Name: deeplabv3plus_r101-d8_512x1024_80k_cityscapes
In Collection: deeplabv3plus
Metadata:
backbone: R-101-D8
crop size: (512,1024)
lr schd: 80000
Name: deeplabv3plus_r101-d8_512x1024_80k_cityscapes
Results:
- Task: Semantic Segmentation
Dataset: Cityscapes
Metrics:
mIoU: 80.97
mIoU(ms+flip): 82.03
Task: Semantic Segmentation
Config: configs/deeplabv3plus/deeplabv3plus_r101-d8_512x1024_80k_cityscapes.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/deeplabv3plus/deeplabv3plus_r101-d8_512x1024_80k_cityscapes/deeplabv3plus_r101-d8_512x1024_80k_cityscapes_20200606_114143-068fcfe9.pth
- Config: configs/deeplabv3plus/deeplabv3plus_r18-d8_769x769_80k_cityscapes.py
- Name: deeplabv3plus_r18-d8_769x769_80k_cityscapes
In Collection: deeplabv3plus
Metadata:
backbone: R-18-D8
crop size: (769,769)
lr schd: 80000
inference time (ms/im):
- backend: PyTorch
batch size: 1
- value: 174.22
hardware: V100
backend: PyTorch
batch size: 1
mode: FP32
resolution: (769,769)
value: 174.22
lr schd: 80000
memory (GB): 2.5
Name: deeplabv3plus_r18-d8_769x769_80k_cityscapes
Results:
- Task: Semantic Segmentation
Dataset: Cityscapes
Metrics:
mIoU: 76.26
mIoU(ms+flip): 77.91
Task: Semantic Segmentation
Config: configs/deeplabv3plus/deeplabv3plus_r18-d8_769x769_80k_cityscapes.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/deeplabv3plus/deeplabv3plus_r18-d8_769x769_80k_cityscapes/deeplabv3plus_r18-d8_769x769_80k_cityscapes_20201226_083346-f326e06a.pth
- Config: configs/deeplabv3plus/deeplabv3plus_r50-d8_769x769_80k_cityscapes.py
- Name: deeplabv3plus_r50-d8_769x769_80k_cityscapes
In Collection: deeplabv3plus
Metadata:
backbone: R-50-D8
crop size: (769,769)
lr schd: 80000
Name: deeplabv3plus_r50-d8_769x769_80k_cityscapes
Results:
- Task: Semantic Segmentation
Dataset: Cityscapes
Metrics:
mIoU: 79.83
mIoU(ms+flip): 81.48
Task: Semantic Segmentation
Config: configs/deeplabv3plus/deeplabv3plus_r50-d8_769x769_80k_cityscapes.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/deeplabv3plus/deeplabv3plus_r50-d8_769x769_80k_cityscapes/deeplabv3plus_r50-d8_769x769_80k_cityscapes_20200606_210233-0e9dfdc4.pth
- Config: configs/deeplabv3plus/deeplabv3plus_r101-d8_769x769_80k_cityscapes.py
- Name: deeplabv3plus_r101-d8_769x769_80k_cityscapes
In Collection: deeplabv3plus
Metadata:
backbone: R-101-D8
crop size: (769,769)
lr schd: 80000
Name: deeplabv3plus_r101-d8_769x769_80k_cityscapes
Results:
- Task: Semantic Segmentation
Dataset: Cityscapes
Metrics:
mIoU: 80.98
mIoU(ms+flip): 82.18
Task: Semantic Segmentation
Config: configs/deeplabv3plus/deeplabv3plus_r101-d8_769x769_80k_cityscapes.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/deeplabv3plus/deeplabv3plus_r101-d8_769x769_80k_cityscapes/deeplabv3plus_r101-d8_769x769_80k_cityscapes_20200607_000405-a7573d20.pth
- Config: configs/deeplabv3plus/deeplabv3plus_r101-d16-mg124_512x1024_40k_cityscapes.py
- Name: deeplabv3plus_r101-d16-mg124_512x1024_40k_cityscapes
In Collection: deeplabv3plus
Metadata:
backbone: R-101-D16-MG124
crop size: (512,1024)
lr schd: 40000
inference time (ms/im):
- backend: PyTorch
batch size: 1
- value: 133.69
hardware: V100
backend: PyTorch
batch size: 1
mode: FP32
resolution: (512,1024)
value: 133.69
lr schd: 40000
memory (GB): 5.8
Name: deeplabv3plus_r101-d16-mg124_512x1024_40k_cityscapes
Results:
- Task: Semantic Segmentation
Dataset: Cityscapes
Metrics:
mIoU: 79.09
mIoU(ms+flip): 80.36
Task: Semantic Segmentation
Config: configs/deeplabv3plus/deeplabv3plus_r101-d16-mg124_512x1024_40k_cityscapes.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/deeplabv3plus/deeplabv3plus_r101-d16-mg124_512x1024_40k_cityscapes/deeplabv3plus_r101-d16-mg124_512x1024_40k_cityscapes_20200908_005644-cf9ce186.pth
- Config: configs/deeplabv3plus/deeplabv3plus_r101-d16-mg124_512x1024_80k_cityscapes.py
- Name: deeplabv3plus_r101-d16-mg124_512x1024_80k_cityscapes
In Collection: deeplabv3plus
Metadata:
backbone: R-101-D16-MG124
crop size: (512,1024)
lr schd: 80000
memory (GB): 9.9
Name: deeplabv3plus_r101-d16-mg124_512x1024_80k_cityscapes
Results:
- Task: Semantic Segmentation
Dataset: Cityscapes
Metrics:
mIoU: 79.9
mIoU(ms+flip): 81.33
Task: Semantic Segmentation
Config: configs/deeplabv3plus/deeplabv3plus_r101-d16-mg124_512x1024_80k_cityscapes.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/deeplabv3plus/deeplabv3plus_r101-d16-mg124_512x1024_80k_cityscapes/deeplabv3plus_r101-d16-mg124_512x1024_80k_cityscapes_20200908_005644-ee6158e0.pth
- Config: configs/deeplabv3plus/deeplabv3plus_r18b-d8_512x1024_80k_cityscapes.py
- Name: deeplabv3plus_r18b-d8_512x1024_80k_cityscapes
In Collection: deeplabv3plus
Metadata:
backbone: R-18b-D8
crop size: (512,1024)
lr schd: 80000
inference time (ms/im):
- backend: PyTorch
batch size: 1
- value: 66.89
hardware: V100
backend: PyTorch
batch size: 1
mode: FP32
resolution: (512,1024)
value: 66.89
lr schd: 80000
memory (GB): 2.1
Name: deeplabv3plus_r18b-d8_512x1024_80k_cityscapes
Results:
- Task: Semantic Segmentation
Dataset: Cityscapes
Metrics:
mIoU: 75.87
mIoU(ms+flip): 77.52
Task: Semantic Segmentation
Config: configs/deeplabv3plus/deeplabv3plus_r18b-d8_512x1024_80k_cityscapes.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/deeplabv3plus/deeplabv3plus_r18b-d8_512x1024_80k_cityscapes/deeplabv3plus_r18b-d8_512x1024_80k_cityscapes_20201226_090828-e451abd9.pth
- Config: configs/deeplabv3plus/deeplabv3plus_r50b-d8_512x1024_80k_cityscapes.py
- Name: deeplabv3plus_r50b-d8_512x1024_80k_cityscapes
In Collection: deeplabv3plus
Metadata:
backbone: R-50b-D8
crop size: (512,1024)
lr schd: 80000
inference time (ms/im):
- backend: PyTorch
batch size: 1
- value: 253.81
hardware: V100
backend: PyTorch
batch size: 1
mode: FP32
resolution: (512,1024)
value: 253.81
lr schd: 80000
memory (GB): 7.4
Name: deeplabv3plus_r50b-d8_512x1024_80k_cityscapes
Results:
- Task: Semantic Segmentation
Dataset: Cityscapes
Metrics:
mIoU: 80.28
mIoU(ms+flip): 81.44
Task: Semantic Segmentation
Config: configs/deeplabv3plus/deeplabv3plus_r50b-d8_512x1024_80k_cityscapes.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/deeplabv3plus/deeplabv3plus_r50b-d8_512x1024_80k_cityscapes/deeplabv3plus_r50b-d8_512x1024_80k_cityscapes_20201225_213645-a97e4e43.pth
- Config: configs/deeplabv3plus/deeplabv3plus_r101b-d8_512x1024_80k_cityscapes.py
- Name: deeplabv3plus_r101b-d8_512x1024_80k_cityscapes
In Collection: deeplabv3plus
Metadata:
backbone: R-101b-D8
crop size: (512,1024)
lr schd: 80000
inference time (ms/im):
- backend: PyTorch
batch size: 1
- value: 384.62
hardware: V100
backend: PyTorch
batch size: 1
mode: FP32
resolution: (512,1024)
value: 384.62
lr schd: 80000
memory (GB): 10.9
Name: deeplabv3plus_r101b-d8_512x1024_80k_cityscapes
Results:
- Task: Semantic Segmentation
Dataset: Cityscapes
Metrics:
mIoU: 80.16
mIoU(ms+flip): 81.41
Task: Semantic Segmentation
Config: configs/deeplabv3plus/deeplabv3plus_r101b-d8_512x1024_80k_cityscapes.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/deeplabv3plus/deeplabv3plus_r101b-d8_512x1024_80k_cityscapes/deeplabv3plus_r101b-d8_512x1024_80k_cityscapes_20201226_190843-9c3c93a4.pth
- Config: configs/deeplabv3plus/deeplabv3plus_r18b-d8_769x769_80k_cityscapes.py
- Name: deeplabv3plus_r18b-d8_769x769_80k_cityscapes
In Collection: deeplabv3plus
Metadata:
backbone: R-18b-D8
crop size: (769,769)
lr schd: 80000
inference time (ms/im):
- backend: PyTorch
batch size: 1
- value: 167.79
hardware: V100
backend: PyTorch
batch size: 1
mode: FP32
resolution: (769,769)
value: 167.79
lr schd: 80000
memory (GB): 2.4
Name: deeplabv3plus_r18b-d8_769x769_80k_cityscapes
Results:
- Task: Semantic Segmentation
Dataset: Cityscapes
Metrics:
mIoU: 76.36
mIoU(ms+flip): 78.24
Task: Semantic Segmentation
Config: configs/deeplabv3plus/deeplabv3plus_r18b-d8_769x769_80k_cityscapes.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/deeplabv3plus/deeplabv3plus_r18b-d8_769x769_80k_cityscapes/deeplabv3plus_r18b-d8_769x769_80k_cityscapes_20201226_151312-2c868aff.pth
- Config: configs/deeplabv3plus/deeplabv3plus_r50b-d8_769x769_80k_cityscapes.py
- Name: deeplabv3plus_r50b-d8_769x769_80k_cityscapes
In Collection: deeplabv3plus
Metadata:
backbone: R-50b-D8
crop size: (769,769)
lr schd: 80000
inference time (ms/im):
- backend: PyTorch
batch size: 1
- value: 581.4
hardware: V100
backend: PyTorch
batch size: 1
mode: FP32
resolution: (769,769)
value: 581.4
lr schd: 80000
memory (GB): 8.4
Name: deeplabv3plus_r50b-d8_769x769_80k_cityscapes
Results:
- Task: Semantic Segmentation
Dataset: Cityscapes
Metrics:
mIoU: 79.41
mIoU(ms+flip): 80.56
Task: Semantic Segmentation
Config: configs/deeplabv3plus/deeplabv3plus_r50b-d8_769x769_80k_cityscapes.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/deeplabv3plus/deeplabv3plus_r50b-d8_769x769_80k_cityscapes/deeplabv3plus_r50b-d8_769x769_80k_cityscapes_20201225_224655-8b596d1c.pth
- Config: configs/deeplabv3plus/deeplabv3plus_r101b-d8_769x769_80k_cityscapes.py
- Name: deeplabv3plus_r101b-d8_769x769_80k_cityscapes
In Collection: deeplabv3plus
Metadata:
backbone: R-101b-D8
crop size: (769,769)
lr schd: 80000
inference time (ms/im):
- backend: PyTorch
batch size: 1
- value: 909.09
hardware: V100
backend: PyTorch
batch size: 1
mode: FP32
resolution: (769,769)
value: 909.09
lr schd: 80000
memory (GB): 12.3
Name: deeplabv3plus_r101b-d8_769x769_80k_cityscapes
Results:
- Task: Semantic Segmentation
Dataset: Cityscapes
Metrics:
mIoU: 79.88
mIoU(ms+flip): 81.46
Task: Semantic Segmentation
Config: configs/deeplabv3plus/deeplabv3plus_r101b-d8_769x769_80k_cityscapes.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/deeplabv3plus/deeplabv3plus_r101b-d8_769x769_80k_cityscapes/deeplabv3plus_r101b-d8_769x769_80k_cityscapes_20201226_205041-227cdf7c.pth
- Config: configs/deeplabv3plus/deeplabv3plus_r50-d8_512x512_80k_ade20k.py
- Name: deeplabv3plus_r50-d8_512x512_80k_ade20k
In Collection: deeplabv3plus
Metadata:
backbone: R-50-D8
crop size: (512,512)
lr schd: 80000
inference time (ms/im):
- backend: PyTorch
batch size: 1
- value: 47.6
hardware: V100
backend: PyTorch
batch size: 1
mode: FP32
resolution: (512,512)
value: 47.6
lr schd: 80000
memory (GB): 10.6
Name: deeplabv3plus_r50-d8_512x512_80k_ade20k
Results:
- Task: Semantic Segmentation
Dataset: ADE20K
Metrics:
mIoU: 42.72
mIoU(ms+flip): 43.75
Task: Semantic Segmentation
Config: configs/deeplabv3plus/deeplabv3plus_r50-d8_512x512_80k_ade20k.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/deeplabv3plus/deeplabv3plus_r50-d8_512x512_80k_ade20k/deeplabv3plus_r50-d8_512x512_80k_ade20k_20200614_185028-bf1400d8.pth
- Config: configs/deeplabv3plus/deeplabv3plus_r101-d8_512x512_80k_ade20k.py
- Name: deeplabv3plus_r101-d8_512x512_80k_ade20k
In Collection: deeplabv3plus
Metadata:
backbone: R-101-D8
crop size: (512,512)
lr schd: 80000
inference time (ms/im):
- backend: PyTorch
batch size: 1
- value: 70.62
hardware: V100
backend: PyTorch
batch size: 1
mode: FP32
resolution: (512,512)
value: 70.62
lr schd: 80000
memory (GB): 14.1
Name: deeplabv3plus_r101-d8_512x512_80k_ade20k
Results:
- Task: Semantic Segmentation
Dataset: ADE20K
Metrics:
mIoU: 44.6
mIoU(ms+flip): 46.06
Task: Semantic Segmentation
Config: configs/deeplabv3plus/deeplabv3plus_r101-d8_512x512_80k_ade20k.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/deeplabv3plus/deeplabv3plus_r101-d8_512x512_80k_ade20k/deeplabv3plus_r101-d8_512x512_80k_ade20k_20200615_014139-d5730af7.pth
- Config: configs/deeplabv3plus/deeplabv3plus_r50-d8_512x512_160k_ade20k.py
- Name: deeplabv3plus_r50-d8_512x512_160k_ade20k
In Collection: deeplabv3plus
Metadata:
backbone: R-50-D8
crop size: (512,512)
lr schd: 160000
Name: deeplabv3plus_r50-d8_512x512_160k_ade20k
Results:
- Task: Semantic Segmentation
Dataset: ADE20K
Metrics:
mIoU: 43.95
mIoU(ms+flip): 44.93
Task: Semantic Segmentation
Config: configs/deeplabv3plus/deeplabv3plus_r50-d8_512x512_160k_ade20k.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/deeplabv3plus/deeplabv3plus_r50-d8_512x512_160k_ade20k/deeplabv3plus_r50-d8_512x512_160k_ade20k_20200615_124504-6135c7e0.pth
- Config: configs/deeplabv3plus/deeplabv3plus_r101-d8_512x512_160k_ade20k.py
- Name: deeplabv3plus_r101-d8_512x512_160k_ade20k
In Collection: deeplabv3plus
Metadata:
backbone: R-101-D8
crop size: (512,512)
lr schd: 160000
Name: deeplabv3plus_r101-d8_512x512_160k_ade20k
Results:
- Task: Semantic Segmentation
Dataset: ADE20K
Metrics:
mIoU: 45.47
mIoU(ms+flip): 46.35
Task: Semantic Segmentation
Config: configs/deeplabv3plus/deeplabv3plus_r101-d8_512x512_160k_ade20k.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/deeplabv3plus/deeplabv3plus_r101-d8_512x512_160k_ade20k/deeplabv3plus_r101-d8_512x512_160k_ade20k_20200615_123232-38ed86bb.pth
- Config: configs/deeplabv3plus/deeplabv3plus_r50-d8_512x512_20k_voc12aug.py
- Name: deeplabv3plus_r50-d8_512x512_20k_voc12aug
In Collection: deeplabv3plus
Metadata:
backbone: R-50-D8
crop size: (512,512)
lr schd: 20000
inference time (ms/im):
- backend: PyTorch
batch size: 1
- value: 47.62
hardware: V100
backend: PyTorch
batch size: 1
mode: FP32
resolution: (512,512)
value: 47.62
lr schd: 20000
memory (GB): 7.6
Name: deeplabv3plus_r50-d8_512x512_20k_voc12aug
Results:
Dataset: ' Pascal VOC 2012 + Aug'
- Task: Semantic Segmentation
Dataset: ADE20K
Metrics:
mIoU: 75.93
mIoU(ms+flip): 77.5
Task: Semantic Segmentation
Config: configs/deeplabv3plus/deeplabv3plus_r50-d8_512x512_20k_voc12aug.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/deeplabv3plus/deeplabv3plus_r50-d8_512x512_20k_voc12aug/deeplabv3plus_r50-d8_512x512_20k_voc12aug_20200617_102323-aad58ef1.pth
- Config: configs/deeplabv3plus/deeplabv3plus_r101-d8_512x512_20k_voc12aug.py
- Name: deeplabv3plus_r101-d8_512x512_20k_voc12aug
In Collection: deeplabv3plus
Metadata:
backbone: R-101-D8
crop size: (512,512)
lr schd: 20000
inference time (ms/im):
- backend: PyTorch
batch size: 1
- value: 72.05
hardware: V100
backend: PyTorch
batch size: 1
mode: FP32
resolution: (512,512)
value: 72.05
lr schd: 20000
memory (GB): 11.0
Name: deeplabv3plus_r101-d8_512x512_20k_voc12aug
Results:
Dataset: ' Pascal VOC 2012 + Aug'
- Task: Semantic Segmentation
Dataset: ADE20K
Metrics:
mIoU: 77.22
mIoU(ms+flip): 78.59
Task: Semantic Segmentation
Config: configs/deeplabv3plus/deeplabv3plus_r101-d8_512x512_20k_voc12aug.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/deeplabv3plus/deeplabv3plus_r101-d8_512x512_20k_voc12aug/deeplabv3plus_r101-d8_512x512_20k_voc12aug_20200617_102345-c7ff3d56.pth
- Config: configs/deeplabv3plus/deeplabv3plus_r50-d8_512x512_40k_voc12aug.py
- Name: deeplabv3plus_r50-d8_512x512_40k_voc12aug
In Collection: deeplabv3plus
Metadata:
backbone: R-50-D8
crop size: (512,512)
lr schd: 40000
Name: deeplabv3plus_r50-d8_512x512_40k_voc12aug
Results:
Dataset: ' Pascal VOC 2012 + Aug'
- Task: Semantic Segmentation
Dataset: ADE20K
Metrics:
mIoU: 76.81
mIoU(ms+flip): 77.57
Task: Semantic Segmentation
Config: configs/deeplabv3plus/deeplabv3plus_r50-d8_512x512_40k_voc12aug.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/deeplabv3plus/deeplabv3plus_r50-d8_512x512_40k_voc12aug/deeplabv3plus_r50-d8_512x512_40k_voc12aug_20200613_161759-e1b43aa9.pth
- Config: configs/deeplabv3plus/deeplabv3plus_r101-d8_512x512_40k_voc12aug.py
- Name: deeplabv3plus_r101-d8_512x512_40k_voc12aug
In Collection: deeplabv3plus
Metadata:
backbone: R-101-D8
crop size: (512,512)
lr schd: 40000
Name: deeplabv3plus_r101-d8_512x512_40k_voc12aug
Results:
Dataset: ' Pascal VOC 2012 + Aug'
- Task: Semantic Segmentation
Dataset: ADE20K
Metrics:
mIoU: 78.62
mIoU(ms+flip): 79.53
Task: Semantic Segmentation
Config: configs/deeplabv3plus/deeplabv3plus_r101-d8_512x512_40k_voc12aug.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/deeplabv3plus/deeplabv3plus_r101-d8_512x512_40k_voc12aug/deeplabv3plus_r101-d8_512x512_40k_voc12aug_20200613_205333-faf03387.pth
- Config: configs/deeplabv3plus/deeplabv3plus_r101-d8_480x480_40k_pascal_context.py
- Name: deeplabv3plus_r101-d8_480x480_40k_pascal_context
In Collection: deeplabv3plus
Metadata:
backbone: R-101-D8
crop size: (480,480)
lr schd: 40000
inference time (ms/im):
- backend: PyTorch
batch size: 1
- value: 110.01
hardware: V100
backend: PyTorch
batch size: 1
mode: FP32
resolution: (480,480)
value: 110.01
lr schd: 40000
Name: deeplabv3plus_r101-d8_480x480_40k_pascal_context
Results:
Dataset: ' Pascal Context'
- Task: Semantic Segmentation
Dataset: ADE20K
Metrics:
mIoU: 47.3
mIoU(ms+flip): 48.47
Task: Semantic Segmentation
Config: configs/deeplabv3plus/deeplabv3plus_r101-d8_480x480_40k_pascal_context.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/deeplabv3plus/deeplabv3plus_r101-d8_480x480_40k_pascal_context/deeplabv3plus_r101-d8_480x480_40k_pascal_context_20200911_165459-d3c8a29e.pth
- Config: configs/deeplabv3plus/deeplabv3plus_r101-d8_480x480_80k_pascal_context.py
- Name: deeplabv3plus_r101-d8_480x480_80k_pascal_context
In Collection: deeplabv3plus
Metadata:
backbone: R-101-D8
crop size: (480,480)
lr schd: 80000
Name: deeplabv3plus_r101-d8_480x480_80k_pascal_context
Results:
Dataset: ' Pascal Context'
- Task: Semantic Segmentation
Dataset: ADE20K
Metrics:
mIoU: 47.23
mIoU(ms+flip): 48.26
Task: Semantic Segmentation
Config: configs/deeplabv3plus/deeplabv3plus_r101-d8_480x480_80k_pascal_context.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/deeplabv3plus/deeplabv3plus_r101-d8_480x480_80k_pascal_context/deeplabv3plus_r101-d8_480x480_80k_pascal_context_20200911_155322-145d3ee8.pth
- Config: configs/deeplabv3plus/deeplabv3plus_r101-d8_480x480_40k_pascal_context_59.py
- Name: deeplabv3plus_r101-d8_480x480_40k_pascal_context_59
In Collection: deeplabv3plus
Metadata:
backbone: R-101-D8
crop size: (480,480)
lr schd: 40000
Name: deeplabv3plus_r101-d8_480x480_40k_pascal_context_59
Results:
Dataset: ' Pascal Context 59'
- Task: Semantic Segmentation
Dataset: ADE20K
Metrics:
mIoU: 52.86
mIoU(ms+flip): 54.54
Task: Semantic Segmentation
Config: configs/deeplabv3plus/deeplabv3plus_r101-d8_480x480_40k_pascal_context_59.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/deeplabv3plus/deeplabv3plus_r101-d8_480x480_40k_pascal_context_59/deeplabv3plus_r101-d8_480x480_40k_pascal_context_59_20210416_111233-ed937f15.pth
- Config: configs/deeplabv3plus/deeplabv3plus_r101-d8_480x480_80k_pascal_context_59.py
- Name: deeplabv3plus_r101-d8_480x480_80k_pascal_context_59
In Collection: deeplabv3plus
Metadata:
backbone: R-101-D8
crop size: (480,480)
lr schd: 80000
Name: deeplabv3plus_r101-d8_480x480_80k_pascal_context_59
Results:
Dataset: ' Pascal Context 59'
- Task: Semantic Segmentation
Dataset: ADE20K
Metrics:
mIoU: 53.2
mIoU(ms+flip): 54.67
Task: Semantic Segmentation
Config: configs/deeplabv3plus/deeplabv3plus_r101-d8_480x480_80k_pascal_context_59.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/deeplabv3plus/deeplabv3plus_r101-d8_480x480_80k_pascal_context_59/deeplabv3plus_r101-d8_480x480_80k_pascal_context_59_20210416_111127-7ca0331d.pth

9
configs/dmnet/README.md
View File

@ -4,6 +4,13 @@
<!-- [ALGORITHM] -->
<a href="https://github.com/Junjun2016/DMNet">Official Repo</a>
<a href="https://github.com/open-mmlab/mmsegmentation/blob/v0.17.0/mmseg/models/decode_heads/dm_head.py#L93">Code Snippet</a>
<details>
<summary align="right"><a href="https://openaccess.thecvf.com/content_ICCV_2019/papers/He_Dynamic_Multi-Scale_Filters_for_Semantic_Segmentation_ICCV_2019_paper.pdf">DMNet (ICCV'2019)</a></summary>
```latex
@InProceedings{He_2019_ICCV,
author = {He, Junjun and Deng, Zhongying and Qiao, Yu},
@ -14,6 +21,8 @@ year = {2019}
}
```
</details>
## Results and models
### Cityscapes

133
configs/dmnet/dmnet.yml
View File

@ -1,223 +1,232 @@
Collections:
- Metadata:
- Name: dmnet
Metadata:
Training Data:
- Cityscapes
- ADE20K
Name: dmnet
Paper:
URL: https://openaccess.thecvf.com/content_ICCV_2019/papers/He_Dynamic_Multi-Scale_Filters_for_Semantic_Segmentation_ICCV_2019_paper.pdf
Title: Dynamic Multi-scale Filters for Semantic Segmentation
README: configs/dmnet/README.md
Code:
URL: https://github.com/open-mmlab/mmsegmentation/blob/v0.17.0/mmseg/models/decode_heads/dm_head.py#L93
Version: v0.17.0
Converted From:
Code: https://github.com/Junjun2016/DMNet
Models:
- Config: configs/dmnet/dmnet_r50-d8_512x1024_40k_cityscapes.py
- Name: dmnet_r50-d8_512x1024_40k_cityscapes
In Collection: dmnet
Metadata:
backbone: R-50-D8
crop size: (512,1024)
lr schd: 40000
inference time (ms/im):
- backend: PyTorch
batch size: 1
- value: 273.22
hardware: V100
backend: PyTorch
batch size: 1
mode: FP32
resolution: (512,1024)
value: 273.22
lr schd: 40000
memory (GB): 7.0
Name: dmnet_r50-d8_512x1024_40k_cityscapes
Results:
- Task: Semantic Segmentation
Dataset: Cityscapes
Metrics:
mIoU: 77.78
mIoU(ms+flip): 79.14
Task: Semantic Segmentation
Config: configs/dmnet/dmnet_r50-d8_512x1024_40k_cityscapes.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/dmnet/dmnet_r50-d8_512x1024_40k_cityscapes/dmnet_r50-d8_512x1024_40k_cityscapes_20201215_042326-615373cf.pth
- Config: configs/dmnet/dmnet_r101-d8_512x1024_40k_cityscapes.py
- Name: dmnet_r101-d8_512x1024_40k_cityscapes
In Collection: dmnet
Metadata:
backbone: R-101-D8
crop size: (512,1024)
lr schd: 40000
inference time (ms/im):
- backend: PyTorch
batch size: 1
- value: 393.7
hardware: V100
backend: PyTorch
batch size: 1
mode: FP32
resolution: (512,1024)
value: 393.7
lr schd: 40000
memory (GB): 10.6
Name: dmnet_r101-d8_512x1024_40k_cityscapes
Results:
- Task: Semantic Segmentation
Dataset: Cityscapes
Metrics:
mIoU: 78.37
mIoU(ms+flip): 79.72
Task: Semantic Segmentation
Config: configs/dmnet/dmnet_r101-d8_512x1024_40k_cityscapes.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/dmnet/dmnet_r101-d8_512x1024_40k_cityscapes/dmnet_r101-d8_512x1024_40k_cityscapes_20201215_043100-8291e976.pth
- Config: configs/dmnet/dmnet_r50-d8_769x769_40k_cityscapes.py
- Name: dmnet_r50-d8_769x769_40k_cityscapes
In Collection: dmnet
Metadata:
backbone: R-50-D8
crop size: (769,769)
lr schd: 40000
inference time (ms/im):
- backend: PyTorch
batch size: 1
- value: 636.94
hardware: V100
backend: PyTorch
batch size: 1
mode: FP32
resolution: (769,769)
value: 636.94
lr schd: 40000
memory (GB): 7.9
Name: dmnet_r50-d8_769x769_40k_cityscapes
Results:
- Task: Semantic Segmentation
Dataset: Cityscapes
Metrics:
mIoU: 78.49
mIoU(ms+flip): 80.27
Task: Semantic Segmentation
Config: configs/dmnet/dmnet_r50-d8_769x769_40k_cityscapes.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/dmnet/dmnet_r50-d8_769x769_40k_cityscapes/dmnet_r50-d8_769x769_40k_cityscapes_20201215_093706-e7f0e23e.pth
- Config: configs/dmnet/dmnet_r101-d8_769x769_40k_cityscapes.py
- Name: dmnet_r101-d8_769x769_40k_cityscapes
In Collection: dmnet
Metadata:
backbone: R-101-D8
crop size: (769,769)
lr schd: 40000
inference time (ms/im):
- backend: PyTorch
batch size: 1
- value: 990.1
hardware: V100
backend: PyTorch
batch size: 1
mode: FP32
resolution: (769,769)
value: 990.1
lr schd: 40000
memory (GB): 12.0
Name: dmnet_r101-d8_769x769_40k_cityscapes
Results:
- Task: Semantic Segmentation
Dataset: Cityscapes
Metrics:
mIoU: 77.62
mIoU(ms+flip): 78.94
Task: Semantic Segmentation
Config: configs/dmnet/dmnet_r101-d8_769x769_40k_cityscapes.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/dmnet/dmnet_r101-d8_769x769_40k_cityscapes/dmnet_r101-d8_769x769_40k_cityscapes_20201215_081348-a74261f6.pth
- Config: configs/dmnet/dmnet_r50-d8_512x1024_80k_cityscapes.py
- Name: dmnet_r50-d8_512x1024_80k_cityscapes
In Collection: dmnet
Metadata:
backbone: R-50-D8
crop size: (512,1024)
lr schd: 80000
Name: dmnet_r50-d8_512x1024_80k_cityscapes
Results:
- Task: Semantic Segmentation
Dataset: Cityscapes
Metrics:
mIoU: 79.07
mIoU(ms+flip): 80.22
Task: Semantic Segmentation
Config: configs/dmnet/dmnet_r50-d8_512x1024_80k_cityscapes.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/dmnet/dmnet_r50-d8_512x1024_80k_cityscapes/dmnet_r50-d8_512x1024_80k_cityscapes_20201215_053728-3c8893b9.pth
- Config: configs/dmnet/dmnet_r101-d8_512x1024_80k_cityscapes.py
- Name: dmnet_r101-d8_512x1024_80k_cityscapes
In Collection: dmnet
Metadata:
backbone: R-101-D8
crop size: (512,1024)
lr schd: 80000
Name: dmnet_r101-d8_512x1024_80k_cityscapes
Results:
- Task: Semantic Segmentation
Dataset: Cityscapes
Metrics:
mIoU: 79.64
mIoU(ms+flip): 80.67
Task: Semantic Segmentation
Config: configs/dmnet/dmnet_r101-d8_512x1024_80k_cityscapes.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/dmnet/dmnet_r101-d8_512x1024_80k_cityscapes/dmnet_r101-d8_512x1024_80k_cityscapes_20201215_031718-fa081cb8.pth
- Config: configs/dmnet/dmnet_r50-d8_769x769_80k_cityscapes.py
- Name: dmnet_r50-d8_769x769_80k_cityscapes
In Collection: dmnet
Metadata:
backbone: R-50-D8
crop size: (769,769)
lr schd: 80000
Name: dmnet_r50-d8_769x769_80k_cityscapes
Results:
- Task: Semantic Segmentation
Dataset: Cityscapes
Metrics:
mIoU: 79.22
mIoU(ms+flip): 80.55
Task: Semantic Segmentation
Config: configs/dmnet/dmnet_r50-d8_769x769_80k_cityscapes.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/dmnet/dmnet_r50-d8_769x769_80k_cityscapes/dmnet_r50-d8_769x769_80k_cityscapes_20201215_034006-6060840e.pth
- Config: configs/dmnet/dmnet_r101-d8_769x769_80k_cityscapes.py
- Name: dmnet_r101-d8_769x769_80k_cityscapes
In Collection: dmnet
Metadata:
backbone: R-101-D8
crop size: (769,769)
lr schd: 80000
Name: dmnet_r101-d8_769x769_80k_cityscapes
Results:
- Task: Semantic Segmentation
Dataset: Cityscapes
Metrics:
mIoU: 79.19
mIoU(ms+flip): 80.65
Task: Semantic Segmentation
Config: configs/dmnet/dmnet_r101-d8_769x769_80k_cityscapes.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/dmnet/dmnet_r101-d8_769x769_80k_cityscapes/dmnet_r101-d8_769x769_80k_cityscapes_20201215_082810-7f0de59a.pth
- Config: configs/dmnet/dmnet_r50-d8_512x512_80k_ade20k.py
- Name: dmnet_r50-d8_512x512_80k_ade20k
In Collection: dmnet
Metadata:
backbone: R-50-D8
crop size: (512,512)
lr schd: 80000
inference time (ms/im):
- backend: PyTorch
batch size: 1
- value: 47.73
hardware: V100
backend: PyTorch
batch size: 1
mode: FP32
resolution: (512,512)
value: 47.73
lr schd: 80000
memory (GB): 9.4
Name: dmnet_r50-d8_512x512_80k_ade20k
Results:
- Task: Semantic Segmentation
Dataset: ADE20K
Metrics:
mIoU: 42.37
mIoU(ms+flip): 43.62
Task: Semantic Segmentation
Config: configs/dmnet/dmnet_r50-d8_512x512_80k_ade20k.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/dmnet/dmnet_r50-d8_512x512_80k_ade20k/dmnet_r50-d8_512x512_80k_ade20k_20201215_144744-f89092a6.pth
- Config: configs/dmnet/dmnet_r101-d8_512x512_80k_ade20k.py
- Name: dmnet_r101-d8_512x512_80k_ade20k
In Collection: dmnet
Metadata:
backbone: R-101-D8
crop size: (512,512)
lr schd: 80000
inference time (ms/im):
- backend: PyTorch
batch size: 1
- value: 72.05
hardware: V100
backend: PyTorch
batch size: 1
mode: FP32
resolution: (512,512)
value: 72.05
lr schd: 80000
memory (GB): 13.0
Name: dmnet_r101-d8_512x512_80k_ade20k
Results:
- Task: Semantic Segmentation
Dataset: ADE20K
Metrics:
mIoU: 45.34
mIoU(ms+flip): 46.13
Task: Semantic Segmentation
Config: configs/dmnet/dmnet_r101-d8_512x512_80k_ade20k.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/dmnet/dmnet_r101-d8_512x512_80k_ade20k/dmnet_r101-d8_512x512_80k_ade20k_20201215_104812-bfa45311.pth
- Config: configs/dmnet/dmnet_r50-d8_512x512_160k_ade20k.py
- Name: dmnet_r50-d8_512x512_160k_ade20k
In Collection: dmnet
Metadata:
backbone: R-50-D8
crop size: (512,512)
lr schd: 160000
Name: dmnet_r50-d8_512x512_160k_ade20k
Results:
- Task: Semantic Segmentation
Dataset: ADE20K
Metrics:
mIoU: 43.15
mIoU(ms+flip): 44.17
Task: Semantic Segmentation
Config: configs/dmnet/dmnet_r50-d8_512x512_160k_ade20k.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/dmnet/dmnet_r50-d8_512x512_160k_ade20k/dmnet_r50-d8_512x512_160k_ade20k_20201215_115313-025ab3f9.pth
- Config: configs/dmnet/dmnet_r101-d8_512x512_160k_ade20k.py
- Name: dmnet_r101-d8_512x512_160k_ade20k
In Collection: dmnet
Metadata:
backbone: R-101-D8
crop size: (512,512)
lr schd: 160000
Name: dmnet_r101-d8_512x512_160k_ade20k
Results:
- Task: Semantic Segmentation
Dataset: ADE20K
Metrics:
mIoU: 45.42
mIoU(ms+flip): 46.76
Task: Semantic Segmentation
Config: configs/dmnet/dmnet_r101-d8_512x512_160k_ade20k.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/dmnet/dmnet_r101-d8_512x512_160k_ade20k/dmnet_r101-d8_512x512_160k_ade20k_20201215_111145-a0bc02ef.pth

9
configs/dnlnet/README.md
View File

@ -4,6 +4,13 @@
<!-- [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>
<details>
<summary align="right"><a href="https://arxiv.org/abs/2006.06668">DNLNet (ECCV'2020)</a></summary>
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
@ -17,6 +24,8 @@ This example is to reproduce ["Disentangled Non-Local Neural Networks"](https://
}
```
</details>
## Results and models (in progress)
### Cityscapes

133
configs/dnlnet/dnlnet.yml
View File

@ -1,219 +1,228 @@
Collections:
- Metadata:
- Name: dnlnet
Metadata:
Training Data:
- Cityscapes
- ADE20K
Name: dnlnet
Paper:
URL: https://arxiv.org/abs/2006.06668
Title: Disentangled Non-Local Neural Networks
README: configs/dnlnet/README.md
Code:
URL: https://github.com/open-mmlab/mmsegmentation/blob/v0.17.0/mmseg/models/decode_heads/dnl_head.py#L88
Version: v0.17.0
Converted From:
Code: https://github.com/yinmh17/DNL-Semantic-Segmentation
Models:
- Config: configs/dnlnet/dnl_r50-d8_512x1024_40k_cityscapes.py
- Name: dnl_r50-d8_512x1024_40k_cityscapes
In Collection: dnlnet
Metadata:
backbone: R-50-D8
crop size: (512,1024)
lr schd: 40000
inference time (ms/im):
- backend: PyTorch
batch size: 1
- value: 390.62
hardware: V100
backend: PyTorch
batch size: 1
mode: FP32
resolution: (512,1024)
value: 390.62
lr schd: 40000
memory (GB): 7.3
Name: dnl_r50-d8_512x1024_40k_cityscapes
Results:
- Task: Semantic Segmentation
Dataset: Cityscapes
Metrics:
mIoU: 78.61
Task: Semantic Segmentation
Config: configs/dnlnet/dnl_r50-d8_512x1024_40k_cityscapes.py
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
- Config: configs/dnlnet/dnl_r101-d8_512x1024_40k_cityscapes.py
- Name: dnl_r101-d8_512x1024_40k_cityscapes
In Collection: dnlnet
Metadata:
backbone: R-101-D8
crop size: (512,1024)
lr schd: 40000
inference time (ms/im):
- backend: PyTorch
batch size: 1
- value: 510.2
hardware: V100
backend: PyTorch
batch size: 1
mode: FP32
resolution: (512,1024)
value: 510.2
lr schd: 40000
memory (GB): 10.9
Name: dnl_r101-d8_512x1024_40k_cityscapes
Results:
- Task: Semantic Segmentation
Dataset: Cityscapes
Metrics:
mIoU: 78.31
Task: Semantic Segmentation
Config: configs/dnlnet/dnl_r101-d8_512x1024_40k_cityscapes.py
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
- Config: configs/dnlnet/dnl_r50-d8_769x769_40k_cityscapes.py
- Name: dnl_r50-d8_769x769_40k_cityscapes
In Collection: dnlnet
Metadata:
backbone: R-50-D8
crop size: (769,769)
lr schd: 40000
inference time (ms/im):
- backend: PyTorch
batch size: 1
- value: 666.67
hardware: V100
backend: PyTorch
batch size: 1
mode: FP32
resolution: (769,769)
value: 666.67
lr schd: 40000
memory (GB): 9.2
Name: dnl_r50-d8_769x769_40k_cityscapes
Results:
- Task: Semantic Segmentation
Dataset: Cityscapes
Metrics:
mIoU: 78.44
mIoU(ms+flip): 80.27
Task: Semantic Segmentation
Config: configs/dnlnet/dnl_r50-d8_769x769_40k_cityscapes.py
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
- Config: configs/dnlnet/dnl_r101-d8_769x769_40k_cityscapes.py
- Name: dnl_r101-d8_769x769_40k_cityscapes
In Collection: dnlnet
Metadata:
backbone: R-101-D8
crop size: (769,769)
lr schd: 40000
inference time (ms/im):
- backend: PyTorch
batch size: 1
- value: 980.39
hardware: V100
backend: PyTorch
batch size: 1
mode: FP32
resolution: (769,769)
value: 980.39
lr schd: 40000
memory (GB): 12.6
Name: dnl_r101-d8_769x769_40k_cityscapes
Results:
- Task: Semantic Segmentation
Dataset: Cityscapes
Metrics:
mIoU: 76.39
mIoU(ms+flip): 77.77
Task: Semantic Segmentation
Config: configs/dnlnet/dnl_r101-d8_769x769_40k_cityscapes.py
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
- Config: configs/dnlnet/dnl_r50-d8_512x1024_80k_cityscapes.py
- Name: dnl_r50-d8_512x1024_80k_cityscapes
In Collection: dnlnet
Metadata:
backbone: R-50-D8
crop size: (512,1024)
lr schd: 80000
Name: dnl_r50-d8_512x1024_80k_cityscapes
Results:
- Task: Semantic Segmentation
Dataset: Cityscapes
Metrics:
mIoU: 79.33
Task: Semantic Segmentation
Config: configs/dnlnet/dnl_r50-d8_512x1024_80k_cityscapes.py
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
- Config: configs/dnlnet/dnl_r101-d8_512x1024_80k_cityscapes.py
- Name: dnl_r101-d8_512x1024_80k_cityscapes
In Collection: dnlnet
Metadata:
backbone: R-101-D8
crop size: (512,1024)
lr schd: 80000
Name: dnl_r101-d8_512x1024_80k_cityscapes
Results:
- Task: Semantic Segmentation
Dataset: Cityscapes
Metrics:
mIoU: 80.41
Task: Semantic Segmentation
Config: configs/dnlnet/dnl_r101-d8_512x1024_80k_cityscapes.py
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
- Config: configs/dnlnet/dnl_r50-d8_769x769_80k_cityscapes.py
- Name: dnl_r50-d8_769x769_80k_cityscapes
In Collection: dnlnet
Metadata:
backbone: R-50-D8
crop size: (769,769)
lr schd: 80000
Name: dnl_r50-d8_769x769_80k_cityscapes
Results:
- Task: Semantic Segmentation
Dataset: Cityscapes
Metrics:
mIoU: 79.36
mIoU(ms+flip): 80.7
Task: Semantic Segmentation
Config: configs/dnlnet/dnl_r50-d8_769x769_80k_cityscapes.py
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
- Config: configs/dnlnet/dnl_r101-d8_769x769_80k_cityscapes.py
- Name: dnl_r101-d8_769x769_80k_cityscapes
In Collection: dnlnet
Metadata:
backbone: R-101-D8
crop size: (769,769)
lr schd: 80000
Name: dnl_r101-d8_769x769_80k_cityscapes
Results:
- Task: Semantic Segmentation
Dataset: Cityscapes
Metrics:
mIoU: 79.41
mIoU(ms+flip): 80.68
Task: Semantic Segmentation
Config: configs/dnlnet/dnl_r101-d8_769x769_80k_cityscapes.py
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
- Config: configs/dnlnet/dnl_r50-d8_512x512_80k_ade20k.py
- Name: dnl_r50-d8_512x512_80k_ade20k
In Collection: dnlnet
Metadata:
backbone: R-50-D8
crop size: (512,512)
lr schd: 80000
inference time (ms/im):
- backend: PyTorch
batch size: 1
- value: 48.4
hardware: V100
backend: PyTorch
batch size: 1
mode: FP32
resolution: (512,512)
value: 48.4
lr schd: 80000
memory (GB): 8.8
Name: dnl_r50-d8_512x512_80k_ade20k
Results:
- Task: Semantic Segmentation
Dataset: ADE20K
Metrics:
mIoU: 41.76
mIoU(ms+flip): 42.99
Task: Semantic Segmentation
Config: configs/dnlnet/dnl_r50-d8_512x512_80k_ade20k.py
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
- Config: configs/dnlnet/dnl_r101-d8_512x512_80k_ade20k.py
- Name: dnl_r101-d8_512x512_80k_ade20k
In Collection: dnlnet
Metadata:
backbone: R-101-D8
crop size: (512,512)
lr schd: 80000
inference time (ms/im):
- backend: PyTorch
batch size: 1
- value: 79.74
hardware: V100
backend: PyTorch
batch size: 1
mode: FP32
resolution: (512,512)
value: 79.74
lr schd: 80000
memory (GB): 12.8
Name: dnl_r101-d8_512x512_80k_ade20k
Results:
- Task: Semantic Segmentation
Dataset: ADE20K
Metrics:
mIoU: 43.76
mIoU(ms+flip): 44.91
Task: Semantic Segmentation
Config: configs/dnlnet/dnl_r101-d8_512x512_80k_ade20k.py
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
- Config: configs/dnlnet/dnl_r50-d8_512x512_160k_ade20k.py
- Name: dnl_r50-d8_512x512_160k_ade20k
In Collection: dnlnet
Metadata:
backbone: R-50-D8
crop size: (512,512)
lr schd: 160000
Name: dnl_r50-d8_512x512_160k_ade20k
Results:
- Task: Semantic Segmentation
Dataset: ADE20K
Metrics:
mIoU: 41.87
mIoU(ms+flip): 43.01
Task: Semantic Segmentation
Config: configs/dnlnet/dnl_r50-d8_512x512_160k_ade20k.py
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
- Config: configs/dnlnet/dnl_r101-d8_512x512_160k_ade20k.py
- Name: dnl_r101-d8_512x512_160k_ade20k
In Collection: dnlnet
Metadata:
backbone: R-101-D8
crop size: (512,512)
lr schd: 160000
Name: dnl_r101-d8_512x512_160k_ade20k
Results:
- Task: Semantic Segmentation
Dataset: ADE20K
Metrics:
mIoU: 44.25
mIoU(ms+flip): 45.78
Task: Semantic Segmentation
Config: configs/dnlnet/dnl_r101-d8_512x512_160k_ade20k.py
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

9
configs/dpt/README.md
View File

@ -4,6 +4,13 @@
<!-- [ALGORITHM] -->
<a href="https://github.com/isl-org/DPT">Official Repo</a>
<a href="https://github.com/open-mmlab/mmsegmentation/blob/v0.17.0/mmseg/models/decode_heads/dpt_head.py#L215">Code Snippet</a>
<details>
<summary align="right"><a href="https://arxiv.org/abs/2103.13413">DPT (ArXiv'2021)</a></summary>
```latex
@article{dosoViTskiy2020,
title={An Image is Worth 16x16 Words: Transformers for Image Recognition at Scale},
@ -20,6 +27,8 @@
}
```
</details>
## Usage
To use other repositories' pre-trained models, it is necessary to convert keys.

27
configs/dpt/dpt.yml
View File

@ -1,28 +1,37 @@
Collections:
- Metadata:
- Name: dpt
Metadata:
Training Data:
- ADE20K
Name: dpt
Paper:
URL: https://arxiv.org/abs/2103.13413
Title: Vision Transformer for Dense Prediction
README: configs/dpt/README.md
Code:
URL: https://github.com/open-mmlab/mmsegmentation/blob/v0.17.0/mmseg/models/decode_heads/dpt_head.py#L215
Version: v0.17.0
Converted From:
Code: https://github.com/isl-org/DPT
Models:
- Config: configs/dpt/dpt_vit-b16_512x512_160k_ade20k.py
- Name: dpt_vit-b16_512x512_160k_ade20k
In Collection: dpt
Metadata:
backbone: ViT-B
crop size: (512,512)
lr schd: 160000
inference time (ms/im):
- backend: PyTorch
batch size: 1
- value: 96.06
hardware: V100
backend: PyTorch
batch size: 1
mode: FP32
resolution: (512,512)
value: 96.06
lr schd: 160000
memory (GB): 8.09
Name: dpt_vit-b16_512x512_160k_ade20k
Results:
- Task: Semantic Segmentation
Dataset: ADE20K
Metrics:
mIoU: 46.97
mIoU(ms+flip): 48.34
Task: Semantic Segmentation
Config: configs/dpt/dpt_vit-b16_512x512_160k_ade20k.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/dpt/dpt_vit-b16_512x512_160k_ade20k/dpt_vit-b16_512x512_160k_ade20k-db31cf52.pth

9
configs/emanet/README.md
View File

@ -4,6 +4,13 @@
<!-- [ALGORITHM] -->
<a href="https://xialipku.github.io/EMANet">Official Repo</a>
<a href="https://github.com/open-mmlab/mmsegmentation/blob/v0.17.0/mmseg/models/decode_heads/ema_head.py#L80">Code Snippet</a>
<details>
<summary align="right"><a href="https://arxiv.org/abs/1907.13426">EMANet (ICCV'2019)</a></summary>
```latex
@inproceedings{li2019expectation,
title={Expectation-maximization attention networks for semantic segmentation},
@ -14,6 +21,8 @@
}
```
</details>
## Results and models
### Cityscapes

69
configs/emanet/emanet.yml
View File

@ -1,94 +1,103 @@
Collections:
- Metadata:
- Name: emanet
Metadata:
Training Data:
- Cityscapes
Name: emanet
Paper:
URL: https://arxiv.org/abs/1907.13426
Title: Expectation-Maximization Attention Networks for Semantic Segmentation
README: configs/emanet/README.md
Code:
URL: https://github.com/open-mmlab/mmsegmentation/blob/v0.17.0/mmseg/models/decode_heads/ema_head.py#L80
Version: v0.17.0
Converted From:
Code: https://xialipku.github.io/EMANet
Models:
- Config: configs/emanet/emanet_r50-d8_512x1024_80k_cityscapes.py
- Name: emanet_r50-d8_512x1024_80k_cityscapes
In Collection: emanet
Metadata:
backbone: R-50-D8
crop size: (512,1024)
lr schd: 80000
inference time (ms/im):
- backend: PyTorch
batch size: 1
- value: 218.34
hardware: V100
backend: PyTorch
batch size: 1
mode: FP32
resolution: (512,1024)
value: 218.34
lr schd: 80000
memory (GB): 5.4
Name: emanet_r50-d8_512x1024_80k_cityscapes
Results:
- Task: Semantic Segmentation
Dataset: Cityscapes
Metrics:
mIoU: 77.59
mIoU(ms+flip): 79.44
Task: Semantic Segmentation
Config: configs/emanet/emanet_r50-d8_512x1024_80k_cityscapes.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/emanet/emanet_r50-d8_512x1024_80k_cityscapes/emanet_r50-d8_512x1024_80k_cityscapes_20200901_100301-c43fcef1.pth
- Config: configs/emanet/emanet_r101-d8_512x1024_80k_cityscapes.py
- Name: emanet_r101-d8_512x1024_80k_cityscapes
In Collection: emanet
Metadata:
backbone: R-101-D8
crop size: (512,1024)
lr schd: 80000
inference time (ms/im):
- backend: PyTorch
batch size: 1
- value: 348.43
hardware: V100
backend: PyTorch
batch size: 1
mode: FP32
resolution: (512,1024)
value: 348.43
lr schd: 80000
memory (GB): 6.2
Name: emanet_r101-d8_512x1024_80k_cityscapes
Results:
- Task: Semantic Segmentation
Dataset: Cityscapes
Metrics:
mIoU: 79.1
mIoU(ms+flip): 81.21
Task: Semantic Segmentation
Config: configs/emanet/emanet_r101-d8_512x1024_80k_cityscapes.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/emanet/emanet_r101-d8_512x1024_80k_cityscapes/emanet_r101-d8_512x1024_80k_cityscapes_20200901_100301-2d970745.pth
- Config: configs/emanet/emanet_r50-d8_769x769_80k_cityscapes.py
- Name: emanet_r50-d8_769x769_80k_cityscapes
In Collection: emanet
Metadata:
backbone: R-50-D8
crop size: (769,769)
lr schd: 80000
inference time (ms/im):
- backend: PyTorch
batch size: 1
- value: 507.61
hardware: V100
backend: PyTorch
batch size: 1
mode: FP32
resolution: (769,769)
value: 507.61
lr schd: 80000
memory (GB): 8.9
Name: emanet_r50-d8_769x769_80k_cityscapes
Results:
- Task: Semantic Segmentation
Dataset: Cityscapes
Metrics:
mIoU: 79.33
mIoU(ms+flip): 80.49
Task: Semantic Segmentation
Config: configs/emanet/emanet_r50-d8_769x769_80k_cityscapes.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/emanet/emanet_r50-d8_769x769_80k_cityscapes/emanet_r50-d8_769x769_80k_cityscapes_20200901_100301-16f8de52.pth
- Config: configs/emanet/emanet_r101-d8_769x769_80k_cityscapes.py
- Name: emanet_r101-d8_769x769_80k_cityscapes
In Collection: emanet
Metadata:
backbone: R-101-D8
crop size: (769,769)
lr schd: 80000
inference time (ms/im):
- backend: PyTorch
batch size: 1
- value: 819.67
hardware: V100
backend: PyTorch
batch size: 1
mode: FP32
resolution: (769,769)
value: 819.67
lr schd: 80000
memory (GB): 10.1
Name: emanet_r101-d8_769x769_80k_cityscapes
Results:
- Task: Semantic Segmentation
Dataset: Cityscapes
Metrics:
mIoU: 79.62
mIoU(ms+flip): 81.0
Task: Semantic Segmentation
Config: configs/emanet/emanet_r101-d8_769x769_80k_cityscapes.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/emanet/emanet_r101-d8_769x769_80k_cityscapes/emanet_r101-d8_769x769_80k_cityscapes_20200901_100301-47a324ce.pth

9
configs/encnet/README.md
View File

@ -4,6 +4,13 @@
<!-- [ALGORITHM] -->
<a href="https://github.com/zhanghang1989/PyTorch-Encoding">Official Repo</a>
<a href="https://github.com/open-mmlab/mmsegmentation/blob/v0.17.0/mmseg/models/decode_heads/enc_head.py#L63">Code Snippet</a>
<details>
<summary align="right"><a href="https://arxiv.org/abs/1803.08904">EncNet (CVPR'2018)</a></summary>
```latex
@InProceedings{Zhang_2018_CVPR,
author = {Zhang, Hang and Dana, Kristin and Shi, Jianping and Zhang, Zhongyue and Wang, Xiaogang and Tyagi, Ambrish and Agrawal, Amit},
@ -14,6 +21,8 @@ year = {2018}
}
```
</details>
## Results and models
### Cityscapes

133
configs/encnet/encnet.yml
View File

@ -1,223 +1,232 @@
Collections:
- Metadata:
- Name: encnet
Metadata:
Training Data:
- Cityscapes
- ADE20K
Name: encnet
Paper:
URL: https://arxiv.org/abs/1803.08904
Title: Context Encoding for Semantic Segmentation
README: configs/encnet/README.md
Code:
URL: https://github.com/open-mmlab/mmsegmentation/blob/v0.17.0/mmseg/models/decode_heads/enc_head.py#L63
Version: v0.17.0
Converted From:
Code: https://github.com/zhanghang1989/PyTorch-Encoding
Models:
- Config: configs/encnet/encnet_r50-d8_512x1024_40k_cityscapes.py
- Name: encnet_r50-d8_512x1024_40k_cityscapes
In Collection: encnet
Metadata:
backbone: R-50-D8
crop size: (512,1024)
lr schd: 40000
inference time (ms/im):
- backend: PyTorch
batch size: 1
- value: 218.34
hardware: V100
backend: PyTorch
batch size: 1
mode: FP32
resolution: (512,1024)
value: 218.34
lr schd: 40000
memory (GB): 8.6
Name: encnet_r50-d8_512x1024_40k_cityscapes
Results:
- Task: Semantic Segmentation
Dataset: Cityscapes
Metrics:
mIoU: 75.67
mIoU(ms+flip): 77.08
Task: Semantic Segmentation
Config: configs/encnet/encnet_r50-d8_512x1024_40k_cityscapes.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/encnet/encnet_r50-d8_512x1024_40k_cityscapes/encnet_r50-d8_512x1024_40k_cityscapes_20200621_220958-68638a47.pth
- Config: configs/encnet/encnet_r101-d8_512x1024_40k_cityscapes.py
- Name: encnet_r101-d8_512x1024_40k_cityscapes
In Collection: encnet
Metadata:
backbone: R-101-D8
crop size: (512,1024)
lr schd: 40000
inference time (ms/im):
- backend: PyTorch
batch size: 1
- value: 375.94
hardware: V100
backend: PyTorch
batch size: 1
mode: FP32
resolution: (512,1024)
value: 375.94
lr schd: 40000
memory (GB): 12.1
Name: encnet_r101-d8_512x1024_40k_cityscapes
Results:
- Task: Semantic Segmentation
Dataset: Cityscapes
Metrics:
mIoU: 75.81
mIoU(ms+flip): 77.21
Task: Semantic Segmentation
Config: configs/encnet/encnet_r101-d8_512x1024_40k_cityscapes.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/encnet/encnet_r101-d8_512x1024_40k_cityscapes/encnet_r101-d8_512x1024_40k_cityscapes_20200621_220933-35e0a3e8.pth
- Config: configs/encnet/encnet_r50-d8_769x769_40k_cityscapes.py
- Name: encnet_r50-d8_769x769_40k_cityscapes
In Collection: encnet
Metadata:
backbone: R-50-D8
crop size: (769,769)
lr schd: 40000
inference time (ms/im):
- backend: PyTorch
batch size: 1
- value: 549.45
hardware: V100
backend: PyTorch
batch size: 1
mode: FP32
resolution: (769,769)
value: 549.45
lr schd: 40000
memory (GB): 9.8
Name: encnet_r50-d8_769x769_40k_cityscapes
Results:
- Task: Semantic Segmentation
Dataset: Cityscapes
Metrics:
mIoU: 76.24
mIoU(ms+flip): 77.85
Task: Semantic Segmentation
Config: configs/encnet/encnet_r50-d8_769x769_40k_cityscapes.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/encnet/encnet_r50-d8_769x769_40k_cityscapes/encnet_r50-d8_769x769_40k_cityscapes_20200621_220958-3bcd2884.pth
- Config: configs/encnet/encnet_r101-d8_769x769_40k_cityscapes.py
- Name: encnet_r101-d8_769x769_40k_cityscapes
In Collection: encnet
Metadata:
backbone: R-101-D8
crop size: (769,769)
lr schd: 40000
inference time (ms/im):
- backend: PyTorch
batch size: 1
- value: 793.65
hardware: V100
backend: PyTorch
batch size: 1
mode: FP32
resolution: (769,769)
value: 793.65
lr schd: 40000
memory (GB): 13.7
Name: encnet_r101-d8_769x769_40k_cityscapes
Results:
- Task: Semantic Segmentation
Dataset: Cityscapes
Metrics:
mIoU: 74.25
mIoU(ms+flip): 76.25
Task: Semantic Segmentation
Config: configs/encnet/encnet_r101-d8_769x769_40k_cityscapes.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/encnet/encnet_r101-d8_769x769_40k_cityscapes/encnet_r101-d8_769x769_40k_cityscapes_20200621_220933-2fafed55.pth
- Config: configs/encnet/encnet_r50-d8_512x1024_80k_cityscapes.py
- Name: encnet_r50-d8_512x1024_80k_cityscapes
In Collection: encnet
Metadata:
backbone: R-50-D8
crop size: (512,1024)
lr schd: 80000
Name: encnet_r50-d8_512x1024_80k_cityscapes
Results:
- Task: Semantic Segmentation
Dataset: Cityscapes
Metrics:
mIoU: 77.94
mIoU(ms+flip): 79.13
Task: Semantic Segmentation
Config: configs/encnet/encnet_r50-d8_512x1024_80k_cityscapes.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/encnet/encnet_r50-d8_512x1024_80k_cityscapes/encnet_r50-d8_512x1024_80k_cityscapes_20200622_003554-fc5c5624.pth
- Config: configs/encnet/encnet_r101-d8_512x1024_80k_cityscapes.py
- Name: encnet_r101-d8_512x1024_80k_cityscapes
In Collection: encnet
Metadata:
backbone: R-101-D8
crop size: (512,1024)
lr schd: 80000
Name: encnet_r101-d8_512x1024_80k_cityscapes
Results:
- Task: Semantic Segmentation
Dataset: Cityscapes
Metrics:
mIoU: 78.55
mIoU(ms+flip): 79.47
Task: Semantic Segmentation
Config: configs/encnet/encnet_r101-d8_512x1024_80k_cityscapes.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/encnet/encnet_r101-d8_512x1024_80k_cityscapes/encnet_r101-d8_512x1024_80k_cityscapes_20200622_003555-1de64bec.pth
- Config: configs/encnet/encnet_r50-d8_769x769_80k_cityscapes.py
- Name: encnet_r50-d8_769x769_80k_cityscapes
In Collection: encnet
Metadata:
backbone: R-50-D8
crop size: (769,769)
lr schd: 80000
Name: encnet_r50-d8_769x769_80k_cityscapes
Results:
- Task: Semantic Segmentation
Dataset: Cityscapes
Metrics:
mIoU: 77.44
mIoU(ms+flip): 78.72
Task: Semantic Segmentation
Config: configs/encnet/encnet_r50-d8_769x769_80k_cityscapes.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/encnet/encnet_r50-d8_769x769_80k_cityscapes/encnet_r50-d8_769x769_80k_cityscapes_20200622_003554-55096dcb.pth
- Config: configs/encnet/encnet_r101-d8_769x769_80k_cityscapes.py
- Name: encnet_r101-d8_769x769_80k_cityscapes
In Collection: encnet
Metadata:
backbone: R-101-D8
crop size: (769,769)
lr schd: 80000
Name: encnet_r101-d8_769x769_80k_cityscapes
Results:
- Task: Semantic Segmentation
Dataset: Cityscapes
Metrics:
mIoU: 76.1
mIoU(ms+flip): 76.97
Task: Semantic Segmentation
Config: configs/encnet/encnet_r101-d8_769x769_80k_cityscapes.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/encnet/encnet_r101-d8_769x769_80k_cityscapes/encnet_r101-d8_769x769_80k_cityscapes_20200622_003555-470ef79d.pth
- Config: configs/encnet/encnet_r50-d8_512x512_80k_ade20k.py
- Name: encnet_r50-d8_512x512_80k_ade20k
In Collection: encnet
Metadata:
backbone: R-50-D8
crop size: (512,512)
lr schd: 80000
inference time (ms/im):
- backend: PyTorch
batch size: 1
- value: 43.84
hardware: V100
backend: PyTorch
batch size: 1
mode: FP32
resolution: (512,512)
value: 43.84
lr schd: 80000
memory (GB): 10.1
Name: encnet_r50-d8_512x512_80k_ade20k
Results:
- Task: Semantic Segmentation
Dataset: ADE20K
Metrics:
mIoU: 39.53
mIoU(ms+flip): 41.17
Task: Semantic Segmentation
Config: configs/encnet/encnet_r50-d8_512x512_80k_ade20k.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/encnet/encnet_r50-d8_512x512_80k_ade20k/encnet_r50-d8_512x512_80k_ade20k_20200622_042412-44b46b04.pth
- Config: configs/encnet/encnet_r101-d8_512x512_80k_ade20k.py
- Name: encnet_r101-d8_512x512_80k_ade20k
In Collection: encnet
Metadata:
backbone: R-101-D8
crop size: (512,512)
lr schd: 80000
inference time (ms/im):
- backend: PyTorch
batch size: 1
- value: 67.25
hardware: V100
backend: PyTorch
batch size: 1
mode: FP32
resolution: (512,512)
value: 67.25
lr schd: 80000
memory (GB): 13.6
Name: encnet_r101-d8_512x512_80k_ade20k
Results:
- Task: Semantic Segmentation
Dataset: ADE20K
Metrics:
mIoU: 42.11
mIoU(ms+flip): 43.61
Task: Semantic Segmentation
Config: configs/encnet/encnet_r101-d8_512x512_80k_ade20k.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/encnet/encnet_r101-d8_512x512_80k_ade20k/encnet_r101-d8_512x512_80k_ade20k_20200622_101128-dd35e237.pth
- Config: configs/encnet/encnet_r50-d8_512x512_160k_ade20k.py
- Name: encnet_r50-d8_512x512_160k_ade20k
In Collection: encnet
Metadata:
backbone: R-50-D8
crop size: (512,512)
lr schd: 160000
Name: encnet_r50-d8_512x512_160k_ade20k
Results:
- Task: Semantic Segmentation
Dataset: ADE20K
Metrics:
mIoU: 40.1
mIoU(ms+flip): 41.71
Task: Semantic Segmentation
Config: configs/encnet/encnet_r50-d8_512x512_160k_ade20k.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/encnet/encnet_r50-d8_512x512_160k_ade20k/encnet_r50-d8_512x512_160k_ade20k_20200622_101059-b2db95e0.pth
- Config: configs/encnet/encnet_r101-d8_512x512_160k_ade20k.py
- Name: encnet_r101-d8_512x512_160k_ade20k
In Collection: encnet
Metadata:
backbone: R-101-D8
crop size: (512,512)
lr schd: 160000
Name: encnet_r101-d8_512x512_160k_ade20k
Results:
- Task: Semantic Segmentation
Dataset: ADE20K
Metrics:
mIoU: 42.61
mIoU(ms+flip): 44.01
Task: Semantic Segmentation
Config: configs/encnet/encnet_r101-d8_512x512_160k_ade20k.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/encnet/encnet_r101-d8_512x512_160k_ade20k/encnet_r101-d8_512x512_160k_ade20k_20200622_073348-7989641f.pth

9
configs/fastscnn/README.md
View File

@ -4,6 +4,13 @@
<!-- [ALGORITHM] -->
<a href="">Official Repo</a>
<a href="https://github.com/open-mmlab/mmsegmentation/blob/v0.17.0/mmseg/models/backbones/fast_scnn.py#L272">Code Snippet</a>
<details>
<summary align="right"><a href="https://arxiv.org/abs/1902.04502">Fast-SCNN (ArXiv'2019)</a></summary>
```latex
@article{poudel2019fast,
title={Fast-scnn: Fast semantic segmentation network},
@ -13,6 +20,8 @@
}
```
</details>
## Results and models
### Cityscapes

25
configs/fastscnn/fastscnn.yml
View File

@ -1,28 +1,35 @@
Collections:
- Metadata:
- Name: fastscnn
Metadata:
Training Data:
- Cityscapes
Name: fastscnn
Paper:
URL: https://arxiv.org/abs/1902.04502
Title: Fast-SCNN for Semantic Segmentation
README: configs/fastscnn/README.md
Code:
URL: https://github.com/open-mmlab/mmsegmentation/blob/v0.17.0/mmseg/models/backbones/fast_scnn.py#L272
Version: v0.17.0
Models:
- Config: configs/fastscnn/fast_scnn_lr0.12_8x4_160k_cityscapes.py
- Name: fast_scnn_lr0.12_8x4_160k_cityscapes
In Collection: fastscnn
Metadata:
backbone: Fast-SCNN
crop size: (512,1024)
lr schd: 160000
inference time (ms/im):
- backend: PyTorch
batch size: 1
- value: 17.71
hardware: V100
backend: PyTorch
batch size: 1
mode: FP32
resolution: (512,1024)
value: 17.71
lr schd: 160000
memory (GB): 3.3
Name: fast_scnn_lr0.12_8x4_160k_cityscapes
Results:
- Task: Semantic Segmentation
Dataset: Cityscapes
Metrics:
mIoU: 70.96
mIoU(ms+flip): 72.65
Task: Semantic Segmentation
Config: configs/fastscnn/fast_scnn_lr0.12_8x4_160k_cityscapes.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/fast_scnn/fast_scnn_lr0.12_8x4_160k_cityscapes/fast_scnn_lr0.12_8x4_160k_cityscapes_20210630_164853-0cec9937.pth

9
configs/fcn/README.md
View File

@ -4,6 +4,13 @@
<!-- [ALGORITHM] -->
<a href="https://github.com/BVLC/caffe/wiki/Model-Zoo#fcn">Official Repo</a>
<a href="https://github.com/open-mmlab/mmsegmentation/blob/v0.17.0/mmseg/models/decode_heads/fcn_head.py#L11">Code Snippet</a>
<details>
<summary align="right"><a href="https://arxiv.org/abs/1411.4038">FCN (CVPR'2015/TPAMI'2017)</a></summary>
```latex
@article{shelhamer2017fully,
title={Fully convolutional networks for semantic segmentation},
@ -17,6 +24,8 @@
}
```
</details>
## Results and models
### Cityscapes

485
configs/fcn/fcn.yml
View File

File diff suppressed because it is too large Load Diff

9
configs/fp16/README.md
View File

@ -4,6 +4,13 @@
<!-- [OTHERS] -->
<a href="https://github.com/baidu-research/DeepBench">Official Repo</a>
<a href="https://github.com/open-mmlab/mmcv/blob/v1.3.14/mmcv/runner/hooks/optimizer.py#L134">Code Snippet</a>
<details>
<summary align="right"><a href="https://arxiv.org/abs/1710.03740">Mixed Precision (FP16) Training (ArXiv'2017)</a></summary>
```latex
@article{micikevicius2017mixed,
title={Mixed precision training},
@ -13,6 +20,8 @@
}
```
</details>
## Results and models
### Cityscapes

69
configs/fp16/fp16.yml
View File

@ -1,90 +1,99 @@
Collections:
- Metadata:
- Name: fp16
Metadata:
Training Data:
- Cityscapes
Name: fp16
Paper:
URL: https://arxiv.org/abs/1710.03740
Title: Mixed Precision Training
README: configs/fp16/README.md
Code:
URL: https://github.com/open-mmlab/mmcv/blob/v1.3.14/mmcv/runner/hooks/optimizer.py#L134
Version: v1.3.14
Converted From:
Code: https://github.com/baidu-research/DeepBench
Models:
- Config: configs/fp16/fcn_r101-d8_512x1024_80k_fp16_cityscapes.py
- Name: fcn_r101-d8_512x1024_80k_fp16_cityscapes
In Collection: fp16
Metadata:
backbone: R-101-D8
crop size: (512,1024)
lr schd: 80000
inference time (ms/im):
- backend: PyTorch
batch size: 1
- value: 115.74
hardware: V100
backend: PyTorch
batch size: 1
mode: FP32
resolution: (512,1024)
value: 115.74
lr schd: 80000
memory (GB): 5.37
Name: fcn_r101-d8_512x1024_80k_fp16_cityscapes
Results:
- Task: Semantic Segmentation
Dataset: Cityscapes
Metrics:
mIoU: 76.8
Task: Semantic Segmentation
Config: configs/fp16/fcn_r101-d8_512x1024_80k_fp16_cityscapes.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/fp16/fcn_r101-d8_512x1024_80k_fp16_cityscapes/fcn_r101-d8_512x1024_80k_fp16_cityscapes_20200717_230921-50245227.pth
- Config: configs/fp16/pspnet_r101-d8_512x1024_80k_fp16_cityscapes.py
- Name: pspnet_r101-d8_512x1024_80k_fp16_cityscapes
In Collection: fp16
Metadata:
backbone: R-101-D8
crop size: (512,1024)
lr schd: 80000
inference time (ms/im):
- backend: PyTorch
batch size: 1
- value: 114.03
hardware: V100
backend: PyTorch
batch size: 1
mode: FP32
resolution: (512,1024)
value: 114.03
lr schd: 80000
memory (GB): 5.34
Name: pspnet_r101-d8_512x1024_80k_fp16_cityscapes
Results:
- Task: Semantic Segmentation
Dataset: Cityscapes
Metrics:
mIoU: 79.46
Task: Semantic Segmentation
Config: configs/fp16/pspnet_r101-d8_512x1024_80k_fp16_cityscapes.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/fp16/pspnet_r101-d8_512x1024_80k_fp16_cityscapes/pspnet_r101-d8_512x1024_80k_fp16_cityscapes_20200717_230919-ade37931.pth
- Config: configs/fp16/deeplabv3_r101-d8_512x1024_80k_fp16_cityscapes.py
- Name: deeplabv3_r101-d8_512x1024_80k_fp16_cityscapes
In Collection: fp16
Metadata:
backbone: R-101-D8
crop size: (512,1024)
lr schd: 80000
inference time (ms/im):
- backend: PyTorch
batch size: 1
- value: 259.07
hardware: V100
backend: PyTorch
batch size: 1
mode: FP32
resolution: (512,1024)
value: 259.07
lr schd: 80000
memory (GB): 5.75
Name: deeplabv3_r101-d8_512x1024_80k_fp16_cityscapes
Results:
- Task: Semantic Segmentation
Dataset: Cityscapes
Metrics:
mIoU: 80.48
Task: Semantic Segmentation
Config: configs/fp16/deeplabv3_r101-d8_512x1024_80k_fp16_cityscapes.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/fp16/deeplabv3_r101-d8_512x1024_80k_fp16_cityscapes/deeplabv3_r101-d8_512x1024_80k_fp16_cityscapes_20200717_230920-bc86dc84.pth
- Config: configs/fp16/deeplabv3plus_r101-d8_512x1024_80k_fp16_cityscapes.py
- Name: deeplabv3plus_r101-d8_512x1024_80k_fp16_cityscapes
In Collection: fp16
Metadata:
backbone: R-101-D8
crop size: (512,1024)
lr schd: 80000
inference time (ms/im):
- backend: PyTorch
batch size: 1
- value: 127.06
hardware: V100
backend: PyTorch
batch size: 1
mode: FP32
resolution: (512,1024)
value: 127.06
lr schd: 80000
memory (GB): 6.35
Name: deeplabv3plus_r101-d8_512x1024_80k_fp16_cityscapes
Results:
- Task: Semantic Segmentation
Dataset: Cityscapes
Metrics:
mIoU: 80.46
Task: Semantic Segmentation
Config: configs/fp16/deeplabv3plus_r101-d8_512x1024_80k_fp16_cityscapes.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/fp16/deeplabv3plus_r101-d8_512x1024_80k_fp16_cityscapes/deeplabv3plus_r101-d8_512x1024_80k_fp16_cityscapes_20200717_230920-cc58bc8d.pth

9
configs/gcnet/README.md
View File

@ -4,6 +4,13 @@
<!-- [ALGORITHM] -->
<a href="https://github.com/xvjiarui/GCNet">Official Repo</a>
<a href="https://github.com/open-mmlab/mmsegmentation/blob/v0.17.0/mmseg/models/decode_heads/gc_head.py#L10">Code Snippet</a>
<details>
<summary align="right"><a href="https://arxiv.org/abs/1904.11492">GCNet (ICCVW'2019/TPAMI'2020)</a></summary>
```latex
@inproceedings{cao2019gcnet,
title={Gcnet: Non-local networks meet squeeze-excitation networks and beyond},
@ -14,6 +21,8 @@
}
```
</details>
## Results and models
### Cityscapes

173
configs/gcnet/gcnet.yml
View File

@ -1,296 +1,305 @@
Collections:
- Metadata:
- Name: gcnet
Metadata:
Training Data:
- Cityscapes
- ADE20K
- Pascal VOC 2012 + Aug
Name: gcnet
Paper:
URL: https://arxiv.org/abs/1904.11492
Title: 'GCNet: Non-local Networks Meet Squeeze-Excitation Networks and Beyond'
README: configs/gcnet/README.md
Code:
URL: https://github.com/open-mmlab/mmsegmentation/blob/v0.17.0/mmseg/models/decode_heads/gc_head.py#L10
Version: v0.17.0
Converted From:
Code: https://github.com/xvjiarui/GCNet
Models:
- Config: configs/gcnet/gcnet_r50-d8_512x1024_40k_cityscapes.py
- Name: gcnet_r50-d8_512x1024_40k_cityscapes
In Collection: gcnet
Metadata:
backbone: R-50-D8
crop size: (512,1024)
lr schd: 40000
inference time (ms/im):
- backend: PyTorch
batch size: 1
- value: 254.45
hardware: V100
backend: PyTorch
batch size: 1
mode: FP32
resolution: (512,1024)
value: 254.45
lr schd: 40000
memory (GB): 5.8
Name: gcnet_r50-d8_512x1024_40k_cityscapes
Results:
- Task: Semantic Segmentation
Dataset: Cityscapes
Metrics:
mIoU: 77.69
mIoU(ms+flip): 78.56
Task: Semantic Segmentation
Config: configs/gcnet/gcnet_r50-d8_512x1024_40k_cityscapes.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/gcnet/gcnet_r50-d8_512x1024_40k_cityscapes/gcnet_r50-d8_512x1024_40k_cityscapes_20200618_074436-4b0fd17b.pth
- Config: configs/gcnet/gcnet_r101-d8_512x1024_40k_cityscapes.py
- Name: gcnet_r101-d8_512x1024_40k_cityscapes
In Collection: gcnet
Metadata:
backbone: R-101-D8
crop size: (512,1024)
lr schd: 40000
inference time (ms/im):
- backend: PyTorch
batch size: 1
- value: 383.14
hardware: V100
backend: PyTorch
batch size: 1
mode: FP32
resolution: (512,1024)
value: 383.14
lr schd: 40000
memory (GB): 9.2
Name: gcnet_r101-d8_512x1024_40k_cityscapes
Results:
- Task: Semantic Segmentation
Dataset: Cityscapes
Metrics:
mIoU: 78.28
mIoU(ms+flip): 79.34
Task: Semantic Segmentation
Config: configs/gcnet/gcnet_r101-d8_512x1024_40k_cityscapes.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/gcnet/gcnet_r101-d8_512x1024_40k_cityscapes/gcnet_r101-d8_512x1024_40k_cityscapes_20200618_074436-5e62567f.pth
- Config: configs/gcnet/gcnet_r50-d8_769x769_40k_cityscapes.py
- Name: gcnet_r50-d8_769x769_40k_cityscapes
In Collection: gcnet
Metadata:
backbone: R-50-D8
crop size: (769,769)
lr schd: 40000
inference time (ms/im):
- backend: PyTorch
batch size: 1
- value: 598.8
hardware: V100
backend: PyTorch
batch size: 1
mode: FP32
resolution: (769,769)
value: 598.8
lr schd: 40000
memory (GB): 6.5
Name: gcnet_r50-d8_769x769_40k_cityscapes
Results:
- Task: Semantic Segmentation
Dataset: Cityscapes
Metrics:
mIoU: 78.12
mIoU(ms+flip): 80.09
Task: Semantic Segmentation
Config: configs/gcnet/gcnet_r50-d8_769x769_40k_cityscapes.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/gcnet/gcnet_r50-d8_769x769_40k_cityscapes/gcnet_r50-d8_769x769_40k_cityscapes_20200618_182814-a26f4471.pth
- Config: configs/gcnet/gcnet_r101-d8_769x769_40k_cityscapes.py
- Name: gcnet_r101-d8_769x769_40k_cityscapes
In Collection: gcnet
Metadata:
backbone: R-101-D8
crop size: (769,769)
lr schd: 40000
inference time (ms/im):
- backend: PyTorch
batch size: 1
- value: 884.96
hardware: V100
backend: PyTorch
batch size: 1
mode: FP32
resolution: (769,769)
value: 884.96
lr schd: 40000
memory (GB): 10.5
Name: gcnet_r101-d8_769x769_40k_cityscapes
Results:
- Task: Semantic Segmentation
Dataset: Cityscapes
Metrics:
mIoU: 78.95
mIoU(ms+flip): 80.71
Task: Semantic Segmentation
Config: configs/gcnet/gcnet_r101-d8_769x769_40k_cityscapes.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/gcnet/gcnet_r101-d8_769x769_40k_cityscapes/gcnet_r101-d8_769x769_40k_cityscapes_20200619_092550-ca4f0a84.pth
- Config: configs/gcnet/gcnet_r50-d8_512x1024_80k_cityscapes.py
- Name: gcnet_r50-d8_512x1024_80k_cityscapes
In Collection: gcnet
Metadata:
backbone: R-50-D8
crop size: (512,1024)
lr schd: 80000
Name: gcnet_r50-d8_512x1024_80k_cityscapes
Results:
- Task: Semantic Segmentation
Dataset: Cityscapes
Metrics:
mIoU: 78.48
mIoU(ms+flip): 80.01
Task: Semantic Segmentation
Config: configs/gcnet/gcnet_r50-d8_512x1024_80k_cityscapes.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/gcnet/gcnet_r50-d8_512x1024_80k_cityscapes/gcnet_r50-d8_512x1024_80k_cityscapes_20200618_074450-ef8f069b.pth
- Config: configs/gcnet/gcnet_r101-d8_512x1024_80k_cityscapes.py
- Name: gcnet_r101-d8_512x1024_80k_cityscapes
In Collection: gcnet
Metadata:
backbone: R-101-D8
crop size: (512,1024)
lr schd: 80000
Name: gcnet_r101-d8_512x1024_80k_cityscapes
Results:
- Task: Semantic Segmentation
Dataset: Cityscapes
Metrics:
mIoU: 79.03
mIoU(ms+flip): 79.84
Task: Semantic Segmentation
Config: configs/gcnet/gcnet_r101-d8_512x1024_80k_cityscapes.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/gcnet/gcnet_r101-d8_512x1024_80k_cityscapes/gcnet_r101-d8_512x1024_80k_cityscapes_20200618_074450-778ebf69.pth
- Config: configs/gcnet/gcnet_r50-d8_769x769_80k_cityscapes.py
- Name: gcnet_r50-d8_769x769_80k_cityscapes
In Collection: gcnet
Metadata:
backbone: R-50-D8
crop size: (769,769)
lr schd: 80000
Name: gcnet_r50-d8_769x769_80k_cityscapes
Results:
- Task: Semantic Segmentation
Dataset: Cityscapes
Metrics:
mIoU: 78.68
mIoU(ms+flip): 80.66
Task: Semantic Segmentation
Config: configs/gcnet/gcnet_r50-d8_769x769_80k_cityscapes.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/gcnet/gcnet_r50-d8_769x769_80k_cityscapes/gcnet_r50-d8_769x769_80k_cityscapes_20200619_092516-4839565b.pth
- Config: configs/gcnet/gcnet_r101-d8_769x769_80k_cityscapes.py
- Name: gcnet_r101-d8_769x769_80k_cityscapes
In Collection: gcnet
Metadata:
backbone: R-101-D8
crop size: (769,769)
lr schd: 80000
Name: gcnet_r101-d8_769x769_80k_cityscapes
Results:
- Task: Semantic Segmentation
Dataset: Cityscapes
Metrics:
mIoU: 79.18
mIoU(ms+flip): 80.71
Task: Semantic Segmentation
Config: configs/gcnet/gcnet_r101-d8_769x769_80k_cityscapes.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/gcnet/gcnet_r101-d8_769x769_80k_cityscapes/gcnet_r101-d8_769x769_80k_cityscapes_20200619_092628-8e043423.pth
- Config: configs/gcnet/gcnet_r50-d8_512x512_80k_ade20k.py
- Name: gcnet_r50-d8_512x512_80k_ade20k
In Collection: gcnet
Metadata:
backbone: R-50-D8
crop size: (512,512)
lr schd: 80000
inference time (ms/im):
- backend: PyTorch
batch size: 1
- value: 42.77
hardware: V100
backend: PyTorch
batch size: 1
mode: FP32
resolution: (512,512)
value: 42.77
lr schd: 80000
memory (GB): 8.5
Name: gcnet_r50-d8_512x512_80k_ade20k
Results:
- Task: Semantic Segmentation
Dataset: ADE20K
Metrics:
mIoU: 41.47
mIoU(ms+flip): 42.85
Task: Semantic Segmentation
Config: configs/gcnet/gcnet_r50-d8_512x512_80k_ade20k.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/gcnet/gcnet_r50-d8_512x512_80k_ade20k/gcnet_r50-d8_512x512_80k_ade20k_20200614_185146-91a6da41.pth
- Config: configs/gcnet/gcnet_r101-d8_512x512_80k_ade20k.py
- Name: gcnet_r101-d8_512x512_80k_ade20k
In Collection: gcnet
Metadata:
backbone: R-101-D8
crop size: (512,512)
lr schd: 80000
inference time (ms/im):
- backend: PyTorch
batch size: 1
- value: 65.79
hardware: V100
backend: PyTorch
batch size: 1
mode: FP32
resolution: (512,512)
value: 65.79
lr schd: 80000
memory (GB): 12.0
Name: gcnet_r101-d8_512x512_80k_ade20k
Results:
- Task: Semantic Segmentation
Dataset: ADE20K
Metrics:
mIoU: 42.82
mIoU(ms+flip): 44.54
Task: Semantic Segmentation
Config: configs/gcnet/gcnet_r101-d8_512x512_80k_ade20k.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/gcnet/gcnet_r101-d8_512x512_80k_ade20k/gcnet_r101-d8_512x512_80k_ade20k_20200615_020811-c3fcb6dd.pth
- Config: configs/gcnet/gcnet_r50-d8_512x512_160k_ade20k.py
- Name: gcnet_r50-d8_512x512_160k_ade20k
In Collection: gcnet
Metadata:
backbone: R-50-D8
crop size: (512,512)
lr schd: 160000
Name: gcnet_r50-d8_512x512_160k_ade20k
Results:
- Task: Semantic Segmentation
Dataset: ADE20K
Metrics:
mIoU: 42.37
mIoU(ms+flip): 43.52
Task: Semantic Segmentation
Config: configs/gcnet/gcnet_r50-d8_512x512_160k_ade20k.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/gcnet/gcnet_r50-d8_512x512_160k_ade20k/gcnet_r50-d8_512x512_160k_ade20k_20200615_224122-d95f3e1f.pth
- Config: configs/gcnet/gcnet_r101-d8_512x512_160k_ade20k.py
- Name: gcnet_r101-d8_512x512_160k_ade20k
In Collection: gcnet
Metadata:
backbone: R-101-D8
crop size: (512,512)
lr schd: 160000
Name: gcnet_r101-d8_512x512_160k_ade20k
Results:
- Task: Semantic Segmentation
Dataset: ADE20K
Metrics:
mIoU: 43.69
mIoU(ms+flip): 45.21
Task: Semantic Segmentation
Config: configs/gcnet/gcnet_r101-d8_512x512_160k_ade20k.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/gcnet/gcnet_r101-d8_512x512_160k_ade20k/gcnet_r101-d8_512x512_160k_ade20k_20200615_225406-615528d7.pth
- Config: configs/gcnet/gcnet_r50-d8_512x512_20k_voc12aug.py
- Name: gcnet_r50-d8_512x512_20k_voc12aug
In Collection: gcnet
Metadata:
backbone: R-50-D8
crop size: (512,512)
lr schd: 20000
inference time (ms/im):
- backend: PyTorch
batch size: 1
- value: 42.83
hardware: V100
backend: PyTorch
batch size: 1
mode: FP32
resolution: (512,512)
value: 42.83
lr schd: 20000
memory (GB): 5.8
Name: gcnet_r50-d8_512x512_20k_voc12aug
Results:
- Task: Semantic Segmentation
Dataset: Pascal VOC 2012 + Aug
Metrics:
mIoU: 76.42
mIoU(ms+flip): 77.51
Task: Semantic Segmentation
Config: configs/gcnet/gcnet_r50-d8_512x512_20k_voc12aug.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/gcnet/gcnet_r50-d8_512x512_20k_voc12aug/gcnet_r50-d8_512x512_20k_voc12aug_20200617_165701-3cbfdab1.pth
- Config: configs/gcnet/gcnet_r101-d8_512x512_20k_voc12aug.py
- Name: gcnet_r101-d8_512x512_20k_voc12aug
In Collection: gcnet
Metadata:
backbone: R-101-D8
crop size: (512,512)
lr schd: 20000
inference time (ms/im):
- backend: PyTorch
batch size: 1
- value: 67.57
hardware: V100
backend: PyTorch
batch size: 1
mode: FP32
resolution: (512,512)
value: 67.57
lr schd: 20000
memory (GB): 9.2
Name: gcnet_r101-d8_512x512_20k_voc12aug
Results:
- Task: Semantic Segmentation
Dataset: Pascal VOC 2012 + Aug
Metrics:
mIoU: 77.41
mIoU(ms+flip): 78.56
Task: Semantic Segmentation
Config: configs/gcnet/gcnet_r101-d8_512x512_20k_voc12aug.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/gcnet/gcnet_r101-d8_512x512_20k_voc12aug/gcnet_r101-d8_512x512_20k_voc12aug_20200617_165713-6c720aa9.pth
- Config: configs/gcnet/gcnet_r50-d8_512x512_40k_voc12aug.py
- Name: gcnet_r50-d8_512x512_40k_voc12aug
In Collection: gcnet
Metadata:
backbone: R-50-D8
crop size: (512,512)
lr schd: 40000
Name: gcnet_r50-d8_512x512_40k_voc12aug
Results:
- Task: Semantic Segmentation
Dataset: Pascal VOC 2012 + Aug
Metrics:
mIoU: 76.24
mIoU(ms+flip): 77.63
Task: Semantic Segmentation
Config: configs/gcnet/gcnet_r50-d8_512x512_40k_voc12aug.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/gcnet/gcnet_r50-d8_512x512_40k_voc12aug/gcnet_r50-d8_512x512_40k_voc12aug_20200613_195105-9797336d.pth
- Config: configs/gcnet/gcnet_r101-d8_512x512_40k_voc12aug.py
- Name: gcnet_r101-d8_512x512_40k_voc12aug
In Collection: gcnet
Metadata:
backbone: R-101-D8
crop size: (512,512)
lr schd: 40000
Name: gcnet_r101-d8_512x512_40k_voc12aug
Results:
- Task: Semantic Segmentation
Dataset: Pascal VOC 2012 + Aug
Metrics:
mIoU: 77.84
mIoU(ms+flip): 78.59
Task: Semantic Segmentation
Config: configs/gcnet/gcnet_r101-d8_512x512_40k_voc12aug.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/gcnet/gcnet_r101-d8_512x512_40k_voc12aug/gcnet_r101-d8_512x512_40k_voc12aug_20200613_185806-1e38208d.pth

9
configs/hrnet/README.md
View File

@ -4,6 +4,13 @@
<!-- [ALGORITHM] -->
<a href="https://github.com/HRNet/HRNet-Semantic-Segmentation">Official Repo</a>
<a href="https://github.com/open-mmlab/mmsegmentation/blob/v0.17.0/mmseg/models/backbones/hrnet.py#L218">Code Snippet</a>
<details>
<summary align="right"><a href="https://arxiv.org/abs/1908.07919">HRNet (CVPR'2019)</a></summary>
```latext
@inproceedings{SunXLW19,
title={Deep High-Resolution Representation Learning for Human Pose Estimation},
@ -13,6 +20,8 @@
}
```
</details>
## Results and models
### Cityscapes

243
configs/hrnet/hrnet.yml
View File

@ -1,440 +1,449 @@
Collections:
- Metadata:
- Name: hrnet
Metadata:
Training Data:
- Cityscapes
- ADE20K
- Pascal VOC 2012 + Aug
- Pascal Context
- Pascal Context 59
Name: hrnet
Paper:
URL: https://arxiv.org/abs/1908.07919
Title: Deep High-Resolution Representation Learning for Human Pose Estimation
README: configs/hrnet/README.md
Code:
URL: https://github.com/open-mmlab/mmsegmentation/blob/v0.17.0/mmseg/models/backbones/hrnet.py#L218
Version: v0.17.0
Converted From:
Code: https://github.com/HRNet/HRNet-Semantic-Segmentation
Models:
- Config: configs/hrnet/fcn_hr18s_512x1024_40k_cityscapes.py
- Name: fcn_hr18s_512x1024_40k_cityscapes
In Collection: hrnet
Metadata:
backbone: HRNetV2p-W18-Small
crop size: (512,1024)
lr schd: 40000
inference time (ms/im):
- backend: PyTorch
batch size: 1
- value: 42.12
hardware: V100
backend: PyTorch
batch size: 1
mode: FP32
resolution: (512,1024)
value: 42.12
lr schd: 40000
memory (GB): 1.7
Name: fcn_hr18s_512x1024_40k_cityscapes
Results:
- Task: Semantic Segmentation
Dataset: Cityscapes
Metrics:
mIoU: 73.86
mIoU(ms+flip): 75.91
Task: Semantic Segmentation
Config: configs/hrnet/fcn_hr18s_512x1024_40k_cityscapes.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/hrnet/fcn_hr18s_512x1024_40k_cityscapes/fcn_hr18s_512x1024_40k_cityscapes_20200601_014216-93db27d0.pth
- Config: configs/hrnet/fcn_hr18_512x1024_40k_cityscapes.py
- Name: fcn_hr18_512x1024_40k_cityscapes
In Collection: hrnet
Metadata:
backbone: HRNetV2p-W18
crop size: (512,1024)
lr schd: 40000
inference time (ms/im):
- backend: PyTorch
batch size: 1
- value: 77.1
hardware: V100
backend: PyTorch
batch size: 1
mode: FP32
resolution: (512,1024)
value: 77.1
lr schd: 40000
memory (GB): 2.9
Name: fcn_hr18_512x1024_40k_cityscapes
Results:
- Task: Semantic Segmentation
Dataset: Cityscapes
Metrics:
mIoU: 77.19
mIoU(ms+flip): 78.92
Task: Semantic Segmentation
Config: configs/hrnet/fcn_hr18_512x1024_40k_cityscapes.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/hrnet/fcn_hr18_512x1024_40k_cityscapes/fcn_hr18_512x1024_40k_cityscapes_20200601_014216-f196fb4e.pth
- Config: configs/hrnet/fcn_hr48_512x1024_40k_cityscapes.py
- Name: fcn_hr48_512x1024_40k_cityscapes
In Collection: hrnet
Metadata:
backbone: HRNetV2p-W48
crop size: (512,1024)
lr schd: 40000
inference time (ms/im):
- backend: PyTorch
batch size: 1
- value: 155.76
hardware: V100
backend: PyTorch
batch size: 1
mode: FP32
resolution: (512,1024)
value: 155.76
lr schd: 40000
memory (GB): 6.2
Name: fcn_hr48_512x1024_40k_cityscapes
Results:
- Task: Semantic Segmentation
Dataset: Cityscapes
Metrics:
mIoU: 78.48
mIoU(ms+flip): 79.69
Task: Semantic Segmentation
Config: configs/hrnet/fcn_hr48_512x1024_40k_cityscapes.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/hrnet/fcn_hr48_512x1024_40k_cityscapes/fcn_hr48_512x1024_40k_cityscapes_20200601_014240-a989b146.pth
- Config: configs/hrnet/fcn_hr18s_512x1024_80k_cityscapes.py
- Name: fcn_hr18s_512x1024_80k_cityscapes
In Collection: hrnet
Metadata:
backbone: HRNetV2p-W18-Small
crop size: (512,1024)
lr schd: 80000
Name: fcn_hr18s_512x1024_80k_cityscapes
Results:
- Task: Semantic Segmentation
Dataset: Cityscapes
Metrics:
mIoU: 75.31
mIoU(ms+flip): 77.48
Task: Semantic Segmentation
Config: configs/hrnet/fcn_hr18s_512x1024_80k_cityscapes.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/hrnet/fcn_hr18s_512x1024_80k_cityscapes/fcn_hr18s_512x1024_80k_cityscapes_20200601_202700-1462b75d.pth
- Config: configs/hrnet/fcn_hr18_512x1024_80k_cityscapes.py
- Name: fcn_hr18_512x1024_80k_cityscapes
In Collection: hrnet
Metadata:
backbone: HRNetV2p-W18
crop size: (512,1024)
lr schd: 80000
Name: fcn_hr18_512x1024_80k_cityscapes
Results:
- Task: Semantic Segmentation
Dataset: Cityscapes
Metrics:
mIoU: 78.65
mIoU(ms+flip): 80.35
Task: Semantic Segmentation
Config: configs/hrnet/fcn_hr18_512x1024_80k_cityscapes.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/hrnet/fcn_hr18_512x1024_80k_cityscapes/fcn_hr18_512x1024_80k_cityscapes_20200601_223255-4e7b345e.pth
- Config: configs/hrnet/fcn_hr48_512x1024_80k_cityscapes.py
- Name: fcn_hr48_512x1024_80k_cityscapes
In Collection: hrnet
Metadata:
backbone: HRNetV2p-W48
crop size: (512,1024)
lr schd: 80000
Name: fcn_hr48_512x1024_80k_cityscapes
Results:
- Task: Semantic Segmentation
Dataset: Cityscapes
Metrics:
mIoU: 79.93
mIoU(ms+flip): 80.72
Task: Semantic Segmentation
Config: configs/hrnet/fcn_hr48_512x1024_80k_cityscapes.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/hrnet/fcn_hr48_512x1024_80k_cityscapes/fcn_hr48_512x1024_80k_cityscapes_20200601_202606-58ea95d6.pth
- Config: configs/hrnet/fcn_hr18s_512x1024_160k_cityscapes.py
- Name: fcn_hr18s_512x1024_160k_cityscapes
In Collection: hrnet
Metadata:
backbone: HRNetV2p-W18-Small
crop size: (512,1024)
lr schd: 160000
Name: fcn_hr18s_512x1024_160k_cityscapes
Results:
- Task: Semantic Segmentation
Dataset: Cityscapes
Metrics:
mIoU: 76.31
mIoU(ms+flip): 78.31
Task: Semantic Segmentation
Config: configs/hrnet/fcn_hr18s_512x1024_160k_cityscapes.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/hrnet/fcn_hr18s_512x1024_160k_cityscapes/fcn_hr18s_512x1024_160k_cityscapes_20200602_190901-4a0797ea.pth
- Config: configs/hrnet/fcn_hr18_512x1024_160k_cityscapes.py
- Name: fcn_hr18_512x1024_160k_cityscapes
In Collection: hrnet
Metadata:
backbone: HRNetV2p-W18
crop size: (512,1024)
lr schd: 160000
Name: fcn_hr18_512x1024_160k_cityscapes
Results:
- Task: Semantic Segmentation
Dataset: Cityscapes
Metrics:
mIoU: 78.8
mIoU(ms+flip): 80.74
Task: Semantic Segmentation
Config: configs/hrnet/fcn_hr18_512x1024_160k_cityscapes.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/hrnet/fcn_hr18_512x1024_160k_cityscapes/fcn_hr18_512x1024_160k_cityscapes_20200602_190822-221e4a4f.pth
- Config: configs/hrnet/fcn_hr48_512x1024_160k_cityscapes.py
- Name: fcn_hr48_512x1024_160k_cityscapes
In Collection: hrnet
Metadata:
backbone: HRNetV2p-W48
crop size: (512,1024)
lr schd: 160000
Name: fcn_hr48_512x1024_160k_cityscapes
Results:
- Task: Semantic Segmentation
Dataset: Cityscapes
Metrics:
mIoU: 80.65
mIoU(ms+flip): 81.92
Task: Semantic Segmentation
Config: configs/hrnet/fcn_hr48_512x1024_160k_cityscapes.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/hrnet/fcn_hr48_512x1024_160k_cityscapes/fcn_hr48_512x1024_160k_cityscapes_20200602_190946-59b7973e.pth
- Config: configs/hrnet/fcn_hr18s_512x512_80k_ade20k.py
- Name: fcn_hr18s_512x512_80k_ade20k
In Collection: hrnet
Metadata:
backbone: HRNetV2p-W18-Small
crop size: (512,512)
lr schd: 80000
inference time (ms/im):
- backend: PyTorch
batch size: 1
- value: 25.87
hardware: V100
backend: PyTorch
batch size: 1
mode: FP32
resolution: (512,512)
value: 25.87
lr schd: 80000
memory (GB): 3.8
Name: fcn_hr18s_512x512_80k_ade20k
Results:
- Task: Semantic Segmentation
Dataset: ADE20K
Metrics:
mIoU: 31.38
mIoU(ms+flip): 32.45
Task: Semantic Segmentation
Config: configs/hrnet/fcn_hr18s_512x512_80k_ade20k.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/hrnet/fcn_hr18s_512x512_80k_ade20k/fcn_hr18s_512x512_80k_ade20k_20200614_144345-77fc814a.pth
- Config: configs/hrnet/fcn_hr18_512x512_80k_ade20k.py
- Name: fcn_hr18_512x512_80k_ade20k
In Collection: hrnet
Metadata:
backbone: HRNetV2p-W18
crop size: (512,512)
lr schd: 80000
inference time (ms/im):
- backend: PyTorch
batch size: 1
- value: 44.31
hardware: V100
backend: PyTorch
batch size: 1
mode: FP32
resolution: (512,512)
value: 44.31
lr schd: 80000
memory (GB): 4.9
Name: fcn_hr18_512x512_80k_ade20k
Results:
- Task: Semantic Segmentation
Dataset: ADE20K
Metrics:
mIoU: 36.27
mIoU(ms+flip): 37.28
Task: Semantic Segmentation
Config: configs/hrnet/fcn_hr18_512x512_80k_ade20k.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/hrnet/fcn_hr18_512x512_80k_ade20k/fcn_hr18_512x512_80k_ade20k_20210827_114910-6c9382c0.pth
- Config: configs/hrnet/fcn_hr48_512x512_80k_ade20k.py
- Name: fcn_hr48_512x512_80k_ade20k
In Collection: hrnet
Metadata:
backbone: HRNetV2p-W48
crop size: (512,512)
lr schd: 80000
inference time (ms/im):
- backend: PyTorch
batch size: 1
- value: 47.1
hardware: V100
backend: PyTorch
batch size: 1
mode: FP32
resolution: (512,512)
value: 47.1
lr schd: 80000
memory (GB): 8.2
Name: fcn_hr48_512x512_80k_ade20k
Results:
- Task: Semantic Segmentation
Dataset: ADE20K
Metrics:
mIoU: 41.9
mIoU(ms+flip): 43.27
Task: Semantic Segmentation
Config: configs/hrnet/fcn_hr48_512x512_80k_ade20k.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/hrnet/fcn_hr48_512x512_80k_ade20k/fcn_hr48_512x512_80k_ade20k_20200614_193946-7ba5258d.pth
- Config: configs/hrnet/fcn_hr18s_512x512_160k_ade20k.py
- Name: fcn_hr18s_512x512_160k_ade20k
In Collection: hrnet
Metadata:
backbone: HRNetV2p-W18-Small
crop size: (512,512)
lr schd: 160000
Name: fcn_hr18s_512x512_160k_ade20k
Results:
- Task: Semantic Segmentation
Dataset: ADE20K
Metrics:
mIoU: 33.07
mIoU(ms+flip): 34.56
Task: Semantic Segmentation
Config: configs/hrnet/fcn_hr18s_512x512_160k_ade20k.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/hrnet/fcn_hr18s_512x512_160k_ade20k/fcn_hr18s_512x512_160k_ade20k_20210829_174739-f1e7c2e7.pth
- Config: configs/hrnet/fcn_hr18_512x512_160k_ade20k.py
- Name: fcn_hr18_512x512_160k_ade20k
In Collection: hrnet
Metadata:
backbone: HRNetV2p-W18
crop size: (512,512)
lr schd: 160000
Name: fcn_hr18_512x512_160k_ade20k
Results:
- Task: Semantic Segmentation
Dataset: ADE20K
Metrics:
mIoU: 36.79
mIoU(ms+flip): 38.58
Task: Semantic Segmentation
Config: configs/hrnet/fcn_hr18_512x512_160k_ade20k.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/hrnet/fcn_hr18_512x512_160k_ade20k/fcn_hr18_512x512_160k_ade20k_20200614_214426-ca961836.pth
- Config: configs/hrnet/fcn_hr48_512x512_160k_ade20k.py
- Name: fcn_hr48_512x512_160k_ade20k
In Collection: hrnet
Metadata:
backbone: HRNetV2p-W48
crop size: (512,512)
lr schd: 160000
Name: fcn_hr48_512x512_160k_ade20k
Results:
- Task: Semantic Segmentation
Dataset: ADE20K
Metrics:
mIoU: 42.02
mIoU(ms+flip): 43.86
Task: Semantic Segmentation
Config: configs/hrnet/fcn_hr48_512x512_160k_ade20k.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/hrnet/fcn_hr48_512x512_160k_ade20k/fcn_hr48_512x512_160k_ade20k_20200614_214407-a52fc02c.pth
- Config: configs/hrnet/fcn_hr18s_512x512_20k_voc12aug.py
- Name: fcn_hr18s_512x512_20k_voc12aug
In Collection: hrnet
Metadata:
backbone: HRNetV2p-W18-Small
crop size: (512,512)
lr schd: 20000
inference time (ms/im):
- backend: PyTorch
batch size: 1
- value: 23.06
hardware: V100
backend: PyTorch
batch size: 1
mode: FP32
resolution: (512,512)
value: 23.06
lr schd: 20000
memory (GB): 1.8
Name: fcn_hr18s_512x512_20k_voc12aug
Results:
- Task: Semantic Segmentation
Dataset: Pascal VOC 2012 + Aug
Metrics:
mIoU: 65.5
mIoU(ms+flip): 68.89
Task: Semantic Segmentation
Config: configs/hrnet/fcn_hr18s_512x512_20k_voc12aug.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/hrnet/fcn_hr18s_512x512_20k_voc12aug/fcn_hr18s_512x512_20k_voc12aug_20210829_174910-0aceadb4.pth
- Config: configs/hrnet/fcn_hr18_512x512_20k_voc12aug.py
- Name: fcn_hr18_512x512_20k_voc12aug
In Collection: hrnet
Metadata:
backbone: HRNetV2p-W18
crop size: (512,512)
lr schd: 20000
inference time (ms/im):
- backend: PyTorch
batch size: 1
- value: 42.59
hardware: V100
backend: PyTorch
batch size: 1
mode: FP32
resolution: (512,512)
value: 42.59
lr schd: 20000
memory (GB): 2.9
Name: fcn_hr18_512x512_20k_voc12aug
Results:
- Task: Semantic Segmentation
Dataset: Pascal VOC 2012 + Aug
Metrics:
mIoU: 72.3
mIoU(ms+flip): 74.71
Task: Semantic Segmentation
Config: configs/hrnet/fcn_hr18_512x512_20k_voc12aug.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/hrnet/fcn_hr18_512x512_20k_voc12aug/fcn_hr18_512x512_20k_voc12aug_20200617_224503-488d45f7.pth
- Config: configs/hrnet/fcn_hr48_512x512_20k_voc12aug.py
- Name: fcn_hr48_512x512_20k_voc12aug
In Collection: hrnet
Metadata:
backbone: HRNetV2p-W48
crop size: (512,512)
lr schd: 20000
inference time (ms/im):
- backend: PyTorch
batch size: 1
- value: 45.35
hardware: V100
backend: PyTorch
batch size: 1
mode: FP32
resolution: (512,512)
value: 45.35
lr schd: 20000
memory (GB): 6.2
Name: fcn_hr48_512x512_20k_voc12aug
Results:
- Task: Semantic Segmentation
Dataset: Pascal VOC 2012 + Aug
Metrics:
mIoU: 75.87
mIoU(ms+flip): 78.58
Task: Semantic Segmentation
Config: configs/hrnet/fcn_hr48_512x512_20k_voc12aug.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/hrnet/fcn_hr48_512x512_20k_voc12aug/fcn_hr48_512x512_20k_voc12aug_20200617_224419-89de05cd.pth
- Config: configs/hrnet/fcn_hr18s_512x512_40k_voc12aug.py
- Name: fcn_hr18s_512x512_40k_voc12aug
In Collection: hrnet
Metadata:
backbone: HRNetV2p-W18-Small
crop size: (512,512)
lr schd: 40000
Name: fcn_hr18s_512x512_40k_voc12aug
Results:
- Task: Semantic Segmentation
Dataset: Pascal VOC 2012 + Aug
Metrics:
mIoU: 66.61
mIoU(ms+flip): 70.0
Task: Semantic Segmentation
Config: configs/hrnet/fcn_hr18s_512x512_40k_voc12aug.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/hrnet/fcn_hr18s_512x512_40k_voc12aug/fcn_hr18s_512x512_40k_voc12aug_20200614_000648-4f8d6e7f.pth
- Config: configs/hrnet/fcn_hr18_512x512_40k_voc12aug.py
- Name: fcn_hr18_512x512_40k_voc12aug
In Collection: hrnet
Metadata:
backbone: HRNetV2p-W18
crop size: (512,512)
lr schd: 40000
Name: fcn_hr18_512x512_40k_voc12aug
Results:
- Task: Semantic Segmentation
Dataset: Pascal VOC 2012 + Aug
Metrics:
mIoU: 72.9
mIoU(ms+flip): 75.59
Task: Semantic Segmentation
Config: configs/hrnet/fcn_hr18_512x512_40k_voc12aug.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/hrnet/fcn_hr18_512x512_40k_voc12aug/fcn_hr18_512x512_40k_voc12aug_20200613_224401-1b4b76cd.pth
- Config: configs/hrnet/fcn_hr48_512x512_40k_voc12aug.py
- Name: fcn_hr48_512x512_40k_voc12aug
In Collection: hrnet
Metadata:
backbone: HRNetV2p-W48
crop size: (512,512)
lr schd: 40000
Name: fcn_hr48_512x512_40k_voc12aug
Results:
- Task: Semantic Segmentation
Dataset: Pascal VOC 2012 + Aug
Metrics:
mIoU: 76.24
mIoU(ms+flip): 78.49
Task: Semantic Segmentation
Config: configs/hrnet/fcn_hr48_512x512_40k_voc12aug.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/hrnet/fcn_hr48_512x512_40k_voc12aug/fcn_hr48_512x512_40k_voc12aug_20200613_222111-1b0f18bc.pth
- Config: configs/hrnet/fcn_hr48_480x480_40k_pascal_context.py
- Name: fcn_hr48_480x480_40k_pascal_context
In Collection: hrnet
Metadata:
backbone: HRNetV2p-W48
crop size: (480,480)
lr schd: 40000
inference time (ms/im):
- backend: PyTorch
batch size: 1
- value: 112.87
hardware: V100
backend: PyTorch
batch size: 1
mode: FP32
resolution: (480,480)
value: 112.87
lr schd: 40000
memory (GB): 6.1
Name: fcn_hr48_480x480_40k_pascal_context
Results:
- Task: Semantic Segmentation
Dataset: Pascal Context
Metrics:
mIoU: 45.14
mIoU(ms+flip): 47.42
Task: Semantic Segmentation
Config: configs/hrnet/fcn_hr48_480x480_40k_pascal_context.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/hrnet/fcn_hr48_480x480_40k_pascal_context/fcn_hr48_480x480_40k_pascal_context_20200911_164852-667d00b0.pth
- Config: configs/hrnet/fcn_hr48_480x480_80k_pascal_context.py
- Name: fcn_hr48_480x480_80k_pascal_context
In Collection: hrnet
Metadata:
backbone: HRNetV2p-W48
crop size: (480,480)
lr schd: 80000
Name: fcn_hr48_480x480_80k_pascal_context
Results:
- Task: Semantic Segmentation
Dataset: Pascal Context
Metrics:
mIoU: 45.84
mIoU(ms+flip): 47.84
Task: Semantic Segmentation
Config: configs/hrnet/fcn_hr48_480x480_80k_pascal_context.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/hrnet/fcn_hr48_480x480_80k_pascal_context/fcn_hr48_480x480_80k_pascal_context_20200911_155322-847a6711.pth
- Config: configs/hrnet/fcn_hr48_480x480_40k_pascal_context_59.py
- Name: fcn_hr48_480x480_40k_pascal_context_59
In Collection: hrnet
Metadata:
backbone: HRNetV2p-W48
crop size: (480,480)
lr schd: 40000
Name: fcn_hr48_480x480_40k_pascal_context_59
Results:
- Task: Semantic Segmentation
Dataset: Pascal Context 59
Metrics:
mIoU: 50.33
mIoU(ms+flip): 52.83
Task: Semantic Segmentation
Config: configs/hrnet/fcn_hr48_480x480_40k_pascal_context_59.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/hrnet/fcn_hr48_480x480_40k_pascal_context_59/fcn_hr48_480x480_40k_pascal_context_59_20210410_122738-b808b8b2.pth
- Config: configs/hrnet/fcn_hr48_480x480_80k_pascal_context_59.py
- Name: fcn_hr48_480x480_80k_pascal_context_59
In Collection: hrnet
Metadata:
backbone: HRNetV2p-W48
crop size: (480,480)
lr schd: 80000
Name: fcn_hr48_480x480_80k_pascal_context_59
Results:
- Task: Semantic Segmentation
Dataset: Pascal Context 59
Metrics:
mIoU: 51.12
mIoU(ms+flip): 53.56
Task: Semantic Segmentation
Config: configs/hrnet/fcn_hr48_480x480_80k_pascal_context_59.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/hrnet/fcn_hr48_480x480_80k_pascal_context_59/fcn_hr48_480x480_80k_pascal_context_59_20210411_003240-3ae7081e.pth

9
configs/isanet/README.md
View File

@ -4,6 +4,13 @@
<!-- [ALGORITHM] -->
<a href="https://github.com/openseg-group/openseg.pytorch">Official Repo</a>
<a href="https://github.com/open-mmlab/mmsegmentation/blob/v0.18.0/mmseg/models/decode_heads/isa_head.py#L58">Code Snippet</a>
<details>
<summary align="right"><a href="https://arxiv.org/abs/1907.12273">ISANet (ArXiv'2019/IJCV'2021)</a></summary>
```
@article{huang2019isa,
title={Interlaced Sparse Self-Attention for Semantic Segmentation},
@ -23,6 +30,8 @@ The technical report above is also presented at:
}
```
</details>
## Results and models
### Cityscapes

237
configs/isanet/isanet.yml
View File

@ -1,360 +1,369 @@
Collections:
- Metadata:
- Name: isanet
Metadata:
Training Data:
- Cityscapes
- ADE20K
- Pascal VOC 2012 + Aug
Name: isanet
Paper:
URL: https://arxiv.org/abs/1907.12273
Title: Interlaced Sparse Self-Attention for Semantic Segmentation
README: configs/isanet/README.md
Code:
URL: https://github.com/open-mmlab/mmsegmentation/blob/v0.18.0/mmseg/models/decode_heads/isa_head.py#L58
Version: v0.18.0
Converted From:
Code: https://github.com/openseg-group/openseg.pytorch
Models:
- Config: configs/isanet/isanet_r50-d8_512x1024_40k_cityscapes.py
- Name: isanet_r50-d8_512x1024_40k_cityscapes
In Collection: isanet
Metadata:
backbone: R-50-D8
crop size: (512,1024)
lr schd: 40000
inference time (ms/im):
- backend: PyTorch
batch size: 1
- value: 343.64
hardware: V100
backend: PyTorch
batch size: 1
mode: FP32
resolution: (512,1024)
value: 343.64
lr schd: 40000
memory (GB): 5.869
Name: isanet_r50-d8_512x1024_40k_cityscapes
Results:
- Task: Semantic Segmentation
Dataset: Cityscapes
Metrics:
mIoU: 78.49
mIoU(ms+flip): 79.44
Task: Semantic Segmentation
Config: configs/isanet/isanet_r50-d8_512x1024_40k_cityscapes.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/isanet/isanet_r50-d8_512x1024_40k_cityscapes/isanet_r50-d8_512x1024_40k_cityscapes_20210901_054739-981bd763.pth
- Config: configs/isanet/isanet_r50-d8_512x1024_80k_cityscapes.py
- Name: isanet_r50-d8_512x1024_80k_cityscapes
In Collection: isanet
Metadata:
backbone: R-50-D8
crop size: (512,1024)
lr schd: 80000
inference time (ms/im):
- backend: PyTorch
batch size: 1
- value: 343.64
hardware: V100
backend: PyTorch
batch size: 1
mode: FP32
resolution: (512,1024)
value: 343.64
lr schd: 80000
memory (GB): 5.869
Name: isanet_r50-d8_512x1024_80k_cityscapes
Results:
- Task: Semantic Segmentation
Dataset: Cityscapes
Metrics:
mIoU: 78.68
mIoU(ms+flip): 80.25
Task: Semantic Segmentation
Config: configs/isanet/isanet_r50-d8_512x1024_80k_cityscapes.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/isanet/isanet_r50-d8_512x1024_80k_cityscapes/isanet_r50-d8_512x1024_80k_cityscapes_20210901_074202-89384497.pth
- Config: configs/isanet/isanet_r50-d8_769x769_40k_cityscapes.py
- Name: isanet_r50-d8_769x769_40k_cityscapes
In Collection: isanet
Metadata:
backbone: R-50-D8
crop size: (769,769)
lr schd: 40000
inference time (ms/im):
- backend: PyTorch
batch size: 1
- value: 649.35
hardware: V100
backend: PyTorch
batch size: 1
mode: FP32
resolution: (769,769)
value: 649.35
lr schd: 40000
memory (GB): 6.759
Name: isanet_r50-d8_769x769_40k_cityscapes
Results:
- Task: Semantic Segmentation
Dataset: Cityscapes
Metrics:
mIoU: 78.7
mIoU(ms+flip): 80.28
Task: Semantic Segmentation
Config: configs/isanet/isanet_r50-d8_769x769_40k_cityscapes.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/isanet/isanet_r50-d8_769x769_40k_cityscapes/isanet_r50-d8_769x769_40k_cityscapes_20210903_050200-4ae7e65b.pth
- Config: configs/isanet/isanet_r50-d8_769x769_80k_cityscapes.py
- Name: isanet_r50-d8_769x769_80k_cityscapes
In Collection: isanet
Metadata:
backbone: R-50-D8
crop size: (769,769)
lr schd: 80000
inference time (ms/im):
- backend: PyTorch
batch size: 1
- value: 649.35
hardware: V100
backend: PyTorch
batch size: 1
mode: FP32
resolution: (769,769)
value: 649.35
lr schd: 80000
memory (GB): 6.759
Name: isanet_r50-d8_769x769_80k_cityscapes
Results:
- Task: Semantic Segmentation
Dataset: Cityscapes
Metrics:
mIoU: 79.29
mIoU(ms+flip): 80.53
Task: Semantic Segmentation
Config: configs/isanet/isanet_r50-d8_769x769_80k_cityscapes.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/isanet/isanet_r50-d8_769x769_80k_cityscapes/isanet_r50-d8_769x769_80k_cityscapes_20210903_101126-99b54519.pth
- Config: configs/isanet/isanet_r101-d8_512x1024_40k_cityscapes.py
- Name: isanet_r101-d8_512x1024_40k_cityscapes
In Collection: isanet
Metadata:
backbone: R-101-D8
crop size: (512,1024)
lr schd: 40000
inference time (ms/im):
- backend: PyTorch
batch size: 1
- value: 425.53
hardware: V100
backend: PyTorch
batch size: 1
mode: FP32
resolution: (512,1024)
value: 425.53
lr schd: 40000
memory (GB): 9.425
Name: isanet_r101-d8_512x1024_40k_cityscapes
Results:
- Task: Semantic Segmentation
Dataset: Cityscapes
Metrics:
mIoU: 79.58
mIoU(ms+flip): 81.05
Task: Semantic Segmentation
Config: configs/isanet/isanet_r101-d8_512x1024_40k_cityscapes.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/isanet/isanet_r101-d8_512x1024_40k_cityscapes/isanet_r101-d8_512x1024_40k_cityscapes_20210901_145553-293e6bd6.pth
- Config: configs/isanet/isanet_r101-d8_512x1024_80k_cityscapes.py
- Name: isanet_r101-d8_512x1024_80k_cityscapes
In Collection: isanet
Metadata:
backbone: R-101-D8
crop size: (512,1024)
lr schd: 80000
inference time (ms/im):
- backend: PyTorch
batch size: 1
- value: 425.53
hardware: V100
backend: PyTorch
batch size: 1
mode: FP32
resolution: (512,1024)
value: 425.53
lr schd: 80000
memory (GB): 9.425
Name: isanet_r101-d8_512x1024_80k_cityscapes
Results:
- Task: Semantic Segmentation
Dataset: Cityscapes
Metrics:
mIoU: 80.32
mIoU(ms+flip): 81.58
Task: Semantic Segmentation
Config: configs/isanet/isanet_r101-d8_512x1024_80k_cityscapes.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/isanet/isanet_r101-d8_512x1024_80k_cityscapes/isanet_r101-d8_512x1024_80k_cityscapes_20210901_145243-5b99c9b2.pth
- Config: configs/isanet/isanet_r101-d8_769x769_40k_cityscapes.py
- Name: isanet_r101-d8_769x769_40k_cityscapes
In Collection: isanet
Metadata:
backbone: R-101-D8
crop size: (769,769)
lr schd: 40000
inference time (ms/im):
- backend: PyTorch
batch size: 1
- value: 1086.96
hardware: V100
backend: PyTorch
batch size: 1
mode: FP32
resolution: (769,769)
value: 1086.96
lr schd: 40000
memory (GB): 10.815
Name: isanet_r101-d8_769x769_40k_cityscapes
Results:
- Task: Semantic Segmentation
Dataset: Cityscapes
Metrics:
mIoU: 79.68
mIoU(ms+flip): 80.95
Task: Semantic Segmentation
Config: configs/isanet/isanet_r101-d8_769x769_40k_cityscapes.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/isanet/isanet_r101-d8_769x769_40k_cityscapes/isanet_r101-d8_769x769_40k_cityscapes_20210903_111320-509e7224.pth
- Config: configs/isanet/isanet_r101-d8_769x769_80k_cityscapes.py
- Name: isanet_r101-d8_769x769_80k_cityscapes
In Collection: isanet
Metadata:
backbone: R-101-D8
crop size: (769,769)
lr schd: 80000
inference time (ms/im):
- backend: PyTorch
batch size: 1
- value: 1086.96
hardware: V100
backend: PyTorch
batch size: 1
mode: FP32
resolution: (769,769)
value: 1086.96
lr schd: 80000
memory (GB): 10.815
Name: isanet_r101-d8_769x769_80k_cityscapes
Results:
- Task: Semantic Segmentation
Dataset: Cityscapes
Metrics:
mIoU: 80.61
mIoU(ms+flip): 81.59
Task: Semantic Segmentation
Config: configs/isanet/isanet_r101-d8_769x769_80k_cityscapes.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/isanet/isanet_r101-d8_769x769_80k_cityscapes/isanet_r101-d8_769x769_80k_cityscapes_20210903_111319-24f71dfa.pth
- Config: configs/isanet/isanet_r50-d8_512x512_80k_ade20k.py
- Name: isanet_r50-d8_512x512_80k_ade20k
In Collection: isanet
Metadata:
backbone: R-50-D8
crop size: (512,512)
lr schd: 80000
inference time (ms/im):
- backend: PyTorch
batch size: 1
- value: 44.35
hardware: V100
backend: PyTorch
batch size: 1
mode: FP32
resolution: (512,512)
value: 44.35
lr schd: 80000
memory (GB): 9.0
Name: isanet_r50-d8_512x512_80k_ade20k
Results:
- Task: Semantic Segmentation
Dataset: ADE20K
Metrics:
mIoU: 41.12
mIoU(ms+flip): 42.35
Task: Semantic Segmentation
Config: configs/isanet/isanet_r50-d8_512x512_80k_ade20k.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/isanet/isanet_r50-d8_512x512_80k_ade20k/isanet_r50-d8_512x512_80k_ade20k_20210903_124557-6ed83a0c.pth
- Config: configs/isanet/isanet_r50-d8_512x512_160k_ade20k.py
- Name: isanet_r50-d8_512x512_160k_ade20k
In Collection: isanet
Metadata:
backbone: R-50-D8
crop size: (512,512)
lr schd: 160000
inference time (ms/im):
- backend: PyTorch
batch size: 1
- value: 44.35
hardware: V100
backend: PyTorch
batch size: 1
mode: FP32
resolution: (512,512)
value: 44.35
lr schd: 160000
memory (GB): 9.0
Name: isanet_r50-d8_512x512_160k_ade20k
Results:
- Task: Semantic Segmentation
Dataset: ADE20K
Metrics:
mIoU: 42.59
mIoU(ms+flip): 43.07
Task: Semantic Segmentation
Config: configs/isanet/isanet_r50-d8_512x512_160k_ade20k.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/isanet/isanet_r50-d8_512x512_160k_ade20k/isanet_r50-d8_512x512_160k_ade20k_20210903_104850-f752d0a3.pth
- Config: configs/isanet/isanet_r101-d8_512x512_80k_ade20k.py
- Name: isanet_r101-d8_512x512_80k_ade20k
In Collection: isanet
Metadata:
backbone: R-101-D8
crop size: (512,512)
lr schd: 80000
inference time (ms/im):
- backend: PyTorch
batch size: 1
- value: 94.7
hardware: V100
backend: PyTorch
batch size: 1
mode: FP32
resolution: (512,512)
value: 94.7
lr schd: 80000
memory (GB): 12.562
Name: isanet_r101-d8_512x512_80k_ade20k
Results:
- Task: Semantic Segmentation
Dataset: ADE20K
Metrics:
mIoU: 43.51
mIoU(ms+flip): 44.38
Task: Semantic Segmentation
Config: configs/isanet/isanet_r101-d8_512x512_80k_ade20k.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/isanet/isanet_r101-d8_512x512_80k_ade20k/isanet_r101-d8_512x512_80k_ade20k_20210903_162056-68b235c2.pth
- Config: configs/isanet/isanet_r101-d8_512x512_160k_ade20k.py
- Name: isanet_r101-d8_512x512_160k_ade20k
In Collection: isanet
Metadata:
backbone: R-101-D8
crop size: (512,512)
lr schd: 160000
inference time (ms/im):
- backend: PyTorch
batch size: 1
- value: 94.7
hardware: V100
backend: PyTorch
batch size: 1
mode: FP32
resolution: (512,512)
value: 94.7
lr schd: 160000
memory (GB): 12.562
Name: isanet_r101-d8_512x512_160k_ade20k
Results:
- Task: Semantic Segmentation
Dataset: ADE20K
Metrics:
mIoU: 43.8
mIoU(ms+flip): 45.4
Task: Semantic Segmentation
Config: configs/isanet/isanet_r101-d8_512x512_160k_ade20k.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/isanet/isanet_r101-d8_512x512_160k_ade20k/isanet_r101-d8_512x512_160k_ade20k_20210903_211431-a7879dcd.pth
- Config: configs/isanet/isanet_r50-d8_512x512_20k_voc12aug.py
- Name: isanet_r50-d8_512x512_20k_voc12aug
In Collection: isanet
Metadata:
backbone: R-50-D8
crop size: (512,512)
lr schd: 20000
inference time (ms/im):
- backend: PyTorch
batch size: 1
- value: 43.33
hardware: V100
backend: PyTorch
batch size: 1
mode: FP32
resolution: (512,512)
value: 43.33
lr schd: 20000
memory (GB): 5.9
Name: isanet_r50-d8_512x512_20k_voc12aug
Results:
- Task: Semantic Segmentation
Dataset: Pascal VOC 2012 + Aug
Metrics:
mIoU: 76.78
mIoU(ms+flip): 77.79
Task: Semantic Segmentation
Config: configs/isanet/isanet_r50-d8_512x512_20k_voc12aug.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/isanet/isanet_r50-d8_512x512_20k_voc12aug/isanet_r50-d8_512x512_20k_voc12aug_20210901_164838-79d59b80.pth
- Config: configs/isanet/isanet_r50-d8_512x512_40k_voc12aug.py
- Name: isanet_r50-d8_512x512_40k_voc12aug
In Collection: isanet
Metadata:
backbone: R-50-D8
crop size: (512,512)
lr schd: 40000
inference time (ms/im):
- backend: PyTorch
batch size: 1
- value: 43.33
hardware: V100
backend: PyTorch
batch size: 1
mode: FP32
resolution: (512,512)
value: 43.33
lr schd: 40000
memory (GB): 5.9
Name: isanet_r50-d8_512x512_40k_voc12aug
Results:
- Task: Semantic Segmentation
Dataset: Pascal VOC 2012 + Aug
Metrics:
mIoU: 76.2
mIoU(ms+flip): 77.22
Task: Semantic Segmentation
Config: configs/isanet/isanet_r50-d8_512x512_40k_voc12aug.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/isanet/isanet_r50-d8_512x512_40k_voc12aug/isanet_r50-d8_512x512_40k_voc12aug_20210901_151349-7d08a54e.pth
- Config: configs/isanet/isanet_r101-d8_512x512_20k_voc12aug.py
- Name: isanet_r101-d8_512x512_20k_voc12aug
In Collection: isanet
Metadata:
backbone: R-101-D8
crop size: (512,512)
lr schd: 20000
inference time (ms/im):
- backend: PyTorch
batch size: 1
- value: 134.77
hardware: V100
backend: PyTorch
batch size: 1
mode: FP32
resolution: (512,512)
value: 134.77
lr schd: 20000
memory (GB): 9.465
Name: isanet_r101-d8_512x512_20k_voc12aug
Results:
- Task: Semantic Segmentation
Dataset: Pascal VOC 2012 + Aug
Metrics:
mIoU: 78.46
mIoU(ms+flip): 79.16
Task: Semantic Segmentation
Config: configs/isanet/isanet_r101-d8_512x512_20k_voc12aug.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/isanet/isanet_r101-d8_512x512_20k_voc12aug/isanet_r101-d8_512x512_20k_voc12aug_20210901_115805-3ccbf355.pth
- Config: configs/isanet/isanet_r101-d8_512x512_40k_voc12aug.py
- Name: isanet_r101-d8_512x512_40k_voc12aug
In Collection: isanet
Metadata:
backbone: R-101-D8
crop size: (512,512)
lr schd: 40000
inference time (ms/im):
- backend: PyTorch
batch size: 1
- value: 134.77
hardware: V100
backend: PyTorch
batch size: 1
mode: FP32
resolution: (512,512)
value: 134.77
lr schd: 40000
memory (GB): 9.465
Name: isanet_r101-d8_512x512_40k_voc12aug
Results:
- Task: Semantic Segmentation
Dataset: Pascal VOC 2012 + Aug
Metrics:
mIoU: 78.12
mIoU(ms+flip): 79.04
Task: Semantic Segmentation
Config: configs/isanet/isanet_r101-d8_512x512_40k_voc12aug.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/isanet/isanet_r101-d8_512x512_40k_voc12aug/isanet_r101-d8_512x512_40k_voc12aug_20210901_145814-bc71233b.pth

9
configs/mobilenet_v2/README.md
View File

@ -4,6 +4,13 @@
<!-- [ALGORITHM] -->
<a href="https://github.com/tensorflow/models/tree/master/research/deeplab">Official Repo</a>
<a href="https://github.com/open-mmlab/mmsegmentation/blob/v0.17.0/mmseg/models/backbones/mobilenet_v2.py#L14">Code Snippet</a>
<details>
<summary align="right"><a href="https://arxiv.org/abs/1801.04381">MobileNetV2 (CVPR'2018)</a></summary>
```latex
@inproceedings{sandler2018mobilenetv2,
title={Mobilenetv2: Inverted residuals and linear bottlenecks},
@ -14,6 +21,8 @@
}
```
</details>
## Results and models
### Cityscapes

125
configs/mobilenet_v2/mobilenet_v2.yml
View File

@ -1,175 +1,184 @@
Collections:
- Metadata:
- Name: mobilenet_v2
Metadata:
Training Data:
- Cityscapes
- ADE20k
Name: mobilenet_v2
Paper:
URL: https://arxiv.org/abs/1801.04381
Title: 'MobileNetV2: Inverted Residuals and Linear Bottlenecks'
README: configs/mobilenet_v2/README.md
Code:
URL: https://github.com/open-mmlab/mmsegmentation/blob/v0.17.0/mmseg/models/backbones/mobilenet_v2.py#L14
Version: v0.17.0
Converted From:
Code: https://github.com/tensorflow/models/tree/master/research/deeplab
Models:
- Config: configs/mobilenet_v2/fcn_m-v2-d8_512x1024_80k_cityscapes.py
- Name: fcn_m-v2-d8_512x1024_80k_cityscapes
In Collection: mobilenet_v2
Metadata:
backbone: M-V2-D8
crop size: (512,1024)
lr schd: 80000
inference time (ms/im):
- backend: PyTorch
batch size: 1
- value: 70.42
hardware: V100
backend: PyTorch
batch size: 1
mode: FP32
resolution: (512,1024)
value: 70.42
lr schd: 80000
memory (GB): 3.4
Name: fcn_m-v2-d8_512x1024_80k_cityscapes
Results:
- Task: Semantic Segmentation
Dataset: Cityscapes
Metrics:
mIoU: 61.54
Task: Semantic Segmentation
Config: configs/mobilenet_v2/fcn_m-v2-d8_512x1024_80k_cityscapes.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/mobilenet_v2/fcn_m-v2-d8_512x1024_80k_cityscapes/fcn_m-v2-d8_512x1024_80k_cityscapes_20200825_124817-d24c28c1.pth
- Config: configs/mobilenet_v2/pspnet_m-v2-d8_512x1024_80k_cityscapes.py
- Name: pspnet_m-v2-d8_512x1024_80k_cityscapes
In Collection: mobilenet_v2
Metadata:
backbone: M-V2-D8
crop size: (512,1024)
lr schd: 80000
inference time (ms/im):
- backend: PyTorch
batch size: 1
- value: 89.29
hardware: V100
backend: PyTorch
batch size: 1
mode: FP32
resolution: (512,1024)
value: 89.29
lr schd: 80000
memory (GB): 3.6
Name: pspnet_m-v2-d8_512x1024_80k_cityscapes
Results:
- Task: Semantic Segmentation
Dataset: Cityscapes
Metrics:
mIoU: 70.23
Task: Semantic Segmentation
Config: configs/mobilenet_v2/pspnet_m-v2-d8_512x1024_80k_cityscapes.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/mobilenet_v2/pspnet_m-v2-d8_512x1024_80k_cityscapes/pspnet_m-v2-d8_512x1024_80k_cityscapes_20200825_124817-19e81d51.pth
- Config: configs/mobilenet_v2/deeplabv3_m-v2-d8_512x1024_80k_cityscapes.py
- Name: deeplabv3_m-v2-d8_512x1024_80k_cityscapes
In Collection: mobilenet_v2
Metadata:
backbone: M-V2-D8
crop size: (512,1024)
lr schd: 80000
inference time (ms/im):
- backend: PyTorch
batch size: 1
- value: 119.05
hardware: V100
backend: PyTorch
batch size: 1
mode: FP32
resolution: (512,1024)
value: 119.05
lr schd: 80000
memory (GB): 3.9
Name: deeplabv3_m-v2-d8_512x1024_80k_cityscapes
Results:
- Task: Semantic Segmentation
Dataset: Cityscapes
Metrics:
mIoU: 73.84
Task: Semantic Segmentation
Config: configs/mobilenet_v2/deeplabv3_m-v2-d8_512x1024_80k_cityscapes.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/mobilenet_v2/deeplabv3_m-v2-d8_512x1024_80k_cityscapes/deeplabv3_m-v2-d8_512x1024_80k_cityscapes_20200825_124836-bef03590.pth
- Config: configs/mobilenet_v2/deeplabv3plus_m-v2-d8_512x1024_80k_cityscapes.py
- Name: deeplabv3plus_m-v2-d8_512x1024_80k_cityscapes
In Collection: mobilenet_v2
Metadata:
backbone: M-V2-D8
crop size: (512,1024)
lr schd: 80000
inference time (ms/im):
- backend: PyTorch
batch size: 1
- value: 119.05
hardware: V100
backend: PyTorch
batch size: 1
mode: FP32
resolution: (512,1024)
value: 119.05
lr schd: 80000
memory (GB): 5.1
Name: deeplabv3plus_m-v2-d8_512x1024_80k_cityscapes
Results:
- Task: Semantic Segmentation
Dataset: Cityscapes
Metrics:
mIoU: 75.2
Task: Semantic Segmentation
Config: configs/mobilenet_v2/deeplabv3plus_m-v2-d8_512x1024_80k_cityscapes.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/mobilenet_v2/deeplabv3plus_m-v2-d8_512x1024_80k_cityscapes/deeplabv3plus_m-v2-d8_512x1024_80k_cityscapes_20200825_124836-d256dd4b.pth
- Config: configs/mobilenet_v2/fcn_m-v2-d8_512x512_160k_ade20k.py
- Name: fcn_m-v2-d8_512x512_160k_ade20k
In Collection: mobilenet_v2
Metadata:
backbone: M-V2-D8
crop size: (512,512)
lr schd: 160000
inference time (ms/im):
- backend: PyTorch
batch size: 1
- value: 15.53
hardware: V100
backend: PyTorch
batch size: 1
mode: FP32
resolution: (512,512)
value: 15.53
lr schd: 160000
memory (GB): 6.5
Name: fcn_m-v2-d8_512x512_160k_ade20k
Results:
- Task: Semantic Segmentation
Dataset: ADE20k
Metrics:
mIoU: 19.71
Task: Semantic Segmentation
Config: configs/mobilenet_v2/fcn_m-v2-d8_512x512_160k_ade20k.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/mobilenet_v2/fcn_m-v2-d8_512x512_160k_ade20k/fcn_m-v2-d8_512x512_160k_ade20k_20200825_214953-c40e1095.pth
- Config: configs/mobilenet_v2/pspnet_m-v2-d8_512x512_160k_ade20k.py
- Name: pspnet_m-v2-d8_512x512_160k_ade20k
In Collection: mobilenet_v2
Metadata:
backbone: M-V2-D8
crop size: (512,512)
lr schd: 160000
inference time (ms/im):
- backend: PyTorch
batch size: 1
- value: 17.33
hardware: V100
backend: PyTorch
batch size: 1
mode: FP32
resolution: (512,512)
value: 17.33
lr schd: 160000
memory (GB): 6.5
Name: pspnet_m-v2-d8_512x512_160k_ade20k
Results:
- Task: Semantic Segmentation
Dataset: ADE20k
Metrics:
mIoU: 29.68
Task: Semantic Segmentation
Config: configs/mobilenet_v2/pspnet_m-v2-d8_512x512_160k_ade20k.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/mobilenet_v2/pspnet_m-v2-d8_512x512_160k_ade20k/pspnet_m-v2-d8_512x512_160k_ade20k_20200825_214953-f5942f7a.pth
- Config: configs/mobilenet_v2/deeplabv3_m-v2-d8_512x512_160k_ade20k.py
- Name: deeplabv3_m-v2-d8_512x512_160k_ade20k
In Collection: mobilenet_v2
Metadata:
backbone: M-V2-D8
crop size: (512,512)
lr schd: 160000
inference time (ms/im):
- backend: PyTorch
batch size: 1
- value: 25.06
hardware: V100
backend: PyTorch
batch size: 1
mode: FP32
resolution: (512,512)
value: 25.06
lr schd: 160000
memory (GB): 6.8
Name: deeplabv3_m-v2-d8_512x512_160k_ade20k
Results:
- Task: Semantic Segmentation
Dataset: ADE20k
Metrics:
mIoU: 34.08
Task: Semantic Segmentation
Config: configs/mobilenet_v2/deeplabv3_m-v2-d8_512x512_160k_ade20k.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/mobilenet_v2/deeplabv3_m-v2-d8_512x512_160k_ade20k/deeplabv3_m-v2-d8_512x512_160k_ade20k_20200825_223255-63986343.pth
- Config: configs/mobilenet_v2/deeplabv3plus_m-v2-d8_512x512_160k_ade20k.py
- Name: deeplabv3plus_m-v2-d8_512x512_160k_ade20k
In Collection: mobilenet_v2
Metadata:
backbone: M-V2-D8
crop size: (512,512)
lr schd: 160000
inference time (ms/im):
- backend: PyTorch
batch size: 1
- value: 23.2
hardware: V100
backend: PyTorch
batch size: 1
mode: FP32
resolution: (512,512)
value: 23.2
lr schd: 160000
memory (GB): 8.2
Name: deeplabv3plus_m-v2-d8_512x512_160k_ade20k
Results:
- Task: Semantic Segmentation
Dataset: ADE20k
Metrics:
mIoU: 34.02
Task: Semantic Segmentation
Config: configs/mobilenet_v2/deeplabv3plus_m-v2-d8_512x512_160k_ade20k.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/mobilenet_v2/deeplabv3plus_m-v2-d8_512x512_160k_ade20k/deeplabv3plus_m-v2-d8_512x512_160k_ade20k_20200825_223255-465a01d4.pth

9
configs/mobilenet_v3/README.md
View File

@ -4,6 +4,13 @@
<!-- [ALGORITHM] -->
<a href="https://github.com/tensorflow/models/tree/master/research/deeplab">Official Repo</a>
<a href="https://github.com/open-mmlab/mmsegmentation/blob/v0.17.0/mmseg/models/backbones/mobilenet_v3.py#L15">Code Snippet</a>
<details>
<summary align="right"><a href="https://arxiv.org/abs/1801.04381">MobileNetV3 (ICCV'2019)</a></summary>
```latex
@inproceedings{Howard_2019_ICCV,
title={Searching for MobileNetV3},
@ -16,6 +23,8 @@
}
```
</details>
## Results and models
### Cityscapes

69
configs/mobilenet_v3/mobilenet_v3.yml
View File

@ -1,94 +1,103 @@
Collections:
- Metadata:
- Name: mobilenet_v3
Metadata:
Training Data:
- Cityscapes
Name: mobilenet_v3
Paper:
URL: https://arxiv.org/abs/1801.04381
Title: Searching for MobileNetV3
README: configs/mobilenet_v3/README.md
Code:
URL: https://github.com/open-mmlab/mmsegmentation/blob/v0.17.0/mmseg/models/backbones/mobilenet_v3.py#L15
Version: v0.17.0
Converted From:
Code: https://github.com/tensorflow/models/tree/master/research/deeplab
Models:
- Config: configs/mobilenet_v3/lraspp_m-v3-d8_512x1024_320k_cityscapes.py
- Name: lraspp_m-v3-d8_512x1024_320k_cityscapes
In Collection: mobilenet_v3
Metadata:
backbone: M-V3-D8
crop size: (512,1024)
lr schd: 320000
inference time (ms/im):
- backend: PyTorch
batch size: 1
- value: 65.7
hardware: V100
backend: PyTorch
batch size: 1
mode: FP32
resolution: (512,1024)
value: 65.7
lr schd: 320000
memory (GB): 8.9
Name: lraspp_m-v3-d8_512x1024_320k_cityscapes
Results:
- Task: Semantic Segmentation
Dataset: Cityscapes
Metrics:
mIoU: 69.54
mIoU(ms+flip): 70.89
Task: Semantic Segmentation
Config: configs/mobilenet_v3/lraspp_m-v3-d8_512x1024_320k_cityscapes.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/mobilenet_v3/lraspp_m-v3-d8_512x1024_320k_cityscapes/lraspp_m-v3-d8_512x1024_320k_cityscapes_20201224_220337-cfe8fb07.pth
- Config: configs/mobilenet_v3/lraspp_m-v3-d8_scratch_512x1024_320k_cityscapes.py
- Name: lraspp_m-v3-d8_scratch_512x1024_320k_cityscapes
In Collection: mobilenet_v3
Metadata:
backbone: M-V3-D8 (scratch)
crop size: (512,1024)
lr schd: 320000
inference time (ms/im):
- backend: PyTorch
batch size: 1
- value: 67.7
hardware: V100
backend: PyTorch
batch size: 1
mode: FP32
resolution: (512,1024)
value: 67.7
lr schd: 320000
memory (GB): 8.9
Name: lraspp_m-v3-d8_scratch_512x1024_320k_cityscapes
Results:
- Task: Semantic Segmentation
Dataset: Cityscapes
Metrics:
mIoU: 67.87
mIoU(ms+flip): 69.78
Task: Semantic Segmentation
Config: configs/mobilenet_v3/lraspp_m-v3-d8_scratch_512x1024_320k_cityscapes.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/mobilenet_v3/lraspp_m-v3-d8_scratch_512x1024_320k_cityscapes/lraspp_m-v3-d8_scratch_512x1024_320k_cityscapes_20201224_220337-9f29cd72.pth
- Config: configs/mobilenet_v3/lraspp_m-v3s-d8_512x1024_320k_cityscapes.py
- Name: lraspp_m-v3s-d8_512x1024_320k_cityscapes
In Collection: mobilenet_v3
Metadata:
backbone: M-V3s-D8
crop size: (512,1024)
lr schd: 320000
inference time (ms/im):
- backend: PyTorch
batch size: 1
- value: 42.3
hardware: V100
backend: PyTorch
batch size: 1
mode: FP32
resolution: (512,1024)
value: 42.3
lr schd: 320000
memory (GB): 5.3
Name: lraspp_m-v3s-d8_512x1024_320k_cityscapes
Results:
- Task: Semantic Segmentation
Dataset: Cityscapes
Metrics:
mIoU: 64.11
mIoU(ms+flip): 66.42
Task: Semantic Segmentation
Config: configs/mobilenet_v3/lraspp_m-v3s-d8_512x1024_320k_cityscapes.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/mobilenet_v3/lraspp_m-v3s-d8_512x1024_320k_cityscapes/lraspp_m-v3s-d8_512x1024_320k_cityscapes_20201224_223935-61565b34.pth
- Config: configs/mobilenet_v3/lraspp_m-v3s-d8_scratch_512x1024_320k_cityscapes.py
- Name: lraspp_m-v3s-d8_scratch_512x1024_320k_cityscapes
In Collection: mobilenet_v3
Metadata:
backbone: M-V3s-D8 (scratch)
crop size: (512,1024)
lr schd: 320000
inference time (ms/im):
- backend: PyTorch
batch size: 1
- value: 40.82
hardware: V100
backend: PyTorch
batch size: 1
mode: FP32
resolution: (512,1024)
value: 40.82
lr schd: 320000
memory (GB): 5.3
Name: lraspp_m-v3s-d8_scratch_512x1024_320k_cityscapes
Results:
- Task: Semantic Segmentation
Dataset: Cityscapes
Metrics:
mIoU: 62.74
mIoU(ms+flip): 65.01
Task: Semantic Segmentation
Config: configs/mobilenet_v3/lraspp_m-v3s-d8_scratch_512x1024_320k_cityscapes.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/mobilenet_v3/lraspp_m-v3s-d8_scratch_512x1024_320k_cityscapes/lraspp_m-v3s-d8_scratch_512x1024_320k_cityscapes_20201224_223935-03daeabb.pth

9
configs/nonlocal_net/README.md
View File

@ -4,6 +4,13 @@
<!-- [ALGORITHM] -->
<a href="https://github.com/facebookresearch/video-nonlocal-net">Official Repo</a>
<a href="https://github.com/open-mmlab/mmsegmentation/blob/v0.17.0/mmseg/models/decode_heads/nl_head.py#L10">Code Snippet</a>
<details>
<summary align="right"><a href="https://arxiv.org/abs/1711.07971">NonLocal Net (CVPR'2018)</a></summary>
```latex
@inproceedings{wang2018non,
title={Non-local neural networks},
@ -14,6 +21,8 @@
}
```
</details>
## Results and models
### Cityscapes

173
configs/nonlocal_net/nonlocal_net.yml
View File

@ -1,292 +1,301 @@
Collections:
- Metadata:
- Name: nonlocal_net
Metadata:
Training Data:
- Cityscapes
- ADE20K
- Pascal VOC 2012 + Aug
Name: nonlocal_net
Paper:
URL: https://arxiv.org/abs/1711.07971
Title: Non-local Neural Networks
README: configs/nonlocal_net/README.md
Code:
URL: https://github.com/open-mmlab/mmsegmentation/blob/v0.17.0/mmseg/models/decode_heads/nl_head.py#L10
Version: v0.17.0
Converted From:
Code: https://github.com/facebookresearch/video-nonlocal-net
Models:
- Config: configs/nonlocal_net/nonlocal_r50-d8_512x1024_40k_cityscapes.py
- Name: nonlocal_r50-d8_512x1024_40k_cityscapes
In Collection: nonlocal_net
Metadata:
backbone: R-50-D8
crop size: (512,1024)
lr schd: 40000
inference time (ms/im):
- backend: PyTorch
batch size: 1
- value: 367.65
hardware: V100
backend: PyTorch
batch size: 1
mode: FP32
resolution: (512,1024)
value: 367.65
lr schd: 40000
memory (GB): 7.4
Name: nonlocal_r50-d8_512x1024_40k_cityscapes
Results:
- Task: Semantic Segmentation
Dataset: Cityscapes
Metrics:
mIoU: 78.24
Task: Semantic Segmentation
Config: configs/nonlocal_net/nonlocal_r50-d8_512x1024_40k_cityscapes.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/nonlocal_net/nonlocal_r50-d8_512x1024_40k_cityscapes/nonlocal_r50-d8_512x1024_40k_cityscapes_20200605_210748-c75e81e3.pth
- Config: configs/nonlocal_net/nonlocal_r101-d8_512x1024_40k_cityscapes.py
- Name: nonlocal_r101-d8_512x1024_40k_cityscapes
In Collection: nonlocal_net
Metadata:
backbone: R-101-D8
crop size: (512,1024)
lr schd: 40000
inference time (ms/im):
- backend: PyTorch
batch size: 1
- value: 512.82
hardware: V100
backend: PyTorch
batch size: 1
mode: FP32
resolution: (512,1024)
value: 512.82
lr schd: 40000
memory (GB): 10.9
Name: nonlocal_r101-d8_512x1024_40k_cityscapes
Results:
- Task: Semantic Segmentation
Dataset: Cityscapes
Metrics:
mIoU: 78.66
Task: Semantic Segmentation
Config: configs/nonlocal_net/nonlocal_r101-d8_512x1024_40k_cityscapes.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/nonlocal_net/nonlocal_r101-d8_512x1024_40k_cityscapes/nonlocal_r101-d8_512x1024_40k_cityscapes_20200605_210748-d63729fa.pth
- Config: configs/nonlocal_net/nonlocal_r50-d8_769x769_40k_cityscapes.py
- Name: nonlocal_r50-d8_769x769_40k_cityscapes
In Collection: nonlocal_net
Metadata:
backbone: R-50-D8
crop size: (769,769)
lr schd: 40000
inference time (ms/im):
- backend: PyTorch
batch size: 1
- value: 657.89
hardware: V100
backend: PyTorch
batch size: 1
mode: FP32
resolution: (769,769)
value: 657.89
lr schd: 40000
memory (GB): 8.9
Name: nonlocal_r50-d8_769x769_40k_cityscapes
Results:
- Task: Semantic Segmentation
Dataset: Cityscapes
Metrics:
mIoU: 78.33
mIoU(ms+flip): 79.92
Task: Semantic Segmentation
Config: configs/nonlocal_net/nonlocal_r50-d8_769x769_40k_cityscapes.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/nonlocal_net/nonlocal_r50-d8_769x769_40k_cityscapes/nonlocal_r50-d8_769x769_40k_cityscapes_20200530_045243-82ef6749.pth
- Config: configs/nonlocal_net/nonlocal_r101-d8_769x769_40k_cityscapes.py
- Name: nonlocal_r101-d8_769x769_40k_cityscapes
In Collection: nonlocal_net
Metadata:
backbone: R-101-D8
crop size: (769,769)
lr schd: 40000
inference time (ms/im):
- backend: PyTorch
batch size: 1
- value: 952.38
hardware: V100
backend: PyTorch
batch size: 1
mode: FP32
resolution: (769,769)
value: 952.38
lr schd: 40000
memory (GB): 12.8
Name: nonlocal_r101-d8_769x769_40k_cityscapes
Results:
- Task: Semantic Segmentation
Dataset: Cityscapes
Metrics:
mIoU: 78.57
mIoU(ms+flip): 80.29
Task: Semantic Segmentation
Config: configs/nonlocal_net/nonlocal_r101-d8_769x769_40k_cityscapes.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/nonlocal_net/nonlocal_r101-d8_769x769_40k_cityscapes/nonlocal_r101-d8_769x769_40k_cityscapes_20200530_045348-8fe9a9dc.pth
- Config: configs/nonlocal_net/nonlocal_r50-d8_512x1024_80k_cityscapes.py
- Name: nonlocal_r50-d8_512x1024_80k_cityscapes
In Collection: nonlocal_net
Metadata:
backbone: R-50-D8
crop size: (512,1024)
lr schd: 80000
Name: nonlocal_r50-d8_512x1024_80k_cityscapes
Results:
- Task: Semantic Segmentation
Dataset: Cityscapes
Metrics:
mIoU: 78.01
Task: Semantic Segmentation
Config: configs/nonlocal_net/nonlocal_r50-d8_512x1024_80k_cityscapes.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/nonlocal_net/nonlocal_r50-d8_512x1024_80k_cityscapes/nonlocal_r50-d8_512x1024_80k_cityscapes_20200607_193518-d6839fae.pth
- Config: configs/nonlocal_net/nonlocal_r101-d8_512x1024_80k_cityscapes.py
- Name: nonlocal_r101-d8_512x1024_80k_cityscapes
In Collection: nonlocal_net
Metadata:
backbone: R-101-D8
crop size: (512,1024)
lr schd: 80000
Name: nonlocal_r101-d8_512x1024_80k_cityscapes
Results:
- Task: Semantic Segmentation
Dataset: Cityscapes
Metrics:
mIoU: 78.93
Task: Semantic Segmentation
Config: configs/nonlocal_net/nonlocal_r101-d8_512x1024_80k_cityscapes.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/nonlocal_net/nonlocal_r101-d8_512x1024_80k_cityscapes/nonlocal_r101-d8_512x1024_80k_cityscapes_20200607_183411-32700183.pth
- Config: configs/nonlocal_net/nonlocal_r50-d8_769x769_80k_cityscapes.py
- Name: nonlocal_r50-d8_769x769_80k_cityscapes
In Collection: nonlocal_net
Metadata:
backbone: R-50-D8
crop size: (769,769)
lr schd: 80000
Name: nonlocal_r50-d8_769x769_80k_cityscapes
Results:
- Task: Semantic Segmentation
Dataset: Cityscapes
Metrics:
mIoU: 79.05
mIoU(ms+flip): 80.68
Task: Semantic Segmentation
Config: configs/nonlocal_net/nonlocal_r50-d8_769x769_80k_cityscapes.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/nonlocal_net/nonlocal_r50-d8_769x769_80k_cityscapes/nonlocal_r50-d8_769x769_80k_cityscapes_20200607_193506-1f9792f6.pth
- Config: configs/nonlocal_net/nonlocal_r101-d8_769x769_80k_cityscapes.py
- Name: nonlocal_r101-d8_769x769_80k_cityscapes
In Collection: nonlocal_net
Metadata:
backbone: R-101-D8
crop size: (769,769)
lr schd: 80000
Name: nonlocal_r101-d8_769x769_80k_cityscapes
Results:
- Task: Semantic Segmentation
Dataset: Cityscapes
Metrics:
mIoU: 79.4
mIoU(ms+flip): 80.85
Task: Semantic Segmentation
Config: configs/nonlocal_net/nonlocal_r101-d8_769x769_80k_cityscapes.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/nonlocal_net/nonlocal_r101-d8_769x769_80k_cityscapes/nonlocal_r101-d8_769x769_80k_cityscapes_20200607_183428-0e1fa4f9.pth
- Config: configs/nonlocal_net/nonlocal_r50-d8_512x512_80k_ade20k.py
- Name: nonlocal_r50-d8_512x512_80k_ade20k
In Collection: nonlocal_net
Metadata:
backbone: R-50-D8
crop size: (512,512)
lr schd: 80000
inference time (ms/im):
- backend: PyTorch
batch size: 1
- value: 46.79
hardware: V100
backend: PyTorch
batch size: 1
mode: FP32
resolution: (512,512)
value: 46.79
lr schd: 80000
memory (GB): 9.1
Name: nonlocal_r50-d8_512x512_80k_ade20k
Results:
- Task: Semantic Segmentation
Dataset: ADE20K
Metrics:
mIoU: 40.75
mIoU(ms+flip): 42.05
Task: Semantic Segmentation
Config: configs/nonlocal_net/nonlocal_r50-d8_512x512_80k_ade20k.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/nonlocal_net/nonlocal_r50-d8_512x512_80k_ade20k/nonlocal_r50-d8_512x512_80k_ade20k_20200615_015801-5ae0aa33.pth
- Config: configs/nonlocal_net/nonlocal_r101-d8_512x512_80k_ade20k.py
- Name: nonlocal_r101-d8_512x512_80k_ade20k
In Collection: nonlocal_net
Metadata:
backbone: R-101-D8
crop size: (512,512)
lr schd: 80000
inference time (ms/im):
- backend: PyTorch
batch size: 1
- value: 71.58
hardware: V100
backend: PyTorch
batch size: 1
mode: FP32
resolution: (512,512)
value: 71.58
lr schd: 80000
memory (GB): 12.6
Name: nonlocal_r101-d8_512x512_80k_ade20k
Results:
- Task: Semantic Segmentation
Dataset: ADE20K
Metrics:
mIoU: 42.9
mIoU(ms+flip): 44.27
Task: Semantic Segmentation
Config: configs/nonlocal_net/nonlocal_r101-d8_512x512_80k_ade20k.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/nonlocal_net/nonlocal_r101-d8_512x512_80k_ade20k/nonlocal_r101-d8_512x512_80k_ade20k_20200615_015758-24105919.pth
- Config: configs/nonlocal_net/nonlocal_r50-d8_512x512_160k_ade20k.py
- Name: nonlocal_r50-d8_512x512_160k_ade20k
In Collection: nonlocal_net
Metadata:
backbone: R-50-D8
crop size: (512,512)
lr schd: 160000
Name: nonlocal_r50-d8_512x512_160k_ade20k
Results:
- Task: Semantic Segmentation
Dataset: ADE20K
Metrics:
mIoU: 42.03
mIoU(ms+flip): 43.04
Task: Semantic Segmentation
Config: configs/nonlocal_net/nonlocal_r50-d8_512x512_160k_ade20k.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/nonlocal_net/nonlocal_r50-d8_512x512_160k_ade20k/nonlocal_r50-d8_512x512_160k_ade20k_20200616_005410-baef45e3.pth
- Config: configs/nonlocal_net/nonlocal_r101-d8_512x512_160k_ade20k.py
- Name: nonlocal_r101-d8_512x512_160k_ade20k
In Collection: nonlocal_net
Metadata:
backbone: R-101-D8
crop size: (512,512)
lr schd: 160000
Name: nonlocal_r101-d8_512x512_160k_ade20k
Results:
- Task: Semantic Segmentation
Dataset: ADE20K
Metrics:
mIoU: 44.63
mIoU(ms+flip): 45.79
Task: Semantic Segmentation
Config: configs/nonlocal_net/nonlocal_r101-d8_512x512_160k_ade20k.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/nonlocal_net/nonlocal_r101-d8_512x512_160k_ade20k/nonlocal_r101-d8_512x512_160k_ade20k_20210827_221502-7881aa1a.pth
- Config: configs/nonlocal_net/nonlocal_r50-d8_512x512_20k_voc12aug.py
- Name: nonlocal_r50-d8_512x512_20k_voc12aug
In Collection: nonlocal_net
Metadata:
backbone: R-50-D8
crop size: (512,512)
lr schd: 20000
inference time (ms/im):
- backend: PyTorch
batch size: 1
- value: 47.15
hardware: V100
backend: PyTorch
batch size: 1
mode: FP32
resolution: (512,512)
value: 47.15
lr schd: 20000
memory (GB): 6.4
Name: nonlocal_r50-d8_512x512_20k_voc12aug
Results:
- Task: Semantic Segmentation
Dataset: Pascal VOC 2012 + Aug
Metrics:
mIoU: 76.2
mIoU(ms+flip): 77.12
Task: Semantic Segmentation
Config: configs/nonlocal_net/nonlocal_r50-d8_512x512_20k_voc12aug.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/nonlocal_net/nonlocal_r50-d8_512x512_20k_voc12aug/nonlocal_r50-d8_512x512_20k_voc12aug_20200617_222613-07f2a57c.pth
- Config: configs/nonlocal_net/nonlocal_r101-d8_512x512_20k_voc12aug.py
- Name: nonlocal_r101-d8_512x512_20k_voc12aug
In Collection: nonlocal_net
Metadata:
backbone: R-101-D8
crop size: (512,512)
lr schd: 20000
inference time (ms/im):
- backend: PyTorch
batch size: 1
- value: 71.38
hardware: V100
backend: PyTorch
batch size: 1
mode: FP32
resolution: (512,512)
value: 71.38
lr schd: 20000
memory (GB): 9.8
Name: nonlocal_r101-d8_512x512_20k_voc12aug
Results:
- Task: Semantic Segmentation
Dataset: Pascal VOC 2012 + Aug
Metrics:
mIoU: 78.15
mIoU(ms+flip): 78.86
Task: Semantic Segmentation
Config: configs/nonlocal_net/nonlocal_r101-d8_512x512_20k_voc12aug.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/nonlocal_net/nonlocal_r101-d8_512x512_20k_voc12aug/nonlocal_r101-d8_512x512_20k_voc12aug_20200617_222615-948c68ab.pth
- Config: configs/nonlocal_net/nonlocal_r50-d8_512x512_40k_voc12aug.py
- Name: nonlocal_r50-d8_512x512_40k_voc12aug
In Collection: nonlocal_net
Metadata:
backbone: R-50-D8
crop size: (512,512)
lr schd: 40000
Name: nonlocal_r50-d8_512x512_40k_voc12aug
Results:
- Task: Semantic Segmentation
Dataset: Pascal VOC 2012 + Aug
Metrics:
mIoU: 76.65
mIoU(ms+flip): 77.47
Task: Semantic Segmentation
Config: configs/nonlocal_net/nonlocal_r50-d8_512x512_40k_voc12aug.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/nonlocal_net/nonlocal_r50-d8_512x512_40k_voc12aug/nonlocal_r50-d8_512x512_40k_voc12aug_20200614_000028-0139d4a9.pth
- Config: configs/nonlocal_net/nonlocal_r101-d8_512x512_40k_voc12aug.py
- Name: nonlocal_r101-d8_512x512_40k_voc12aug
In Collection: nonlocal_net
Metadata:
backbone: R-101-D8
crop size: (512,512)
lr schd: 40000
Name: nonlocal_r101-d8_512x512_40k_voc12aug
Results:
- Task: Semantic Segmentation
Dataset: Pascal VOC 2012 + Aug
Metrics:
mIoU: 78.27
mIoU(ms+flip): 79.12
Task: Semantic Segmentation
Config: configs/nonlocal_net/nonlocal_r101-d8_512x512_40k_voc12aug.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/nonlocal_net/nonlocal_r101-d8_512x512_40k_voc12aug/nonlocal_r101-d8_512x512_40k_voc12aug_20200614_000028-7e5ff470.pth

9
configs/ocrnet/README.md
View File

@ -4,6 +4,13 @@
<!-- [ALGORITHM] -->
<a href="https://github.com/openseg-group/OCNet.pytorch">Official Repo</a>
<a href="https://github.com/open-mmlab/mmsegmentation/blob/v0.17.0/mmseg/models/decode_heads/ocr_head.py#L86">Code Snippet</a>
<details>
<summary align="right"><a href="https://arxiv.org/abs/1909.11065">OCRNet (ECCV'2020)</a></summary>
```latex
@article{YuanW18,
title={Ocnet: Object context network for scene parsing},
@ -20,6 +27,8 @@
}
```
</details>
## Results and models
### Cityscapes

271
configs/ocrnet/ocrnet.yml
View File

@ -1,431 +1,438 @@
Collections:
- Metadata:
- Name: ocrnet
Metadata:
Training Data:
- Cityscapes
- ' HRNet backbone'
- ' ResNet backbone'
- ADE20K
- Pascal VOC 2012 + Aug
Name: ocrnet
Paper:
URL: https://arxiv.org/abs/1909.11065
Title: Object-Contextual Representations for Semantic Segmentation
README: configs/ocrnet/README.md
Code:
URL: https://github.com/open-mmlab/mmsegmentation/blob/v0.17.0/mmseg/models/decode_heads/ocr_head.py#L86
Version: v0.17.0
Converted From:
Code: https://github.com/openseg-group/OCNet.pytorch
Models:
- Config: configs/ocrnet/ocrnet_hr18s_512x1024_40k_cityscapes.py
- Name: ocrnet_hr18s_512x1024_40k_cityscapes
In Collection: ocrnet
Metadata:
backbone: HRNetV2p-W18-Small
crop size: (512,1024)
lr schd: 40000
inference time (ms/im):
- backend: PyTorch
batch size: 1
- value: 95.69
hardware: V100
backend: PyTorch
batch size: 1
mode: FP32
resolution: (512,1024)
value: 95.69
lr schd: 40000
memory (GB): 3.5
Name: ocrnet_hr18s_512x1024_40k_cityscapes
Results:
Dataset: ' HRNet backbone'
- Task: Semantic Segmentation
Dataset: Cityscapes
Metrics:
mIoU: 74.3
mIoU(ms+flip): 75.95
Task: Semantic Segmentation
Config: configs/ocrnet/ocrnet_hr18s_512x1024_40k_cityscapes.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/ocrnet/ocrnet_hr18s_512x1024_40k_cityscapes/ocrnet_hr18s_512x1024_40k_cityscapes_20200601_033304-fa2436c2.pth
- Config: configs/ocrnet/ocrnet_hr18_512x1024_40k_cityscapes.py
- Name: ocrnet_hr18_512x1024_40k_cityscapes
In Collection: ocrnet
Metadata:
backbone: HRNetV2p-W18
crop size: (512,1024)
lr schd: 40000
inference time (ms/im):
- backend: PyTorch
batch size: 1
- value: 133.33
hardware: V100
backend: PyTorch
batch size: 1
mode: FP32
resolution: (512,1024)
value: 133.33
lr schd: 40000
memory (GB): 4.7
Name: ocrnet_hr18_512x1024_40k_cityscapes
Results:
Dataset: ' HRNet backbone'
- Task: Semantic Segmentation
Dataset: Cityscapes
Metrics:
mIoU: 77.72
mIoU(ms+flip): 79.49
Task: Semantic Segmentation
Config: configs/ocrnet/ocrnet_hr18_512x1024_40k_cityscapes.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/ocrnet/ocrnet_hr18_512x1024_40k_cityscapes/ocrnet_hr18_512x1024_40k_cityscapes_20200601_033320-401c5bdd.pth
- Config: configs/ocrnet/ocrnet_hr48_512x1024_40k_cityscapes.py
- Name: ocrnet_hr48_512x1024_40k_cityscapes
In Collection: ocrnet
Metadata:
backbone: HRNetV2p-W48
crop size: (512,1024)
lr schd: 40000
inference time (ms/im):
- backend: PyTorch
batch size: 1
- value: 236.97
hardware: V100
backend: PyTorch
batch size: 1
mode: FP32
resolution: (512,1024)
value: 236.97
lr schd: 40000
memory (GB): 8.0
Name: ocrnet_hr48_512x1024_40k_cityscapes
Results:
Dataset: ' HRNet backbone'
- Task: Semantic Segmentation
Dataset: Cityscapes
Metrics:
mIoU: 80.58
mIoU(ms+flip): 81.79
Task: Semantic Segmentation
Config: configs/ocrnet/ocrnet_hr48_512x1024_40k_cityscapes.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/ocrnet/ocrnet_hr48_512x1024_40k_cityscapes/ocrnet_hr48_512x1024_40k_cityscapes_20200601_033336-55b32491.pth
- Config: configs/ocrnet/ocrnet_hr18s_512x1024_80k_cityscapes.py
- Name: ocrnet_hr18s_512x1024_80k_cityscapes
In Collection: ocrnet
Metadata:
backbone: HRNetV2p-W18-Small
crop size: (512,1024)
lr schd: 80000
Name: ocrnet_hr18s_512x1024_80k_cityscapes
Results:
Dataset: ' HRNet backbone'
- Task: Semantic Segmentation
Dataset: Cityscapes
Metrics:
mIoU: 77.16
mIoU(ms+flip): 78.66
Task: Semantic Segmentation
Config: configs/ocrnet/ocrnet_hr18s_512x1024_80k_cityscapes.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/ocrnet/ocrnet_hr18s_512x1024_80k_cityscapes/ocrnet_hr18s_512x1024_80k_cityscapes_20200601_222735-55979e63.pth
- Config: configs/ocrnet/ocrnet_hr18_512x1024_80k_cityscapes.py
- Name: ocrnet_hr18_512x1024_80k_cityscapes
In Collection: ocrnet
Metadata:
backbone: HRNetV2p-W18
crop size: (512,1024)
lr schd: 80000
Name: ocrnet_hr18_512x1024_80k_cityscapes
Results:
Dataset: ' HRNet backbone'
- Task: Semantic Segmentation
Dataset: Cityscapes
Metrics:
mIoU: 78.57
mIoU(ms+flip): 80.46
Task: Semantic Segmentation
Config: configs/ocrnet/ocrnet_hr18_512x1024_80k_cityscapes.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/ocrnet/ocrnet_hr18_512x1024_80k_cityscapes/ocrnet_hr18_512x1024_80k_cityscapes_20200614_230521-c2e1dd4a.pth
- Config: configs/ocrnet/ocrnet_hr48_512x1024_80k_cityscapes.py
- Name: ocrnet_hr48_512x1024_80k_cityscapes
In Collection: ocrnet
Metadata:
backbone: HRNetV2p-W48
crop size: (512,1024)
lr schd: 80000
Name: ocrnet_hr48_512x1024_80k_cityscapes
Results:
Dataset: ' HRNet backbone'
- Task: Semantic Segmentation
Dataset: Cityscapes
Metrics:
mIoU: 80.7
mIoU(ms+flip): 81.87
Task: Semantic Segmentation
Config: configs/ocrnet/ocrnet_hr48_512x1024_80k_cityscapes.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/ocrnet/ocrnet_hr48_512x1024_80k_cityscapes/ocrnet_hr48_512x1024_80k_cityscapes_20200601_222752-9076bcdf.pth
- Config: configs/ocrnet/ocrnet_hr18s_512x1024_160k_cityscapes.py
- Name: ocrnet_hr18s_512x1024_160k_cityscapes
In Collection: ocrnet
Metadata:
backbone: HRNetV2p-W18-Small
crop size: (512,1024)
lr schd: 160000
Name: ocrnet_hr18s_512x1024_160k_cityscapes
Results:
Dataset: ' HRNet backbone'
- Task: Semantic Segmentation
Dataset: Cityscapes
Metrics:
mIoU: 78.45
mIoU(ms+flip): 79.97
Task: Semantic Segmentation
Config: configs/ocrnet/ocrnet_hr18s_512x1024_160k_cityscapes.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/ocrnet/ocrnet_hr18s_512x1024_160k_cityscapes/ocrnet_hr18s_512x1024_160k_cityscapes_20200602_191005-f4a7af28.pth
- Config: configs/ocrnet/ocrnet_hr18_512x1024_160k_cityscapes.py
- Name: ocrnet_hr18_512x1024_160k_cityscapes
In Collection: ocrnet
Metadata:
backbone: HRNetV2p-W18
crop size: (512,1024)
lr schd: 160000
Name: ocrnet_hr18_512x1024_160k_cityscapes
Results:
Dataset: ' HRNet backbone'
- Task: Semantic Segmentation
Dataset: Cityscapes
Metrics:
mIoU: 79.47
mIoU(ms+flip): 80.91
Task: Semantic Segmentation
Config: configs/ocrnet/ocrnet_hr18_512x1024_160k_cityscapes.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/ocrnet/ocrnet_hr18_512x1024_160k_cityscapes/ocrnet_hr18_512x1024_160k_cityscapes_20200602_191001-b9172d0c.pth
- Config: configs/ocrnet/ocrnet_hr48_512x1024_160k_cityscapes.py
- Name: ocrnet_hr48_512x1024_160k_cityscapes
In Collection: ocrnet
Metadata:
backbone: HRNetV2p-W48
crop size: (512,1024)
lr schd: 160000
Name: ocrnet_hr48_512x1024_160k_cityscapes
Results:
Dataset: ' HRNet backbone'
- Task: Semantic Segmentation
Dataset: Cityscapes
Metrics:
mIoU: 81.35
mIoU(ms+flip): 82.7
Task: Semantic Segmentation
Config: configs/ocrnet/ocrnet_hr48_512x1024_160k_cityscapes.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/ocrnet/ocrnet_hr48_512x1024_160k_cityscapes/ocrnet_hr48_512x1024_160k_cityscapes_20200602_191037-dfbf1b0c.pth
- Config: configs/ocrnet/ocrnet_r101-d8_512x1024_40k_b8_cityscapes.py
- Name: ocrnet_r101-d8_512x1024_40k_b8_cityscapes
In Collection: ocrnet
Metadata:
backbone: R-101-D8
crop size: (512,1024)
lr schd: 40000
Name: ocrnet_r101-d8_512x1024_40k_b8_cityscapes
Results:
Dataset: ' ResNet backbone'
- Task: Semantic Segmentation
Dataset: Cityscapes
Metrics:
mIoU: 80.09
Task: Semantic Segmentation
Config: configs/ocrnet/ocrnet_r101-d8_512x1024_40k_b8_cityscapes.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/ocrnet/ocrnet_r101-d8_512x1024_40k_b8_cityscapes/ocrnet_r101-d8_512x1024_40k_b8_cityscapes_20200717_110721-02ac0f13.pth
- Config: configs/ocrnet/ocrnet_r101-d8_512x1024_40k_b16_cityscapes.py
- Name: ocrnet_r101-d8_512x1024_40k_b16_cityscapes
In Collection: ocrnet
Metadata:
backbone: R-101-D8
crop size: (512,1024)
lr schd: 40000
inference time (ms/im):
- backend: PyTorch
batch size: 1
- value: 331.13
hardware: V100
backend: PyTorch
batch size: 1
mode: FP32
resolution: (512,1024)
value: 331.13
lr schd: 40000
memory (GB): 8.8
Name: ocrnet_r101-d8_512x1024_40k_b16_cityscapes
Results:
Dataset: ' ResNet backbone'
- Task: Semantic Segmentation
Dataset: Cityscapes
Metrics:
mIoU: 80.3
Task: Semantic Segmentation
Config: configs/ocrnet/ocrnet_r101-d8_512x1024_40k_b16_cityscapes.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/ocrnet/ocrnet_r101-d8_512x1024_40k_b16_cityscapes/ocrnet_r101-d8_512x1024_40k_b16_cityscapes_20200723_193726-db500f80.pth
- Config: configs/ocrnet/ocrnet_r101-d8_512x1024_80k_b16_cityscapes.py
- Name: ocrnet_r101-d8_512x1024_80k_b16_cityscapes
In Collection: ocrnet
Metadata:
backbone: R-101-D8
crop size: (512,1024)
lr schd: 80000
inference time (ms/im):
- backend: PyTorch
batch size: 1
- value: 331.13
hardware: V100
backend: PyTorch
batch size: 1
mode: FP32
resolution: (512,1024)
value: 331.13
lr schd: 80000
memory (GB): 8.8
Name: ocrnet_r101-d8_512x1024_80k_b16_cityscapes
Results:
Dataset: ' ResNet backbone'
- Task: Semantic Segmentation
Dataset: Cityscapes
Metrics:
mIoU: 80.81
Task: Semantic Segmentation
Config: configs/ocrnet/ocrnet_r101-d8_512x1024_80k_b16_cityscapes.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/ocrnet/ocrnet_r101-d8_512x1024_80k_b16_cityscapes/ocrnet_r101-d8_512x1024_80k_b16_cityscapes_20200723_192421-78688424.pth
- Config: configs/ocrnet/ocrnet_hr18s_512x512_80k_ade20k.py
- Name: ocrnet_hr18s_512x512_80k_ade20k
In Collection: ocrnet
Metadata:
backbone: HRNetV2p-W18-Small
crop size: (512,512)
lr schd: 80000
inference time (ms/im):
- backend: PyTorch
batch size: 1
- value: 34.51
hardware: V100
backend: PyTorch
batch size: 1
mode: FP32
resolution: (512,512)
value: 34.51
lr schd: 80000
memory (GB): 6.7
Name: ocrnet_hr18s_512x512_80k_ade20k
Results:
- Task: Semantic Segmentation
Dataset: ADE20K
Metrics:
mIoU: 35.06
mIoU(ms+flip): 35.8
Task: Semantic Segmentation
Config: configs/ocrnet/ocrnet_hr18s_512x512_80k_ade20k.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/ocrnet/ocrnet_hr18s_512x512_80k_ade20k/ocrnet_hr18s_512x512_80k_ade20k_20200615_055600-e80b62af.pth
- Config: configs/ocrnet/ocrnet_hr18_512x512_80k_ade20k.py
- Name: ocrnet_hr18_512x512_80k_ade20k
In Collection: ocrnet
Metadata:
backbone: HRNetV2p-W18
crop size: (512,512)
lr schd: 80000
inference time (ms/im):
- backend: PyTorch
batch size: 1
- value: 52.83
hardware: V100
backend: PyTorch
batch size: 1
mode: FP32
resolution: (512,512)
value: 52.83
lr schd: 80000
memory (GB): 7.9
Name: ocrnet_hr18_512x512_80k_ade20k
Results:
- Task: Semantic Segmentation
Dataset: ADE20K
Metrics:
mIoU: 37.79
mIoU(ms+flip): 39.16
Task: Semantic Segmentation
Config: configs/ocrnet/ocrnet_hr18_512x512_80k_ade20k.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/ocrnet/ocrnet_hr18_512x512_80k_ade20k/ocrnet_hr18_512x512_80k_ade20k_20200615_053157-d173d83b.pth
- Config: configs/ocrnet/ocrnet_hr48_512x512_80k_ade20k.py
- Name: ocrnet_hr48_512x512_80k_ade20k
In Collection: ocrnet
Metadata:
backbone: HRNetV2p-W48
crop size: (512,512)
lr schd: 80000
inference time (ms/im):
- backend: PyTorch
batch size: 1
- value: 58.86
hardware: V100
backend: PyTorch
batch size: 1
mode: FP32
resolution: (512,512)
value: 58.86
lr schd: 80000
memory (GB): 11.2
Name: ocrnet_hr48_512x512_80k_ade20k
Results:
- Task: Semantic Segmentation
Dataset: ADE20K
Metrics:
mIoU: 43.0
mIoU(ms+flip): 44.3
Task: Semantic Segmentation
Config: configs/ocrnet/ocrnet_hr48_512x512_80k_ade20k.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/ocrnet/ocrnet_hr48_512x512_80k_ade20k/ocrnet_hr48_512x512_80k_ade20k_20200615_021518-d168c2d1.pth
- Config: configs/ocrnet/ocrnet_hr18s_512x512_160k_ade20k.py
- Name: ocrnet_hr18s_512x512_160k_ade20k
In Collection: ocrnet
Metadata:
backbone: HRNetV2p-W18-Small
crop size: (512,512)
lr schd: 160000
Name: ocrnet_hr18s_512x512_160k_ade20k
Results:
- Task: Semantic Segmentation
Dataset: ADE20K
Metrics:
mIoU: 37.19
mIoU(ms+flip): 38.4
Task: Semantic Segmentation
Config: configs/ocrnet/ocrnet_hr18s_512x512_160k_ade20k.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/ocrnet/ocrnet_hr18s_512x512_160k_ade20k/ocrnet_hr18s_512x512_160k_ade20k_20200615_184505-8e913058.pth
- Config: configs/ocrnet/ocrnet_hr18_512x512_160k_ade20k.py
- Name: ocrnet_hr18_512x512_160k_ade20k
In Collection: ocrnet
Metadata:
backbone: HRNetV2p-W18
crop size: (512,512)
lr schd: 160000
Name: ocrnet_hr18_512x512_160k_ade20k
Results:
- Task: Semantic Segmentation
Dataset: ADE20K
Metrics:
mIoU: 39.32
mIoU(ms+flip): 40.8
Task: Semantic Segmentation
Config: configs/ocrnet/ocrnet_hr18_512x512_160k_ade20k.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/ocrnet/ocrnet_hr18_512x512_160k_ade20k/ocrnet_hr18_512x512_160k_ade20k_20200615_200940-d8fcd9d1.pth
- Config: configs/ocrnet/ocrnet_hr48_512x512_160k_ade20k.py
- Name: ocrnet_hr48_512x512_160k_ade20k
In Collection: ocrnet
Metadata:
backbone: HRNetV2p-W48
crop size: (512,512)
lr schd: 160000
Name: ocrnet_hr48_512x512_160k_ade20k
Results:
- Task: Semantic Segmentation
Dataset: ADE20K
Metrics:
mIoU: 43.25
mIoU(ms+flip): 44.88
Task: Semantic Segmentation
Config: configs/ocrnet/ocrnet_hr48_512x512_160k_ade20k.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/ocrnet/ocrnet_hr48_512x512_160k_ade20k/ocrnet_hr48_512x512_160k_ade20k_20200615_184705-a073726d.pth
- Config: configs/ocrnet/ocrnet_hr18s_512x512_20k_voc12aug.py
- Name: ocrnet_hr18s_512x512_20k_voc12aug
In Collection: ocrnet
Metadata:
backbone: HRNetV2p-W18-Small
crop size: (512,512)
lr schd: 20000
inference time (ms/im):
- backend: PyTorch
batch size: 1
- value: 31.7
hardware: V100
backend: PyTorch
batch size: 1
mode: FP32
resolution: (512,512)
value: 31.7
lr schd: 20000
memory (GB): 3.5
Name: ocrnet_hr18s_512x512_20k_voc12aug
Results:
- Task: Semantic Segmentation
Dataset: Pascal VOC 2012 + Aug
Metrics:
mIoU: 71.7
mIoU(ms+flip): 73.84
Task: Semantic Segmentation
Config: configs/ocrnet/ocrnet_hr18s_512x512_20k_voc12aug.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/ocrnet/ocrnet_hr18s_512x512_20k_voc12aug/ocrnet_hr18s_512x512_20k_voc12aug_20200617_233913-02b04fcb.pth
- Config: configs/ocrnet/ocrnet_hr18_512x512_20k_voc12aug.py
- Name: ocrnet_hr18_512x512_20k_voc12aug
In Collection: ocrnet
Metadata:
backbone: HRNetV2p-W18
crop size: (512,512)
lr schd: 20000
inference time (ms/im):
- backend: PyTorch
batch size: 1
- value: 50.23
hardware: V100
backend: PyTorch
batch size: 1
mode: FP32
resolution: (512,512)
value: 50.23
lr schd: 20000
memory (GB): 4.7
Name: ocrnet_hr18_512x512_20k_voc12aug
Results:
- Task: Semantic Segmentation
Dataset: Pascal VOC 2012 + Aug
Metrics:
mIoU: 74.75
mIoU(ms+flip): 77.11
Task: Semantic Segmentation
Config: configs/ocrnet/ocrnet_hr18_512x512_20k_voc12aug.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/ocrnet/ocrnet_hr18_512x512_20k_voc12aug/ocrnet_hr18_512x512_20k_voc12aug_20200617_233932-8954cbb7.pth
- Config: configs/ocrnet/ocrnet_hr48_512x512_20k_voc12aug.py
- Name: ocrnet_hr48_512x512_20k_voc12aug
In Collection: ocrnet
Metadata:
backbone: HRNetV2p-W48
crop size: (512,512)
lr schd: 20000
inference time (ms/im):
- backend: PyTorch
batch size: 1
- value: 56.09
hardware: V100
backend: PyTorch
batch size: 1
mode: FP32
resolution: (512,512)
value: 56.09
lr schd: 20000
memory (GB): 8.1
Name: ocrnet_hr48_512x512_20k_voc12aug
Results:
- Task: Semantic Segmentation
Dataset: Pascal VOC 2012 + Aug
Metrics:
mIoU: 77.72
mIoU(ms+flip): 79.87
Task: Semantic Segmentation
Config: configs/ocrnet/ocrnet_hr48_512x512_20k_voc12aug.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/ocrnet/ocrnet_hr48_512x512_20k_voc12aug/ocrnet_hr48_512x512_20k_voc12aug_20200617_233932-9e82080a.pth
- Config: configs/ocrnet/ocrnet_hr18s_512x512_40k_voc12aug.py
- Name: ocrnet_hr18s_512x512_40k_voc12aug
In Collection: ocrnet
Metadata:
backbone: HRNetV2p-W18-Small
crop size: (512,512)
lr schd: 40000
Name: ocrnet_hr18s_512x512_40k_voc12aug
Results:
- Task: Semantic Segmentation
Dataset: Pascal VOC 2012 + Aug
Metrics:
mIoU: 72.76
mIoU(ms+flip): 74.6
Task: Semantic Segmentation
Config: configs/ocrnet/ocrnet_hr18s_512x512_40k_voc12aug.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/ocrnet/ocrnet_hr18s_512x512_40k_voc12aug/ocrnet_hr18s_512x512_40k_voc12aug_20200614_002025-42b587ac.pth
- Config: configs/ocrnet/ocrnet_hr18_512x512_40k_voc12aug.py
- Name: ocrnet_hr18_512x512_40k_voc12aug
In Collection: ocrnet
Metadata:
backbone: HRNetV2p-W18
crop size: (512,512)
lr schd: 40000
Name: ocrnet_hr18_512x512_40k_voc12aug
Results:
- Task: Semantic Segmentation
Dataset: Pascal VOC 2012 + Aug
Metrics:
mIoU: 74.98
mIoU(ms+flip): 77.4
Task: Semantic Segmentation
Config: configs/ocrnet/ocrnet_hr18_512x512_40k_voc12aug.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/ocrnet/ocrnet_hr18_512x512_40k_voc12aug/ocrnet_hr18_512x512_40k_voc12aug_20200614_015958-714302be.pth
- Config: configs/ocrnet/ocrnet_hr48_512x512_40k_voc12aug.py
- Name: ocrnet_hr48_512x512_40k_voc12aug
In Collection: ocrnet
Metadata:
backbone: HRNetV2p-W48
crop size: (512,512)
lr schd: 40000
Name: ocrnet_hr48_512x512_40k_voc12aug
Results:
- Task: Semantic Segmentation
Dataset: Pascal VOC 2012 + Aug
Metrics:
mIoU: 77.14
mIoU(ms+flip): 79.71
Task: Semantic Segmentation
Config: configs/ocrnet/ocrnet_hr48_512x512_40k_voc12aug.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/ocrnet/ocrnet_hr48_512x512_40k_voc12aug/ocrnet_hr48_512x512_40k_voc12aug_20200614_015958-255bc5ce.pth

9
configs/point_rend/README.md
View File

@ -4,6 +4,13 @@
<!-- [ALGORITHM] -->
<a href="https://github.com/facebookresearch/detectron2/tree/master/projects/PointRend">Official Repo</a>
<a href="https://github.com/open-mmlab/mmsegmentation/blob/v0.17.0/mmseg/models/decode_heads/point_head.py#L36">Code Snippet</a>
<details>
<summary align="right"><a href="https://arxiv.org/abs/1912.08193">PointRend (CVPR'2020)</a></summary>
```
@inproceedings{kirillov2020pointrend,
title={Pointrend: Image segmentation as rendering},
@ -14,6 +21,8 @@
}
```
</details>
## Results and models
### Cityscapes

69
configs/point_rend/point_rend.yml
View File

@ -1,95 +1,104 @@
Collections:
- Metadata:
- Name: point_rend
Metadata:
Training Data:
- Cityscapes
- ADE20K
Name: point_rend
Paper:
URL: https://arxiv.org/abs/1912.08193
Title: 'PointRend: Image Segmentation as Rendering'
README: configs/point_rend/README.md
Code:
URL: https://github.com/open-mmlab/mmsegmentation/blob/v0.17.0/mmseg/models/decode_heads/point_head.py#L36
Version: v0.17.0
Converted From:
Code: https://github.com/facebookresearch/detectron2/tree/master/projects/PointRend
Models:
- Config: configs/point_rend/pointrend_r50_512x1024_80k_cityscapes.py
- Name: pointrend_r50_512x1024_80k_cityscapes
In Collection: point_rend
Metadata:
backbone: R-50
crop size: (512,1024)
lr schd: 80000
inference time (ms/im):
- backend: PyTorch
batch size: 1
- value: 117.92
hardware: V100
backend: PyTorch
batch size: 1
mode: FP32
resolution: (512,1024)
value: 117.92
lr schd: 80000
memory (GB): 3.1
Name: pointrend_r50_512x1024_80k_cityscapes
Results:
- Task: Semantic Segmentation
Dataset: Cityscapes
Metrics:
mIoU: 76.47
mIoU(ms+flip): 78.13
Task: Semantic Segmentation
Config: configs/point_rend/pointrend_r50_512x1024_80k_cityscapes.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/point_rend/pointrend_r50_512x1024_80k_cityscapes/pointrend_r50_512x1024_80k_cityscapes_20200711_015821-bb1ff523.pth
- Config: configs/point_rend/pointrend_r101_512x1024_80k_cityscapes.py
- Name: pointrend_r101_512x1024_80k_cityscapes
In Collection: point_rend
Metadata:
backbone: R-101
crop size: (512,1024)
lr schd: 80000
inference time (ms/im):
- backend: PyTorch
batch size: 1
- value: 142.86
hardware: V100
backend: PyTorch
batch size: 1
mode: FP32
resolution: (512,1024)
value: 142.86
lr schd: 80000
memory (GB): 4.2
Name: pointrend_r101_512x1024_80k_cityscapes
Results:
- Task: Semantic Segmentation
Dataset: Cityscapes
Metrics:
mIoU: 78.3
mIoU(ms+flip): 79.97
Task: Semantic Segmentation
Config: configs/point_rend/pointrend_r101_512x1024_80k_cityscapes.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/point_rend/pointrend_r101_512x1024_80k_cityscapes/pointrend_r101_512x1024_80k_cityscapes_20200711_170850-d0ca84be.pth
- Config: configs/point_rend/pointrend_r50_512x512_160k_ade20k.py
- Name: pointrend_r50_512x512_160k_ade20k
In Collection: point_rend
Metadata:
backbone: R-50
crop size: (512,512)
lr schd: 160000
inference time (ms/im):
- backend: PyTorch
batch size: 1
- value: 57.77
hardware: V100
backend: PyTorch
batch size: 1
mode: FP32
resolution: (512,512)
value: 57.77
lr schd: 160000
memory (GB): 5.1
Name: pointrend_r50_512x512_160k_ade20k
Results:
- Task: Semantic Segmentation
Dataset: ADE20K
Metrics:
mIoU: 37.64
mIoU(ms+flip): 39.17
Task: Semantic Segmentation
Config: configs/point_rend/pointrend_r50_512x512_160k_ade20k.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/point_rend/pointrend_r50_512x512_160k_ade20k/pointrend_r50_512x512_160k_ade20k_20200807_232644-ac3febf2.pth
- Config: configs/point_rend/pointrend_r101_512x512_160k_ade20k.py
- Name: pointrend_r101_512x512_160k_ade20k
In Collection: point_rend
Metadata:
backbone: R-101
crop size: (512,512)
lr schd: 160000
inference time (ms/im):
- backend: PyTorch
batch size: 1
- value: 64.52
hardware: V100
backend: PyTorch
batch size: 1
mode: FP32
resolution: (512,512)
value: 64.52
lr schd: 160000
memory (GB): 6.1
Name: pointrend_r101_512x512_160k_ade20k
Results:
- Task: Semantic Segmentation
Dataset: ADE20K
Metrics:
mIoU: 40.02
mIoU(ms+flip): 41.6
Task: Semantic Segmentation
Config: configs/point_rend/pointrend_r101_512x512_160k_ade20k.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/point_rend/pointrend_r101_512x512_160k_ade20k/pointrend_r101_512x512_160k_ade20k_20200808_030852-8834902a.pth

9
configs/psanet/README.md
View File

@ -4,6 +4,13 @@
<!-- [ALGORITHM] -->
<a href="https://github.com/hszhao/PSANet">Official Repo</a>
<a href="https://github.com/open-mmlab/mmsegmentation/blob/v0.17.0/mmseg/models/decode_heads/psa_head.py#L18">Code Snippet</a>
<details>
<summary align="right"><a href="https://openaccess.thecvf.com/content_ECCV_2018/papers/Hengshuang_Zhao_PSANet_Point-wise_Spatial_ECCV_2018_paper.pdf">PSANet (ECCV'2018)</a></summary>
```latex
@inproceedings{zhao2018psanet,
title={Psanet: Point-wise spatial attention network for scene parsing},
@ -14,6 +21,8 @@
}
```
</details>
## Results and models
### Cityscapes

173
configs/psanet/psanet.yml
View File

@ -1,296 +1,305 @@
Collections:
- Metadata:
- Name: psanet
Metadata:
Training Data:
- Cityscapes
- ADE20K
- Pascal VOC 2012 + Aug
Name: psanet
Paper:
URL: https://openaccess.thecvf.com/content_ECCV_2018/papers/Hengshuang_Zhao_PSANet_Point-wise_Spatial_ECCV_2018_paper.pdf
Title: 'PSANet: Point-wise Spatial Attention Network for Scene Parsing'
README: configs/psanet/README.md
Code:
URL: https://github.com/open-mmlab/mmsegmentation/blob/v0.17.0/mmseg/models/decode_heads/psa_head.py#L18
Version: v0.17.0
Converted From:
Code: https://github.com/hszhao/PSANet
Models:
- Config: configs/psanet/psanet_r50-d8_512x1024_40k_cityscapes.py
- Name: psanet_r50-d8_512x1024_40k_cityscapes
In Collection: psanet
Metadata:
backbone: R-50-D8
crop size: (512,1024)
lr schd: 40000
inference time (ms/im):
- backend: PyTorch
batch size: 1
- value: 315.46
hardware: V100
backend: PyTorch
batch size: 1
mode: FP32
resolution: (512,1024)
value: 315.46
lr schd: 40000
memory (GB): 7.0
Name: psanet_r50-d8_512x1024_40k_cityscapes
Results:
- Task: Semantic Segmentation
Dataset: Cityscapes
Metrics:
mIoU: 77.63
mIoU(ms+flip): 79.04
Task: Semantic Segmentation
Config: configs/psanet/psanet_r50-d8_512x1024_40k_cityscapes.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/psanet/psanet_r50-d8_512x1024_40k_cityscapes/psanet_r50-d8_512x1024_40k_cityscapes_20200606_103117-99fac37c.pth
- Config: configs/psanet/psanet_r101-d8_512x1024_40k_cityscapes.py
- Name: psanet_r101-d8_512x1024_40k_cityscapes
In Collection: psanet
Metadata:
backbone: R-101-D8
crop size: (512,1024)
lr schd: 40000
inference time (ms/im):
- backend: PyTorch
batch size: 1
- value: 454.55
hardware: V100
backend: PyTorch
batch size: 1
mode: FP32
resolution: (512,1024)
value: 454.55
lr schd: 40000
memory (GB): 10.5
Name: psanet_r101-d8_512x1024_40k_cityscapes
Results:
- Task: Semantic Segmentation
Dataset: Cityscapes
Metrics:
mIoU: 79.14
mIoU(ms+flip): 80.19
Task: Semantic Segmentation
Config: configs/psanet/psanet_r101-d8_512x1024_40k_cityscapes.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/psanet/psanet_r101-d8_512x1024_40k_cityscapes/psanet_r101-d8_512x1024_40k_cityscapes_20200606_001418-27b9cfa7.pth
- Config: configs/psanet/psanet_r50-d8_769x769_40k_cityscapes.py
- Name: psanet_r50-d8_769x769_40k_cityscapes
In Collection: psanet
Metadata:
backbone: R-50-D8
crop size: (769,769)
lr schd: 40000
inference time (ms/im):
- backend: PyTorch
batch size: 1
- value: 714.29
hardware: V100
backend: PyTorch
batch size: 1
mode: FP32
resolution: (769,769)
value: 714.29
lr schd: 40000
memory (GB): 7.9
Name: psanet_r50-d8_769x769_40k_cityscapes
Results:
- Task: Semantic Segmentation
Dataset: Cityscapes
Metrics:
mIoU: 77.99
mIoU(ms+flip): 79.64
Task: Semantic Segmentation
Config: configs/psanet/psanet_r50-d8_769x769_40k_cityscapes.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/psanet/psanet_r50-d8_769x769_40k_cityscapes/psanet_r50-d8_769x769_40k_cityscapes_20200530_033717-d5365506.pth
- Config: configs/psanet/psanet_r101-d8_769x769_40k_cityscapes.py
- Name: psanet_r101-d8_769x769_40k_cityscapes
In Collection: psanet
Metadata:
backbone: R-101-D8
crop size: (769,769)
lr schd: 40000
inference time (ms/im):
- backend: PyTorch
batch size: 1
- value: 1020.41
hardware: V100
backend: PyTorch
batch size: 1
mode: FP32
resolution: (769,769)
value: 1020.41
lr schd: 40000
memory (GB): 11.9
Name: psanet_r101-d8_769x769_40k_cityscapes
Results:
- Task: Semantic Segmentation
Dataset: Cityscapes
Metrics:
mIoU: 78.43
mIoU(ms+flip): 80.26
Task: Semantic Segmentation
Config: configs/psanet/psanet_r101-d8_769x769_40k_cityscapes.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/psanet/psanet_r101-d8_769x769_40k_cityscapes/psanet_r101-d8_769x769_40k_cityscapes_20200530_035107-997da1e6.pth
- Config: configs/psanet/psanet_r50-d8_512x1024_80k_cityscapes.py
- Name: psanet_r50-d8_512x1024_80k_cityscapes
In Collection: psanet
Metadata:
backbone: R-50-D8
crop size: (512,1024)
lr schd: 80000
Name: psanet_r50-d8_512x1024_80k_cityscapes
Results:
- Task: Semantic Segmentation
Dataset: Cityscapes
Metrics:
mIoU: 77.24
mIoU(ms+flip): 78.69
Task: Semantic Segmentation
Config: configs/psanet/psanet_r50-d8_512x1024_80k_cityscapes.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/psanet/psanet_r50-d8_512x1024_80k_cityscapes/psanet_r50-d8_512x1024_80k_cityscapes_20200606_161842-ab60a24f.pth
- Config: configs/psanet/psanet_r101-d8_512x1024_80k_cityscapes.py
- Name: psanet_r101-d8_512x1024_80k_cityscapes
In Collection: psanet
Metadata:
backbone: R-101-D8
crop size: (512,1024)
lr schd: 80000
Name: psanet_r101-d8_512x1024_80k_cityscapes
Results:
- Task: Semantic Segmentation
Dataset: Cityscapes
Metrics:
mIoU: 79.31
mIoU(ms+flip): 80.53
Task: Semantic Segmentation
Config: configs/psanet/psanet_r101-d8_512x1024_80k_cityscapes.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/psanet/psanet_r101-d8_512x1024_80k_cityscapes/psanet_r101-d8_512x1024_80k_cityscapes_20200606_161823-0f73a169.pth
- Config: configs/psanet/psanet_r50-d8_769x769_80k_cityscapes.py
- Name: psanet_r50-d8_769x769_80k_cityscapes
In Collection: psanet
Metadata:
backbone: R-50-D8
crop size: (769,769)
lr schd: 80000
Name: psanet_r50-d8_769x769_80k_cityscapes
Results:
- Task: Semantic Segmentation
Dataset: Cityscapes
Metrics:
mIoU: 79.31
mIoU(ms+flip): 80.91
Task: Semantic Segmentation
Config: configs/psanet/psanet_r50-d8_769x769_80k_cityscapes.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/psanet/psanet_r50-d8_769x769_80k_cityscapes/psanet_r50-d8_769x769_80k_cityscapes_20200606_225134-fe42f49e.pth
- Config: configs/psanet/psanet_r101-d8_769x769_80k_cityscapes.py
- Name: psanet_r101-d8_769x769_80k_cityscapes
In Collection: psanet
Metadata:
backbone: R-101-D8
crop size: (769,769)
lr schd: 80000
Name: psanet_r101-d8_769x769_80k_cityscapes
Results:
- Task: Semantic Segmentation
Dataset: Cityscapes
Metrics:
mIoU: 79.69
mIoU(ms+flip): 80.89
Task: Semantic Segmentation
Config: configs/psanet/psanet_r101-d8_769x769_80k_cityscapes.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/psanet/psanet_r101-d8_769x769_80k_cityscapes/psanet_r101-d8_769x769_80k_cityscapes_20200606_214550-7665827b.pth
- Config: configs/psanet/psanet_r50-d8_512x512_80k_ade20k.py
- Name: psanet_r50-d8_512x512_80k_ade20k
In Collection: psanet
Metadata:
backbone: R-50-D8
crop size: (512,512)
lr schd: 80000
inference time (ms/im):
- backend: PyTorch
batch size: 1
- value: 52.88
hardware: V100
backend: PyTorch
batch size: 1
mode: FP32
resolution: (512,512)
value: 52.88
lr schd: 80000
memory (GB): 9.0
Name: psanet_r50-d8_512x512_80k_ade20k
Results:
- Task: Semantic Segmentation
Dataset: ADE20K
Metrics:
mIoU: 41.14
mIoU(ms+flip): 41.91
Task: Semantic Segmentation
Config: configs/psanet/psanet_r50-d8_512x512_80k_ade20k.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/psanet/psanet_r50-d8_512x512_80k_ade20k/psanet_r50-d8_512x512_80k_ade20k_20200614_144141-835e4b97.pth
- Config: configs/psanet/psanet_r101-d8_512x512_80k_ade20k.py
- Name: psanet_r101-d8_512x512_80k_ade20k
In Collection: psanet
Metadata:
backbone: R-101-D8
crop size: (512,512)
lr schd: 80000
inference time (ms/im):
- backend: PyTorch
batch size: 1
- value: 76.16
hardware: V100
backend: PyTorch
batch size: 1
mode: FP32
resolution: (512,512)
value: 76.16
lr schd: 80000
memory (GB): 12.5
Name: psanet_r101-d8_512x512_80k_ade20k
Results:
- Task: Semantic Segmentation
Dataset: ADE20K
Metrics:
mIoU: 43.8
mIoU(ms+flip): 44.75
Task: Semantic Segmentation
Config: configs/psanet/psanet_r101-d8_512x512_80k_ade20k.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/psanet/psanet_r101-d8_512x512_80k_ade20k/psanet_r101-d8_512x512_80k_ade20k_20200614_185117-1fab60d4.pth
- Config: configs/psanet/psanet_r50-d8_512x512_160k_ade20k.py
- Name: psanet_r50-d8_512x512_160k_ade20k
In Collection: psanet
Metadata:
backbone: R-50-D8
crop size: (512,512)
lr schd: 160000
Name: psanet_r50-d8_512x512_160k_ade20k
Results:
- Task: Semantic Segmentation
Dataset: ADE20K
Metrics:
mIoU: 41.67
mIoU(ms+flip): 42.95
Task: Semantic Segmentation
Config: configs/psanet/psanet_r50-d8_512x512_160k_ade20k.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/psanet/psanet_r50-d8_512x512_160k_ade20k/psanet_r50-d8_512x512_160k_ade20k_20200615_161258-148077dd.pth
- Config: configs/psanet/psanet_r101-d8_512x512_160k_ade20k.py
- Name: psanet_r101-d8_512x512_160k_ade20k
In Collection: psanet
Metadata:
backbone: R-101-D8
crop size: (512,512)
lr schd: 160000
Name: psanet_r101-d8_512x512_160k_ade20k
Results:
- Task: Semantic Segmentation
Dataset: ADE20K
Metrics:
mIoU: 43.74
mIoU(ms+flip): 45.38
Task: Semantic Segmentation
Config: configs/psanet/psanet_r101-d8_512x512_160k_ade20k.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/psanet/psanet_r101-d8_512x512_160k_ade20k/psanet_r101-d8_512x512_160k_ade20k_20200615_161537-dbfa564c.pth
- Config: configs/psanet/psanet_r50-d8_512x512_20k_voc12aug.py
- Name: psanet_r50-d8_512x512_20k_voc12aug
In Collection: psanet
Metadata:
backbone: R-50-D8
crop size: (512,512)
lr schd: 20000
inference time (ms/im):
- backend: PyTorch
batch size: 1
- value: 54.82
hardware: V100
backend: PyTorch
batch size: 1
mode: FP32
resolution: (512,512)
value: 54.82
lr schd: 20000
memory (GB): 6.9
Name: psanet_r50-d8_512x512_20k_voc12aug
Results:
- Task: Semantic Segmentation
Dataset: Pascal VOC 2012 + Aug
Metrics:
mIoU: 76.39
mIoU(ms+flip): 77.34
Task: Semantic Segmentation
Config: configs/psanet/psanet_r50-d8_512x512_20k_voc12aug.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/psanet/psanet_r50-d8_512x512_20k_voc12aug/psanet_r50-d8_512x512_20k_voc12aug_20200617_102413-2f1bbaa1.pth
- Config: configs/psanet/psanet_r101-d8_512x512_20k_voc12aug.py
- Name: psanet_r101-d8_512x512_20k_voc12aug
In Collection: psanet
Metadata:
backbone: R-101-D8
crop size: (512,512)
lr schd: 20000
inference time (ms/im):
- backend: PyTorch
batch size: 1
- value: 79.18
hardware: V100
backend: PyTorch
batch size: 1
mode: FP32
resolution: (512,512)
value: 79.18
lr schd: 20000
memory (GB): 10.4
Name: psanet_r101-d8_512x512_20k_voc12aug
Results:
- Task: Semantic Segmentation
Dataset: Pascal VOC 2012 + Aug
Metrics:
mIoU: 77.91
mIoU(ms+flip): 79.3
Task: Semantic Segmentation
Config: configs/psanet/psanet_r101-d8_512x512_20k_voc12aug.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/psanet/psanet_r101-d8_512x512_20k_voc12aug/psanet_r101-d8_512x512_20k_voc12aug_20200617_110624-946fef11.pth
- Config: configs/psanet/psanet_r50-d8_512x512_40k_voc12aug.py
- Name: psanet_r50-d8_512x512_40k_voc12aug
In Collection: psanet
Metadata:
backbone: R-50-D8
crop size: (512,512)
lr schd: 40000
Name: psanet_r50-d8_512x512_40k_voc12aug
Results:
- Task: Semantic Segmentation
Dataset: Pascal VOC 2012 + Aug
Metrics:
mIoU: 76.3
mIoU(ms+flip): 77.35
Task: Semantic Segmentation
Config: configs/psanet/psanet_r50-d8_512x512_40k_voc12aug.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/psanet/psanet_r50-d8_512x512_40k_voc12aug/psanet_r50-d8_512x512_40k_voc12aug_20200613_161946-f596afb5.pth
- Config: configs/psanet/psanet_r101-d8_512x512_40k_voc12aug.py
- Name: psanet_r101-d8_512x512_40k_voc12aug
In Collection: psanet
Metadata:
backbone: R-101-D8
crop size: (512,512)
lr schd: 40000
Name: psanet_r101-d8_512x512_40k_voc12aug
Results:
- Task: Semantic Segmentation
Dataset: Pascal VOC 2012 + Aug
Metrics:
mIoU: 77.73
mIoU(ms+flip): 79.05
Task: Semantic Segmentation
Config: configs/psanet/psanet_r101-d8_512x512_40k_voc12aug.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/psanet/psanet_r101-d8_512x512_40k_voc12aug/psanet_r101-d8_512x512_40k_voc12aug_20200613_161946-1f560f9e.pth

9
configs/pspnet/README.md
View File

@ -4,6 +4,13 @@
<!-- [ALGORITHM] -->
<a href="https://github.com/hszhao/PSPNet">Official Repo</a>
<a href="https://github.com/open-mmlab/mmsegmentation/blob/v0.17.0/mmseg/models/decode_heads/psp_head.py#L63">Code Snippet</a>
<details>
<summary align="right"><a href="https://arxiv.org/abs/1612.01105">PSPNet (CVPR'2017)</a></summary>
```latex
@inproceedings{zhao2017pspnet,
title={Pyramid Scene Parsing Network},
@ -13,6 +20,8 @@
}
```
</details>
## Results and models
### Cityscapes

409
configs/pspnet/pspnet.yml
View File

@ -1,5 +1,6 @@
Collections:
- Metadata:
- Name: pspnet
Metadata:
Training Data:
- Cityscapes
- ADE20K
@ -9,705 +10,713 @@ Collections:
- Dark Zurich and Nighttime Driving
- COCO-Stuff 10k
- COCO-Stuff 164k
Name: pspnet
Paper:
URL: https://arxiv.org/abs/1612.01105
Title: Pyramid Scene Parsing Network
README: configs/pspnet/README.md
Code:
URL: https://github.com/open-mmlab/mmsegmentation/blob/v0.17.0/mmseg/models/decode_heads/psp_head.py#L63
Version: v0.17.0
Converted From:
Code: https://github.com/hszhao/PSPNet
Models:
- Config: configs/pspnet/pspnet_r50-d8_512x1024_40k_cityscapes.py
- Name: pspnet_r50-d8_512x1024_40k_cityscapes
In Collection: pspnet
Metadata:
backbone: R-50-D8
crop size: (512,1024)
lr schd: 40000
inference time (ms/im):
- backend: PyTorch
batch size: 1
- value: 245.7
hardware: V100
backend: PyTorch
batch size: 1
mode: FP32
resolution: (512,1024)
value: 245.7
lr schd: 40000
memory (GB): 6.1
Name: pspnet_r50-d8_512x1024_40k_cityscapes
Results:
- Task: Semantic Segmentation
Dataset: Cityscapes
Metrics:
mIoU: 77.85
mIoU(ms+flip): 79.18
Task: Semantic Segmentation
Config: configs/pspnet/pspnet_r50-d8_512x1024_40k_cityscapes.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/pspnet/pspnet_r50-d8_512x1024_40k_cityscapes/pspnet_r50-d8_512x1024_40k_cityscapes_20200605_003338-2966598c.pth
- Config: configs/pspnet/pspnet_r101-d8_512x1024_40k_cityscapes.py
- Name: pspnet_r101-d8_512x1024_40k_cityscapes
In Collection: pspnet
Metadata:
backbone: R-101-D8
crop size: (512,1024)
lr schd: 40000
inference time (ms/im):
- backend: PyTorch
batch size: 1
- value: 373.13
hardware: V100
backend: PyTorch
batch size: 1
mode: FP32
resolution: (512,1024)
value: 373.13
lr schd: 40000
memory (GB): 9.6
Name: pspnet_r101-d8_512x1024_40k_cityscapes
Results:
- Task: Semantic Segmentation
Dataset: Cityscapes
Metrics:
mIoU: 78.34
mIoU(ms+flip): 79.74
Task: Semantic Segmentation
Config: configs/pspnet/pspnet_r101-d8_512x1024_40k_cityscapes.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/pspnet/pspnet_r101-d8_512x1024_40k_cityscapes/pspnet_r101-d8_512x1024_40k_cityscapes_20200604_232751-467e7cf4.pth
- Config: configs/pspnet/pspnet_r50-d8_769x769_40k_cityscapes.py
- Name: pspnet_r50-d8_769x769_40k_cityscapes
In Collection: pspnet
Metadata:
backbone: R-50-D8
crop size: (769,769)
lr schd: 40000
inference time (ms/im):
- backend: PyTorch
batch size: 1
- value: 568.18
hardware: V100
backend: PyTorch
batch size: 1
mode: FP32
resolution: (769,769)
value: 568.18
lr schd: 40000
memory (GB): 6.9
Name: pspnet_r50-d8_769x769_40k_cityscapes
Results:
- Task: Semantic Segmentation
Dataset: Cityscapes
Metrics:
mIoU: 78.26
mIoU(ms+flip): 79.88
Task: Semantic Segmentation
Config: configs/pspnet/pspnet_r50-d8_769x769_40k_cityscapes.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/pspnet/pspnet_r50-d8_769x769_40k_cityscapes/pspnet_r50-d8_769x769_40k_cityscapes_20200606_112725-86638686.pth
- Config: configs/pspnet/pspnet_r101-d8_769x769_40k_cityscapes.py
- Name: pspnet_r101-d8_769x769_40k_cityscapes
In Collection: pspnet
Metadata:
backbone: R-101-D8
crop size: (769,769)
lr schd: 40000
inference time (ms/im):
- backend: PyTorch
batch size: 1
- value: 869.57
hardware: V100
backend: PyTorch
batch size: 1
mode: FP32
resolution: (769,769)
value: 869.57
lr schd: 40000
memory (GB): 10.9
Name: pspnet_r101-d8_769x769_40k_cityscapes
Results:
- Task: Semantic Segmentation
Dataset: Cityscapes
Metrics:
mIoU: 79.08
mIoU(ms+flip): 80.28
Task: Semantic Segmentation
Config: configs/pspnet/pspnet_r101-d8_769x769_40k_cityscapes.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/pspnet/pspnet_r101-d8_769x769_40k_cityscapes/pspnet_r101-d8_769x769_40k_cityscapes_20200606_112753-61c6f5be.pth
- Config: configs/pspnet/pspnet_r18-d8_512x1024_80k_cityscapes.py
- Name: pspnet_r18-d8_512x1024_80k_cityscapes
In Collection: pspnet
Metadata:
backbone: R-18-D8
crop size: (512,1024)
lr schd: 80000
inference time (ms/im):
- backend: PyTorch
batch size: 1
- value: 63.65
hardware: V100
backend: PyTorch
batch size: 1
mode: FP32
resolution: (512,1024)
value: 63.65
lr schd: 80000
memory (GB): 1.7
Name: pspnet_r18-d8_512x1024_80k_cityscapes
Results:
- Task: Semantic Segmentation
Dataset: Cityscapes
Metrics:
mIoU: 74.87
mIoU(ms+flip): 76.04
Task: Semantic Segmentation
Config: configs/pspnet/pspnet_r18-d8_512x1024_80k_cityscapes.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/pspnet/pspnet_r18-d8_512x1024_80k_cityscapes/pspnet_r18-d8_512x1024_80k_cityscapes_20201225_021458-09ffa746.pth
- Config: configs/pspnet/pspnet_r50-d8_512x1024_80k_cityscapes.py
- Name: pspnet_r50-d8_512x1024_80k_cityscapes
In Collection: pspnet
Metadata:
backbone: R-50-D8
crop size: (512,1024)
lr schd: 80000
Name: pspnet_r50-d8_512x1024_80k_cityscapes
Results:
- Task: Semantic Segmentation
Dataset: Cityscapes
Metrics:
mIoU: 78.55
mIoU(ms+flip): 79.79
Task: Semantic Segmentation
Config: configs/pspnet/pspnet_r50-d8_512x1024_80k_cityscapes.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/pspnet/pspnet_r50-d8_512x1024_80k_cityscapes/pspnet_r50-d8_512x1024_80k_cityscapes_20200606_112131-2376f12b.pth
- Config: configs/pspnet/pspnet_r101-d8_512x1024_80k_cityscapes.py
- Name: pspnet_r101-d8_512x1024_80k_cityscapes
In Collection: pspnet
Metadata:
backbone: R-101-D8
crop size: (512,1024)
lr schd: 80000
Name: pspnet_r101-d8_512x1024_80k_cityscapes
Results:
- Task: Semantic Segmentation
Dataset: Cityscapes
Metrics:
mIoU: 79.76
mIoU(ms+flip): 81.01
Task: Semantic Segmentation
Config: configs/pspnet/pspnet_r101-d8_512x1024_80k_cityscapes.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/pspnet/pspnet_r101-d8_512x1024_80k_cityscapes/pspnet_r101-d8_512x1024_80k_cityscapes_20200606_112211-e1e1100f.pth
- Config: configs/pspnet/pspnet_r18-d8_769x769_80k_cityscapes.py
- Name: pspnet_r18-d8_769x769_80k_cityscapes
In Collection: pspnet
Metadata:
backbone: R-18-D8
crop size: (769,769)
lr schd: 80000
inference time (ms/im):
- backend: PyTorch
batch size: 1
- value: 161.29
hardware: V100
backend: PyTorch
batch size: 1
mode: FP32
resolution: (769,769)
value: 161.29
lr schd: 80000
memory (GB): 1.9
Name: pspnet_r18-d8_769x769_80k_cityscapes
Results:
- Task: Semantic Segmentation
Dataset: Cityscapes
Metrics:
mIoU: 75.9
mIoU(ms+flip): 77.86
Task: Semantic Segmentation
Config: configs/pspnet/pspnet_r18-d8_769x769_80k_cityscapes.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/pspnet/pspnet_r18-d8_769x769_80k_cityscapes/pspnet_r18-d8_769x769_80k_cityscapes_20201225_021458-3deefc62.pth
- Config: configs/pspnet/pspnet_r50-d8_769x769_80k_cityscapes.py
- Name: pspnet_r50-d8_769x769_80k_cityscapes
In Collection: pspnet
Metadata:
backbone: R-50-D8
crop size: (769,769)
lr schd: 80000
Name: pspnet_r50-d8_769x769_80k_cityscapes
Results:
- Task: Semantic Segmentation
Dataset: Cityscapes
Metrics:
mIoU: 79.59
mIoU(ms+flip): 80.69
Task: Semantic Segmentation
Config: configs/pspnet/pspnet_r50-d8_769x769_80k_cityscapes.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/pspnet/pspnet_r50-d8_769x769_80k_cityscapes/pspnet_r50-d8_769x769_80k_cityscapes_20200606_210121-5ccf03dd.pth
- Config: configs/pspnet/pspnet_r101-d8_769x769_80k_cityscapes.py
- Name: pspnet_r101-d8_769x769_80k_cityscapes
In Collection: pspnet
Metadata:
backbone: R-101-D8
crop size: (769,769)
lr schd: 80000
Name: pspnet_r101-d8_769x769_80k_cityscapes
Results:
- Task: Semantic Segmentation
Dataset: Cityscapes
Metrics:
mIoU: 79.77
mIoU(ms+flip): 81.06
Task: Semantic Segmentation
Config: configs/pspnet/pspnet_r101-d8_769x769_80k_cityscapes.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/pspnet/pspnet_r101-d8_769x769_80k_cityscapes/pspnet_r101-d8_769x769_80k_cityscapes_20200606_225055-dba412fa.pth
- Config: configs/pspnet/pspnet_r18b-d8_512x1024_80k_cityscapes.py
- Name: pspnet_r18b-d8_512x1024_80k_cityscapes
In Collection: pspnet
Metadata:
backbone: R-18b-D8
crop size: (512,1024)
lr schd: 80000
inference time (ms/im):
- backend: PyTorch
batch size: 1
- value: 61.43
hardware: V100
backend: PyTorch
batch size: 1
mode: FP32
resolution: (512,1024)
value: 61.43
lr schd: 80000
memory (GB): 1.5
Name: pspnet_r18b-d8_512x1024_80k_cityscapes
Results:
- Task: Semantic Segmentation
Dataset: Cityscapes
Metrics:
mIoU: 74.23
mIoU(ms+flip): 75.79
Task: Semantic Segmentation
Config: configs/pspnet/pspnet_r18b-d8_512x1024_80k_cityscapes.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/pspnet/pspnet_r18b-d8_512x1024_80k_cityscapes/pspnet_r18b-d8_512x1024_80k_cityscapes_20201226_063116-26928a60.pth
- Config: configs/pspnet/pspnet_r50b-d8_512x1024_80k_cityscapes.py
- Name: pspnet_r50b-d8_512x1024_80k_cityscapes
In Collection: pspnet
Metadata:
backbone: R-50b-D8
crop size: (512,1024)
lr schd: 80000
inference time (ms/im):
- backend: PyTorch
batch size: 1
- value: 232.56
hardware: V100
backend: PyTorch
batch size: 1
mode: FP32
resolution: (512,1024)
value: 232.56
lr schd: 80000
memory (GB): 6.0
Name: pspnet_r50b-d8_512x1024_80k_cityscapes
Results:
- Task: Semantic Segmentation
Dataset: Cityscapes
Metrics:
mIoU: 78.22
mIoU(ms+flip): 79.46
Task: Semantic Segmentation
Config: configs/pspnet/pspnet_r50b-d8_512x1024_80k_cityscapes.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/pspnet/pspnet_r50b-d8_512x1024_80k_cityscapes/pspnet_r50b-d8_512x1024_80k_cityscapes_20201225_094315-6344287a.pth
- Config: configs/pspnet/pspnet_r101b-d8_512x1024_80k_cityscapes.py
- Name: pspnet_r101b-d8_512x1024_80k_cityscapes
In Collection: pspnet
Metadata:
backbone: R-101b-D8
crop size: (512,1024)
lr schd: 80000
inference time (ms/im):
- backend: PyTorch
batch size: 1
- value: 362.32
hardware: V100
backend: PyTorch
batch size: 1
mode: FP32
resolution: (512,1024)
value: 362.32
lr schd: 80000
memory (GB): 9.5
Name: pspnet_r101b-d8_512x1024_80k_cityscapes
Results:
- Task: Semantic Segmentation
Dataset: Cityscapes
Metrics:
mIoU: 79.69
mIoU(ms+flip): 80.79
Task: Semantic Segmentation
Config: configs/pspnet/pspnet_r101b-d8_512x1024_80k_cityscapes.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/pspnet/pspnet_r101b-d8_512x1024_80k_cityscapes/pspnet_r101b-d8_512x1024_80k_cityscapes_20201226_170012-3a4d38ab.pth
- Config: configs/pspnet/pspnet_r18b-d8_769x769_80k_cityscapes.py
- Name: pspnet_r18b-d8_769x769_80k_cityscapes
In Collection: pspnet
Metadata:
backbone: R-18b-D8
crop size: (769,769)
lr schd: 80000
inference time (ms/im):
- backend: PyTorch
batch size: 1
- value: 156.01
hardware: V100
backend: PyTorch
batch size: 1
mode: FP32
resolution: (769,769)
value: 156.01
lr schd: 80000
memory (GB): 1.7
Name: pspnet_r18b-d8_769x769_80k_cityscapes
Results:
- Task: Semantic Segmentation
Dataset: Cityscapes
Metrics:
mIoU: 74.92
mIoU(ms+flip): 76.9
Task: Semantic Segmentation
Config: configs/pspnet/pspnet_r18b-d8_769x769_80k_cityscapes.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/pspnet/pspnet_r18b-d8_769x769_80k_cityscapes/pspnet_r18b-d8_769x769_80k_cityscapes_20201226_080942-bf98d186.pth
- Config: configs/pspnet/pspnet_r50b-d8_769x769_80k_cityscapes.py
- Name: pspnet_r50b-d8_769x769_80k_cityscapes
In Collection: pspnet
Metadata:
backbone: R-50b-D8
crop size: (769,769)
lr schd: 80000
inference time (ms/im):
- backend: PyTorch
batch size: 1
- value: 531.91
hardware: V100
backend: PyTorch
batch size: 1
mode: FP32
resolution: (769,769)
value: 531.91
lr schd: 80000
memory (GB): 6.8
Name: pspnet_r50b-d8_769x769_80k_cityscapes
Results:
- Task: Semantic Segmentation
Dataset: Cityscapes
Metrics:
mIoU: 78.5
mIoU(ms+flip): 79.96
Task: Semantic Segmentation
Config: configs/pspnet/pspnet_r50b-d8_769x769_80k_cityscapes.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/pspnet/pspnet_r50b-d8_769x769_80k_cityscapes/pspnet_r50b-d8_769x769_80k_cityscapes_20201225_094316-4c643cf6.pth
- Config: configs/pspnet/pspnet_r101b-d8_769x769_80k_cityscapes.py
- Name: pspnet_r101b-d8_769x769_80k_cityscapes
In Collection: pspnet
Metadata:
backbone: R-101b-D8
crop size: (769,769)
lr schd: 80000
inference time (ms/im):
- backend: PyTorch
batch size: 1
- value: 854.7
hardware: V100
backend: PyTorch
batch size: 1
mode: FP32
resolution: (769,769)
value: 854.7
lr schd: 80000
memory (GB): 10.8
Name: pspnet_r101b-d8_769x769_80k_cityscapes
Results:
- Task: Semantic Segmentation
Dataset: Cityscapes
Metrics:
mIoU: 78.87
mIoU(ms+flip): 80.04
Task: Semantic Segmentation
Config: configs/pspnet/pspnet_r101b-d8_769x769_80k_cityscapes.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/pspnet/pspnet_r101b-d8_769x769_80k_cityscapes/pspnet_r101b-d8_769x769_80k_cityscapes_20201226_171823-f0e7c293.pth
- Config: configs/pspnet/pspnet_r50-d8_512x512_80k_ade20k.py
- Name: pspnet_r50-d8_512x512_80k_ade20k
In Collection: pspnet
Metadata:
backbone: R-50-D8
crop size: (512,512)
lr schd: 80000
inference time (ms/im):
- backend: PyTorch
batch size: 1
- value: 42.5
hardware: V100
backend: PyTorch
batch size: 1
mode: FP32
resolution: (512,512)
value: 42.5
lr schd: 80000
memory (GB): 8.5
Name: pspnet_r50-d8_512x512_80k_ade20k
Results:
- Task: Semantic Segmentation
Dataset: ADE20K
Metrics:
mIoU: 41.13
mIoU(ms+flip): 41.94
Task: Semantic Segmentation
Config: configs/pspnet/pspnet_r50-d8_512x512_80k_ade20k.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/pspnet/pspnet_r50-d8_512x512_80k_ade20k/pspnet_r50-d8_512x512_80k_ade20k_20200615_014128-15a8b914.pth
- Config: configs/pspnet/pspnet_r101-d8_512x512_80k_ade20k.py
- Name: pspnet_r101-d8_512x512_80k_ade20k
In Collection: pspnet
Metadata:
backbone: R-101-D8
crop size: (512,512)
lr schd: 80000
inference time (ms/im):
- backend: PyTorch
batch size: 1
- value: 65.36
hardware: V100
backend: PyTorch
batch size: 1
mode: FP32
resolution: (512,512)
value: 65.36
lr schd: 80000
memory (GB): 12.0
Name: pspnet_r101-d8_512x512_80k_ade20k
Results:
- Task: Semantic Segmentation
Dataset: ADE20K
Metrics:
mIoU: 43.57
mIoU(ms+flip): 44.35
Task: Semantic Segmentation
Config: configs/pspnet/pspnet_r101-d8_512x512_80k_ade20k.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/pspnet/pspnet_r101-d8_512x512_80k_ade20k/pspnet_r101-d8_512x512_80k_ade20k_20200614_031423-b6e782f0.pth
- Config: configs/pspnet/pspnet_r50-d8_512x512_160k_ade20k.py
- Name: pspnet_r50-d8_512x512_160k_ade20k
In Collection: pspnet
Metadata:
backbone: R-50-D8
crop size: (512,512)
lr schd: 160000
Name: pspnet_r50-d8_512x512_160k_ade20k
Results:
- Task: Semantic Segmentation
Dataset: ADE20K
Metrics:
mIoU: 42.48
mIoU(ms+flip): 43.44
Task: Semantic Segmentation
Config: configs/pspnet/pspnet_r50-d8_512x512_160k_ade20k.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/pspnet/pspnet_r50-d8_512x512_160k_ade20k/pspnet_r50-d8_512x512_160k_ade20k_20200615_184358-1890b0bd.pth
- Config: configs/pspnet/pspnet_r101-d8_512x512_160k_ade20k.py
- Name: pspnet_r101-d8_512x512_160k_ade20k
In Collection: pspnet
Metadata:
backbone: R-101-D8
crop size: (512,512)
lr schd: 160000
Name: pspnet_r101-d8_512x512_160k_ade20k
Results:
- Task: Semantic Segmentation
Dataset: ADE20K
Metrics:
mIoU: 44.39
mIoU(ms+flip): 45.35
Task: Semantic Segmentation
Config: configs/pspnet/pspnet_r101-d8_512x512_160k_ade20k.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/pspnet/pspnet_r101-d8_512x512_160k_ade20k/pspnet_r101-d8_512x512_160k_ade20k_20200615_100650-967c316f.pth
- Config: configs/pspnet/pspnet_r50-d8_512x512_20k_voc12aug.py
- Name: pspnet_r50-d8_512x512_20k_voc12aug
In Collection: pspnet
Metadata:
backbone: R-50-D8
crop size: (512,512)
lr schd: 20000
inference time (ms/im):
- backend: PyTorch
batch size: 1
- value: 42.39
hardware: V100
backend: PyTorch
batch size: 1
mode: FP32
resolution: (512,512)
value: 42.39
lr schd: 20000
memory (GB): 6.1
Name: pspnet_r50-d8_512x512_20k_voc12aug
Results:
- Task: Semantic Segmentation
Dataset: Pascal VOC 2012 + Aug
Metrics:
mIoU: 76.78
mIoU(ms+flip): 77.61
Task: Semantic Segmentation
Config: configs/pspnet/pspnet_r50-d8_512x512_20k_voc12aug.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/pspnet/pspnet_r50-d8_512x512_20k_voc12aug/pspnet_r50-d8_512x512_20k_voc12aug_20200617_101958-ed5dfbd9.pth
- Config: configs/pspnet/pspnet_r101-d8_512x512_20k_voc12aug.py
- Name: pspnet_r101-d8_512x512_20k_voc12aug
In Collection: pspnet
Metadata:
backbone: R-101-D8
crop size: (512,512)
lr schd: 20000
inference time (ms/im):
- backend: PyTorch
batch size: 1
- value: 66.58
hardware: V100
backend: PyTorch
batch size: 1
mode: FP32
resolution: (512,512)
value: 66.58
lr schd: 20000
memory (GB): 9.6
Name: pspnet_r101-d8_512x512_20k_voc12aug
Results:
- Task: Semantic Segmentation
Dataset: Pascal VOC 2012 + Aug
Metrics:
mIoU: 78.47
mIoU(ms+flip): 79.25
Task: Semantic Segmentation
Config: configs/pspnet/pspnet_r101-d8_512x512_20k_voc12aug.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/pspnet/pspnet_r101-d8_512x512_20k_voc12aug/pspnet_r101-d8_512x512_20k_voc12aug_20200617_102003-4aef3c9a.pth
- Config: configs/pspnet/pspnet_r50-d8_512x512_40k_voc12aug.py
- Name: pspnet_r50-d8_512x512_40k_voc12aug
In Collection: pspnet
Metadata:
backbone: R-50-D8
crop size: (512,512)
lr schd: 40000
Name: pspnet_r50-d8_512x512_40k_voc12aug
Results:
- Task: Semantic Segmentation
Dataset: Pascal VOC 2012 + Aug
Metrics:
mIoU: 77.29
mIoU(ms+flip): 78.48
Task: Semantic Segmentation
Config: configs/pspnet/pspnet_r50-d8_512x512_40k_voc12aug.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/pspnet/pspnet_r50-d8_512x512_40k_voc12aug/pspnet_r50-d8_512x512_40k_voc12aug_20200613_161222-ae9c1b8c.pth
- Config: configs/pspnet/pspnet_r101-d8_512x512_40k_voc12aug.py
- Name: pspnet_r101-d8_512x512_40k_voc12aug
In Collection: pspnet
Metadata:
backbone: R-101-D8
crop size: (512,512)
lr schd: 40000
Name: pspnet_r101-d8_512x512_40k_voc12aug
Results:
- Task: Semantic Segmentation
Dataset: Pascal VOC 2012 + Aug
Metrics:
mIoU: 78.52
mIoU(ms+flip): 79.57
Task: Semantic Segmentation
Config: configs/pspnet/pspnet_r101-d8_512x512_40k_voc12aug.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/pspnet/pspnet_r101-d8_512x512_40k_voc12aug/pspnet_r101-d8_512x512_40k_voc12aug_20200613_161222-bc933b18.pth
- Config: configs/pspnet/pspnet_r101-d8_480x480_40k_pascal_context.py
- Name: pspnet_r101-d8_480x480_40k_pascal_context
In Collection: pspnet
Metadata:
backbone: R-101-D8
crop size: (480,480)
lr schd: 40000
inference time (ms/im):
- backend: PyTorch
batch size: 1
- value: 103.31
hardware: V100
backend: PyTorch
batch size: 1
mode: FP32
resolution: (480,480)
value: 103.31
lr schd: 40000
memory (GB): 8.8
Name: pspnet_r101-d8_480x480_40k_pascal_context
Results:
- Task: Semantic Segmentation
Dataset: Pascal Context
Metrics:
mIoU: 46.6
mIoU(ms+flip): 47.78
Task: Semantic Segmentation
Config: configs/pspnet/pspnet_r101-d8_480x480_40k_pascal_context.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/pspnet/pspnet_r101-d8_480x480_40k_pascal_context/pspnet_r101-d8_480x480_40k_pascal_context_20200911_211210-bf0f5d7c.pth
- Config: configs/pspnet/pspnet_r101-d8_480x480_80k_pascal_context.py
- Name: pspnet_r101-d8_480x480_80k_pascal_context
In Collection: pspnet
Metadata:
backbone: R-101-D8
crop size: (480,480)
lr schd: 80000
Name: pspnet_r101-d8_480x480_80k_pascal_context
Results:
- Task: Semantic Segmentation
Dataset: Pascal Context
Metrics:
mIoU: 46.03
mIoU(ms+flip): 47.15
Task: Semantic Segmentation
Config: configs/pspnet/pspnet_r101-d8_480x480_80k_pascal_context.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/pspnet/pspnet_r101-d8_480x480_80k_pascal_context/pspnet_r101-d8_480x480_80k_pascal_context_20200911_190530-c86d6233.pth
- Config: configs/pspnet/pspnet_r101-d8_480x480_40k_pascal_context_59.py
- Name: pspnet_r101-d8_480x480_40k_pascal_context_59
In Collection: pspnet
Metadata:
backbone: R-101-D8
crop size: (480,480)
lr schd: 40000
Name: pspnet_r101-d8_480x480_40k_pascal_context_59
Results:
- Task: Semantic Segmentation
Dataset: Pascal Context 59
Metrics:
mIoU: 52.02
mIoU(ms+flip): 53.54
Task: Semantic Segmentation
Config: configs/pspnet/pspnet_r101-d8_480x480_40k_pascal_context_59.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/pspnet/pspnet_r101-d8_480x480_40k_pascal_context_59/pspnet_r101-d8_480x480_40k_pascal_context_59_20210416_114524-86d44cd4.pth
- Config: configs/pspnet/pspnet_r101-d8_480x480_80k_pascal_context_59.py
- Name: pspnet_r101-d8_480x480_80k_pascal_context_59
In Collection: pspnet
Metadata:
backbone: R-101-D8
crop size: (480,480)
lr schd: 80000
Name: pspnet_r101-d8_480x480_80k_pascal_context_59
Results:
- Task: Semantic Segmentation
Dataset: Pascal Context 59
Metrics:
mIoU: 52.47
mIoU(ms+flip): 53.99
Task: Semantic Segmentation
Config: configs/pspnet/pspnet_r101-d8_480x480_80k_pascal_context_59.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/pspnet/pspnet_r101-d8_480x480_80k_pascal_context_59/pspnet_r101-d8_480x480_80k_pascal_context_59_20210416_114418-fa6caaa2.pth
- Config: configs/pspnet/pspnet_r50-d8_512x512_4x4_20k_coco-stuff10k.py
- Name: pspnet_r50-d8_512x512_4x4_20k_coco-stuff10k
In Collection: pspnet
Metadata:
backbone: R-50-D8
crop size: (512,512)
lr schd: 20000
inference time (ms/im):
- backend: PyTorch
batch size: 1
- value: 48.78
hardware: V100
backend: PyTorch
batch size: 1
mode: FP32
resolution: (512,512)
value: 48.78
lr schd: 20000
memory (GB): 9.6
Name: pspnet_r50-d8_512x512_4x4_20k_coco-stuff10k
Results:
- Task: Semantic Segmentation
Dataset: COCO-Stuff 10k
Metrics:
mIoU: 35.69
mIoU(ms+flip): 36.62
Task: Semantic Segmentation
Config: configs/pspnet/pspnet_r50-d8_512x512_4x4_20k_coco-stuff10k.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/pspnet/pspnet_r50-d8_512x512_4x4_20k_coco-stuff10k/pspnet_r50-d8_512x512_4x4_20k_coco-stuff10k_20210820_203258-b88df27f.pth
- Config: configs/pspnet/pspnet_r101-d8_512x512_4x4_20k_coco-stuff10k.py
- Name: pspnet_r101-d8_512x512_4x4_20k_coco-stuff10k
In Collection: pspnet
Metadata:
backbone: R-101-D8
crop size: (512,512)
lr schd: 20000
inference time (ms/im):
- backend: PyTorch
batch size: 1
- value: 90.09
hardware: V100
backend: PyTorch
batch size: 1
mode: FP32
resolution: (512,512)
value: 90.09
lr schd: 20000
memory (GB): 13.2
Name: pspnet_r101-d8_512x512_4x4_20k_coco-stuff10k
Results:
- Task: Semantic Segmentation
Dataset: COCO-Stuff 10k
Metrics:
mIoU: 37.26
mIoU(ms+flip): 38.52
Task: Semantic Segmentation
Config: configs/pspnet/pspnet_r101-d8_512x512_4x4_20k_coco-stuff10k.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/pspnet/pspnet_r101-d8_512x512_4x4_20k_coco-stuff10k/pspnet_r101-d8_512x512_4x4_20k_coco-stuff10k_20210820_232135-76aae482.pth
- Config: configs/pspnet/pspnet_r50-d8_512x512_4x4_40k_coco-stuff10k.py
- Name: pspnet_r50-d8_512x512_4x4_40k_coco-stuff10k
In Collection: pspnet
Metadata:
backbone: R-50-D8
crop size: (512,512)
lr schd: 40000
Name: pspnet_r50-d8_512x512_4x4_40k_coco-stuff10k
Results:
- Task: Semantic Segmentation
Dataset: COCO-Stuff 10k
Metrics:
mIoU: 36.33
mIoU(ms+flip): 37.24
Task: Semantic Segmentation
Config: configs/pspnet/pspnet_r50-d8_512x512_4x4_40k_coco-stuff10k.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/pspnet/pspnet_r50-d8_512x512_4x4_40k_coco-stuff10k/pspnet_r50-d8_512x512_4x4_40k_coco-stuff10k_20210821_030857-92e2902b.pth
- Config: configs/pspnet/pspnet_r101-d8_512x512_4x4_40k_coco-stuff10k.py
- Name: pspnet_r101-d8_512x512_4x4_40k_coco-stuff10k
In Collection: pspnet
Metadata:
backbone: R-101-D8
crop size: (512,512)
lr schd: 40000
Name: pspnet_r101-d8_512x512_4x4_40k_coco-stuff10k
Results:
- Task: Semantic Segmentation
Dataset: COCO-Stuff 10k
Metrics:
mIoU: 37.76
mIoU(ms+flip): 38.86
Task: Semantic Segmentation
Config: configs/pspnet/pspnet_r101-d8_512x512_4x4_40k_coco-stuff10k.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/pspnet/pspnet_r101-d8_512x512_4x4_40k_coco-stuff10k/pspnet_r101-d8_512x512_4x4_40k_coco-stuff10k_20210821_014022-831aec95.pth
- Config: configs/pspnet/pspnet_r50-d8_512x512_4x4_80k_coco-stuff164k.py
- Name: pspnet_r50-d8_512x512_4x4_80k_coco-stuff164k
In Collection: pspnet
Metadata:
backbone: R-50-D8
crop size: (512,512)
lr schd: 80000
inference time (ms/im):
- backend: PyTorch
batch size: 1
- value: 48.78
hardware: V100
backend: PyTorch
batch size: 1
mode: FP32
resolution: (512,512)
value: 48.78
lr schd: 80000
memory (GB): 9.6
Name: pspnet_r50-d8_512x512_4x4_80k_coco-stuff164k
Results:
- Task: Semantic Segmentation
Dataset: COCO-Stuff 164k
Metrics:
mIoU: 38.8
mIoU(ms+flip): 39.19
Task: Semantic Segmentation
Config: configs/pspnet/pspnet_r50-d8_512x512_4x4_80k_coco-stuff164k.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/pspnet/pspnet_r50-d8_512x512_4x4_80k_coco-stuff164k/pspnet_r50-d8_512x512_4x4_80k_coco-stuff164k_20210707_152034-0e41b2db.pth
- Config: configs/pspnet/pspnet_r101-d8_512x512_4x4_80k_coco-stuff164k.py
- Name: pspnet_r101-d8_512x512_4x4_80k_coco-stuff164k
In Collection: pspnet
Metadata:
backbone: R-101-D8
crop size: (512,512)
lr schd: 80000
inference time (ms/im):
- backend: PyTorch
batch size: 1
- value: 90.09
hardware: V100
backend: PyTorch
batch size: 1
mode: FP32
resolution: (512,512)
value: 90.09
lr schd: 80000
memory (GB): 13.2
Name: pspnet_r101-d8_512x512_4x4_80k_coco-stuff164k
Results:
- Task: Semantic Segmentation
Dataset: COCO-Stuff 164k
Metrics:
mIoU: 40.34
mIoU(ms+flip): 40.79
Task: Semantic Segmentation
Config: configs/pspnet/pspnet_r101-d8_512x512_4x4_80k_coco-stuff164k.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/pspnet/pspnet_r101-d8_512x512_4x4_80k_coco-stuff164k/pspnet_r101-d8_512x512_4x4_80k_coco-stuff164k_20210707_152034-7eb41789.pth
- Config: configs/pspnet/pspnet_r50-d8_512x512_4x4_160k_coco-stuff164k.py
- Name: pspnet_r50-d8_512x512_4x4_160k_coco-stuff164k
In Collection: pspnet
Metadata:
backbone: R-50-D8
crop size: (512,512)
lr schd: 160000
Name: pspnet_r50-d8_512x512_4x4_160k_coco-stuff164k
Results:
- Task: Semantic Segmentation
Dataset: COCO-Stuff 164k
Metrics:
mIoU: 39.64
mIoU(ms+flip): 39.97
Task: Semantic Segmentation
Config: configs/pspnet/pspnet_r50-d8_512x512_4x4_160k_coco-stuff164k.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/pspnet/pspnet_r50-d8_512x512_4x4_160k_coco-stuff164k/pspnet_r50-d8_512x512_4x4_160k_coco-stuff164k_20210707_152004-51276a57.pth
- Config: configs/pspnet/pspnet_r101-d8_512x512_4x4_160k_coco-stuff164k.py
- Name: pspnet_r101-d8_512x512_4x4_160k_coco-stuff164k
In Collection: pspnet
Metadata:
backbone: R-101-D8
crop size: (512,512)
lr schd: 160000
Name: pspnet_r101-d8_512x512_4x4_160k_coco-stuff164k
Results:
- Task: Semantic Segmentation
Dataset: COCO-Stuff 164k
Metrics:
mIoU: 41.28
mIoU(ms+flip): 41.66
Task: Semantic Segmentation
Config: configs/pspnet/pspnet_r101-d8_512x512_4x4_160k_coco-stuff164k.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/pspnet/pspnet_r101-d8_512x512_4x4_160k_coco-stuff164k/pspnet_r101-d8_512x512_4x4_160k_coco-stuff164k_20210707_152004-4af9621b.pth
- Config: configs/pspnet/pspnet_r50-d8_512x512_4x4_320k_coco-stuff164k.py
- Name: pspnet_r50-d8_512x512_4x4_320k_coco-stuff164k
In Collection: pspnet
Metadata:
backbone: R-50-D8
crop size: (512,512)
lr schd: 320000
Name: pspnet_r50-d8_512x512_4x4_320k_coco-stuff164k
Results:
- Task: Semantic Segmentation
Dataset: COCO-Stuff 164k
Metrics:
mIoU: 40.53
mIoU(ms+flip): 40.75
Task: Semantic Segmentation
Config: configs/pspnet/pspnet_r50-d8_512x512_4x4_320k_coco-stuff164k.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/pspnet/pspnet_r50-d8_512x512_4x4_320k_coco-stuff164k/pspnet_r50-d8_512x512_4x4_320k_coco-stuff164k_20210707_152004-be9610cc.pth
- Config: configs/pspnet/pspnet_r101-d8_512x512_4x4_320k_coco-stuff164k.py
- Name: pspnet_r101-d8_512x512_4x4_320k_coco-stuff164k
In Collection: pspnet
Metadata:
backbone: R-101-D8
crop size: (512,512)
lr schd: 320000
Name: pspnet_r101-d8_512x512_4x4_320k_coco-stuff164k
Results:
- Task: Semantic Segmentation
Dataset: COCO-Stuff 164k
Metrics:
mIoU: 41.95
mIoU(ms+flip): 42.42
Task: Semantic Segmentation
Config: configs/pspnet/pspnet_r101-d8_512x512_4x4_320k_coco-stuff164k.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/pspnet/pspnet_r101-d8_512x512_4x4_320k_coco-stuff164k/pspnet_r101-d8_512x512_4x4_320k_coco-stuff164k_20210707_152004-72220c60.pth

9
configs/resnest/README.md
View File

@ -4,6 +4,13 @@
<!-- [ALGORITHM] -->
<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>
<details>
<summary align="right"><a href="https://arxiv.org/abs/2004.08955">ResNeSt (ArXiv'2020)</a></summary>
```latex
@article{zhang2020resnest,
title={ResNeSt: Split-Attention Networks},
@ -13,6 +20,8 @@ year={2020}
}
```
</details>
## Results and models
### Cityscapes

125
configs/resnest/resnest.yml
View File

@ -1,183 +1,192 @@
Collections:
- Metadata:
- Name: resnest
Metadata:
Training Data:
- Cityscapes
- ADE20k
Name: resnest
Paper:
URL: https://arxiv.org/abs/2004.08955
Title: 'ResNeSt: Split-Attention Networks'
README: configs/resnest/README.md
Code:
URL: https://github.com/open-mmlab/mmsegmentation/blob/v0.17.0/mmseg/models/backbones/resnest.py#L271
Version: v0.17.0
Converted From:
Code: https://github.com/zhanghang1989/ResNeSt
Models:
- Config: configs/resnest/fcn_s101-d8_512x1024_80k_cityscapes.py
- Name: fcn_s101-d8_512x1024_80k_cityscapes
In Collection: resnest
Metadata:
backbone: S-101-D8
crop size: (512,1024)
lr schd: 80000
inference time (ms/im):
- backend: PyTorch
batch size: 1
- value: 418.41
hardware: V100
backend: PyTorch
batch size: 1
mode: FP32
resolution: (512,1024)
value: 418.41
lr schd: 80000
memory (GB): 11.4
Name: fcn_s101-d8_512x1024_80k_cityscapes
Results:
- Task: Semantic Segmentation
Dataset: Cityscapes
Metrics:
mIoU: 77.56
mIoU(ms+flip): 78.98
Task: Semantic Segmentation
Config: configs/resnest/fcn_s101-d8_512x1024_80k_cityscapes.py
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
- Config: configs/resnest/pspnet_s101-d8_512x1024_80k_cityscapes.py
- Name: pspnet_s101-d8_512x1024_80k_cityscapes
In Collection: resnest
Metadata:
backbone: S-101-D8
crop size: (512,1024)
lr schd: 80000
inference time (ms/im):
- backend: PyTorch
batch size: 1
- value: 396.83
hardware: V100
backend: PyTorch
batch size: 1
mode: FP32
resolution: (512,1024)
value: 396.83
lr schd: 80000
memory (GB): 11.8
Name: pspnet_s101-d8_512x1024_80k_cityscapes
Results:
- Task: Semantic Segmentation
Dataset: Cityscapes
Metrics:
mIoU: 78.57
mIoU(ms+flip): 79.19
Task: Semantic Segmentation
Config: configs/resnest/pspnet_s101-d8_512x1024_80k_cityscapes.py
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
- Config: configs/resnest/deeplabv3_s101-d8_512x1024_80k_cityscapes.py
- Name: deeplabv3_s101-d8_512x1024_80k_cityscapes
In Collection: resnest
Metadata:
backbone: S-101-D8
crop size: (512,1024)
lr schd: 80000
inference time (ms/im):
- backend: PyTorch
batch size: 1
- value: 531.91
hardware: V100
backend: PyTorch
batch size: 1
mode: FP32
resolution: (512,1024)
value: 531.91
lr schd: 80000
memory (GB): 11.9
Name: deeplabv3_s101-d8_512x1024_80k_cityscapes
Results:
- Task: Semantic Segmentation
Dataset: Cityscapes
Metrics:
mIoU: 79.67
mIoU(ms+flip): 80.51
Task: Semantic Segmentation
Config: configs/resnest/deeplabv3_s101-d8_512x1024_80k_cityscapes.py
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
- Config: configs/resnest/deeplabv3plus_s101-d8_512x1024_80k_cityscapes.py
- Name: deeplabv3plus_s101-d8_512x1024_80k_cityscapes
In Collection: resnest
Metadata:
backbone: S-101-D8
crop size: (512,1024)
lr schd: 80000
inference time (ms/im):
- backend: PyTorch
batch size: 1
- value: 423.73
hardware: V100
backend: PyTorch
batch size: 1
mode: FP32
resolution: (512,1024)
value: 423.73
lr schd: 80000
memory (GB): 13.2
Name: deeplabv3plus_s101-d8_512x1024_80k_cityscapes
Results:
- Task: Semantic Segmentation
Dataset: Cityscapes
Metrics:
mIoU: 79.62
mIoU(ms+flip): 80.27
Task: Semantic Segmentation
Config: configs/resnest/deeplabv3plus_s101-d8_512x1024_80k_cityscapes.py
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
- Config: configs/resnest/fcn_s101-d8_512x512_160k_ade20k.py
- Name: fcn_s101-d8_512x512_160k_ade20k
In Collection: resnest
Metadata:
backbone: S-101-D8
crop size: (512,512)
lr schd: 160000
inference time (ms/im):
- backend: PyTorch
batch size: 1
- value: 77.76
hardware: V100
backend: PyTorch
batch size: 1
mode: FP32
resolution: (512,512)
value: 77.76
lr schd: 160000
memory (GB): 14.2
Name: fcn_s101-d8_512x512_160k_ade20k
Results:
- Task: Semantic Segmentation
Dataset: ADE20k
Metrics:
mIoU: 45.62
mIoU(ms+flip): 46.16
Task: Semantic Segmentation
Config: configs/resnest/fcn_s101-d8_512x512_160k_ade20k.py
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
- Config: configs/resnest/pspnet_s101-d8_512x512_160k_ade20k.py
- Name: pspnet_s101-d8_512x512_160k_ade20k
In Collection: resnest
Metadata:
backbone: S-101-D8
crop size: (512,512)
lr schd: 160000
inference time (ms/im):
- backend: PyTorch
batch size: 1
- value: 76.8
hardware: V100
backend: PyTorch
batch size: 1
mode: FP32
resolution: (512,512)
value: 76.8
lr schd: 160000
memory (GB): 14.2
Name: pspnet_s101-d8_512x512_160k_ade20k
Results:
- Task: Semantic Segmentation
Dataset: ADE20k
Metrics:
mIoU: 45.44
mIoU(ms+flip): 46.28
Task: Semantic Segmentation
Config: configs/resnest/pspnet_s101-d8_512x512_160k_ade20k.py
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
- Config: configs/resnest/deeplabv3_s101-d8_512x512_160k_ade20k.py
- Name: deeplabv3_s101-d8_512x512_160k_ade20k
In Collection: resnest
Metadata:
backbone: S-101-D8
crop size: (512,512)
lr schd: 160000
inference time (ms/im):
- backend: PyTorch
batch size: 1
- value: 107.76
hardware: V100
backend: PyTorch
batch size: 1
mode: FP32
resolution: (512,512)
value: 107.76
lr schd: 160000
memory (GB): 14.6
Name: deeplabv3_s101-d8_512x512_160k_ade20k
Results:
- Task: Semantic Segmentation
Dataset: ADE20k
Metrics:
mIoU: 45.71
mIoU(ms+flip): 46.59
Task: Semantic Segmentation
Config: configs/resnest/deeplabv3_s101-d8_512x512_160k_ade20k.py
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
- Config: configs/resnest/deeplabv3plus_s101-d8_512x512_160k_ade20k.py
- Name: deeplabv3plus_s101-d8_512x512_160k_ade20k
In Collection: resnest
Metadata:
backbone: S-101-D8
crop size: (512,512)
lr schd: 160000
inference time (ms/im):
- backend: PyTorch
batch size: 1
- value: 83.61
hardware: V100
backend: PyTorch
batch size: 1
mode: FP32
resolution: (512,512)
value: 83.61
lr schd: 160000
memory (GB): 16.2
Name: deeplabv3plus_s101-d8_512x512_160k_ade20k
Results:
- Task: Semantic Segmentation
Dataset: ADE20k
Metrics:
mIoU: 46.47
mIoU(ms+flip): 47.27
Task: Semantic Segmentation
Config: configs/resnest/deeplabv3plus_s101-d8_512x512_160k_ade20k.py
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

9
configs/segformer/README.md
View File

@ -4,6 +4,13 @@
<!-- [ALGORITHM] -->
<a href="https://github.com/NVlabs/SegFormer">Official Repo</a>
<a href="https://github.com/open-mmlab/mmsegmentation/blob/v0.17.0/mmseg/models/backbones/mit.py#L246">Code Snippet</a>
<details>
<summary align="right"><a href="https://arxiv.org/abs/2105.15203">SegFormer (ArXiv'2021)</a></summary>
```latex
@article{xie2021segformer,
title={SegFormer: Simple and Efficient Design for Semantic Segmentation with Transformers},
@ -13,6 +20,8 @@
}
```
</details>
## Usage
To use other repositories' pre-trained models, it is necessary to convert keys.

111
configs/segformer/segformer.yml
View File

@ -1,160 +1,169 @@
Collections:
- Metadata:
- Name: segformer
Metadata:
Training Data:
- ADE20k
Name: segformer
Paper:
URL: https://arxiv.org/abs/2105.15203
Title: resize image to multiple of 32, improve SegFormer by 0.5-1.0 mIoU.
README: configs/segformer/README.md
Code:
URL: https://github.com/open-mmlab/mmsegmentation/blob/v0.17.0/mmseg/models/backbones/mit.py#L246
Version: v0.17.0
Converted From:
Code: https://github.com/NVlabs/SegFormer
Models:
- Config: configs/segformer/segformer_mit-b0_512x512_160k_ade20k.py
- Name: segformer_mit-b0_512x512_160k_ade20k
In Collection: segformer
Metadata:
backbone: MIT-B0
crop size: (512,512)
lr schd: 160000
inference time (ms/im):
- backend: PyTorch
batch size: 1
- value: 19.49
hardware: V100
backend: PyTorch
batch size: 1
mode: FP32
resolution: (512,512)
value: 19.49
lr schd: 160000
memory (GB): 2.1
Name: segformer_mit-b0_512x512_160k_ade20k
Results:
- Task: Semantic Segmentation
Dataset: ADE20k
Metrics:
mIoU: 37.41
mIoU(ms+flip): 38.34
Task: Semantic Segmentation
Config: configs/segformer/segformer_mit-b0_512x512_160k_ade20k.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/segformer/segformer_mit-b0_512x512_160k_ade20k/segformer_mit-b0_512x512_160k_ade20k_20210726_101530-8ffa8fda.pth
- Config: configs/segformer/segformer_mit-b1_512x512_160k_ade20k.py
- Name: segformer_mit-b1_512x512_160k_ade20k
In Collection: segformer
Metadata:
backbone: MIT-B1
crop size: (512,512)
lr schd: 160000
inference time (ms/im):
- backend: PyTorch
batch size: 1
- value: 20.98
hardware: V100
backend: PyTorch
batch size: 1
mode: FP32
resolution: (512,512)
value: 20.98
lr schd: 160000
memory (GB): 2.6
Name: segformer_mit-b1_512x512_160k_ade20k
Results:
- Task: Semantic Segmentation
Dataset: ADE20k
Metrics:
mIoU: 40.97
mIoU(ms+flip): 42.54
Task: Semantic Segmentation
Config: configs/segformer/segformer_mit-b1_512x512_160k_ade20k.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/segformer/segformer_mit-b1_512x512_160k_ade20k/segformer_mit-b1_512x512_160k_ade20k_20210726_112106-d70e859d.pth
- Config: configs/segformer/segformer_mit-b2_512x512_160k_ade20k.py
- Name: segformer_mit-b2_512x512_160k_ade20k
In Collection: segformer
Metadata:
backbone: MIT-B2
crop size: (512,512)
lr schd: 160000
inference time (ms/im):
- backend: PyTorch
batch size: 1
- value: 32.38
hardware: V100
backend: PyTorch
batch size: 1
mode: FP32
resolution: (512,512)
value: 32.38
lr schd: 160000
memory (GB): 3.6
Name: segformer_mit-b2_512x512_160k_ade20k
Results:
- Task: Semantic Segmentation
Dataset: ADE20k
Metrics:
mIoU: 45.58
mIoU(ms+flip): 47.03
Task: Semantic Segmentation
Config: configs/segformer/segformer_mit-b2_512x512_160k_ade20k.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/segformer/segformer_mit-b2_512x512_160k_ade20k/segformer_mit-b2_512x512_160k_ade20k_20210726_112103-cbd414ac.pth
- Config: configs/segformer/segformer_mit-b3_512x512_160k_ade20k.py
- Name: segformer_mit-b3_512x512_160k_ade20k
In Collection: segformer
Metadata:
backbone: MIT-B3
crop size: (512,512)
lr schd: 160000
inference time (ms/im):
- backend: PyTorch
batch size: 1
- value: 45.23
hardware: V100
backend: PyTorch
batch size: 1
mode: FP32
resolution: (512,512)
value: 45.23
lr schd: 160000
memory (GB): 4.8
Name: segformer_mit-b3_512x512_160k_ade20k
Results:
- Task: Semantic Segmentation
Dataset: ADE20k
Metrics:
mIoU: 47.82
mIoU(ms+flip): 48.81
Task: Semantic Segmentation
Config: configs/segformer/segformer_mit-b3_512x512_160k_ade20k.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/segformer/segformer_mit-b3_512x512_160k_ade20k/segformer_mit-b3_512x512_160k_ade20k_20210726_081410-962b98d2.pth
- Config: configs/segformer/segformer_mit-b4_512x512_160k_ade20k.py
- Name: segformer_mit-b4_512x512_160k_ade20k
In Collection: segformer
Metadata:
backbone: MIT-B4
crop size: (512,512)
lr schd: 160000
inference time (ms/im):
- backend: PyTorch
batch size: 1
- value: 64.72
hardware: V100
backend: PyTorch
batch size: 1
mode: FP32
resolution: (512,512)
value: 64.72
lr schd: 160000
memory (GB): 6.1
Name: segformer_mit-b4_512x512_160k_ade20k
Results:
- Task: Semantic Segmentation
Dataset: ADE20k
Metrics:
mIoU: 48.46
mIoU(ms+flip): 49.76
Task: Semantic Segmentation
Config: configs/segformer/segformer_mit-b4_512x512_160k_ade20k.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/segformer/segformer_mit-b4_512x512_160k_ade20k/segformer_mit-b4_512x512_160k_ade20k_20210728_183055-7f509d7d.pth
- Config: configs/segformer/segformer_mit-b5_512x512_160k_ade20k.py
- Name: segformer_mit-b5_512x512_160k_ade20k
In Collection: segformer
Metadata:
backbone: MIT-B5
crop size: (512,512)
lr schd: 160000
inference time (ms/im):
- backend: PyTorch
batch size: 1
- value: 84.1
hardware: V100
backend: PyTorch
batch size: 1
mode: FP32
resolution: (512,512)
value: 84.1
lr schd: 160000
memory (GB): 7.2
Name: segformer_mit-b5_512x512_160k_ade20k
Results:
- Task: Semantic Segmentation
Dataset: ADE20k
Metrics:
mIoU: 49.13
mIoU(ms+flip): 50.22
Task: Semantic Segmentation
Config: configs/segformer/segformer_mit-b5_512x512_160k_ade20k.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/segformer/segformer_mit-b5_512x512_160k_ade20k/segformer_mit-b5_512x512_160k_ade20k_20210726_145235-94cedf59.pth
- Config: configs/segformer/segformer_mit-b5_640x640_160k_ade20k.py
- Name: segformer_mit-b5_640x640_160k_ade20k
In Collection: segformer
Metadata:
backbone: MIT-B5
crop size: (640,640)
lr schd: 160000
inference time (ms/im):
- backend: PyTorch
batch size: 1
- value: 88.5
hardware: V100
backend: PyTorch
batch size: 1
mode: FP32
resolution: (640,640)
value: 88.5
lr schd: 160000
memory (GB): 11.5
Name: segformer_mit-b5_640x640_160k_ade20k
Results:
- Task: Semantic Segmentation
Dataset: ADE20k
Metrics:
mIoU: 49.62
mIoU(ms+flip): 50.36
Task: Semantic Segmentation
Config: configs/segformer/segformer_mit-b5_640x640_160k_ade20k.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/segformer/segformer_mit-b5_640x640_160k_ade20k/segformer_mit-b5_640x640_160k_ade20k_20210801_121243-41d2845b.pth

9
configs/sem_fpn/README.md
View File

@ -4,6 +4,13 @@
<!-- [ALGORITHM] -->
<a href="https://github.com/facebookresearch/detectron2">Official Repo</a>
<a href="https://github.com/open-mmlab/mmsegmentation/blob/v0.17.0/mmseg/models/decode_heads/fpn_head.py#L12">Code Snippet</a>
<details>
<summary align="right"><a href="https://arxiv.org/abs/1901.02446">Semantic FPN (CVPR'2019)</a></summary>
```latex
@article{Kirillov_2019,
title={Panoptic Feature Pyramid Networks},
@ -18,6 +25,8 @@
}
```
</details>
## Results and models
### Cityscapes

69
configs/sem_fpn/sem_fpn.yml
View File

@ -1,95 +1,104 @@
Collections:
- Metadata:
- Name: sem_fpn
Metadata:
Training Data:
- Cityscapes
- ADE20K
Name: sem_fpn
Paper:
URL: https://arxiv.org/abs/1901.02446
Title: Panoptic Feature Pyramid Networks
README: configs/sem_fpn/README.md
Code:
URL: https://github.com/open-mmlab/mmsegmentation/blob/v0.17.0/mmseg/models/decode_heads/fpn_head.py#L12
Version: v0.17.0
Converted From:
Code: https://github.com/facebookresearch/detectron2
Models:
- Config: configs/sem_fpn/fpn_r50_512x1024_80k_cityscapes.py
- Name: fpn_r50_512x1024_80k_cityscapes
In Collection: sem_fpn
Metadata:
backbone: R-50
crop size: (512,1024)
lr schd: 80000
inference time (ms/im):
- backend: PyTorch
batch size: 1
- value: 73.86
hardware: V100
backend: PyTorch
batch size: 1
mode: FP32
resolution: (512,1024)
value: 73.86
lr schd: 80000
memory (GB): 2.8
Name: fpn_r50_512x1024_80k_cityscapes
Results:
- Task: Semantic Segmentation
Dataset: Cityscapes
Metrics:
mIoU: 74.52
mIoU(ms+flip): 76.08
Task: Semantic Segmentation
Config: configs/sem_fpn/fpn_r50_512x1024_80k_cityscapes.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/sem_fpn/fpn_r50_512x1024_80k_cityscapes/fpn_r50_512x1024_80k_cityscapes_20200717_021437-94018a0d.pth
- Config: configs/sem_fpn/fpn_r101_512x1024_80k_cityscapes.py
- Name: fpn_r101_512x1024_80k_cityscapes
In Collection: sem_fpn
Metadata:
backbone: R-101
crop size: (512,1024)
lr schd: 80000
inference time (ms/im):
- backend: PyTorch
batch size: 1
- value: 97.18
hardware: V100
backend: PyTorch
batch size: 1
mode: FP32
resolution: (512,1024)
value: 97.18
lr schd: 80000
memory (GB): 3.9
Name: fpn_r101_512x1024_80k_cityscapes
Results:
- Task: Semantic Segmentation
Dataset: Cityscapes
Metrics:
mIoU: 75.8
mIoU(ms+flip): 77.4
Task: Semantic Segmentation
Config: configs/sem_fpn/fpn_r101_512x1024_80k_cityscapes.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/sem_fpn/fpn_r101_512x1024_80k_cityscapes/fpn_r101_512x1024_80k_cityscapes_20200717_012416-c5800d4c.pth
- Config: configs/sem_fpn/fpn_r50_512x512_160k_ade20k.py
- Name: fpn_r50_512x512_160k_ade20k
In Collection: sem_fpn
Metadata:
backbone: R-50
crop size: (512,512)
lr schd: 160000
inference time (ms/im):
- backend: PyTorch
batch size: 1
- value: 17.93
hardware: V100
backend: PyTorch
batch size: 1
mode: FP32
resolution: (512,512)
value: 17.93
lr schd: 160000
memory (GB): 4.9
Name: fpn_r50_512x512_160k_ade20k
Results:
- Task: Semantic Segmentation
Dataset: ADE20K
Metrics:
mIoU: 37.49
mIoU(ms+flip): 39.09
Task: Semantic Segmentation
Config: configs/sem_fpn/fpn_r50_512x512_160k_ade20k.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/sem_fpn/fpn_r50_512x512_160k_ade20k/fpn_r50_512x512_160k_ade20k_20200718_131734-5b5a6ab9.pth
- Config: configs/sem_fpn/fpn_r101_512x512_160k_ade20k.py
- Name: fpn_r101_512x512_160k_ade20k
In Collection: sem_fpn
Metadata:
backbone: R-101
crop size: (512,512)
lr schd: 160000
inference time (ms/im):
- backend: PyTorch
batch size: 1
- value: 24.64
hardware: V100
backend: PyTorch
batch size: 1
mode: FP32
resolution: (512,512)
value: 24.64
lr schd: 160000
memory (GB): 5.9
Name: fpn_r101_512x512_160k_ade20k
Results:
- Task: Semantic Segmentation
Dataset: ADE20K
Metrics:
mIoU: 39.35
mIoU(ms+flip): 40.72
Task: Semantic Segmentation
Config: configs/sem_fpn/fpn_r101_512x512_160k_ade20k.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/sem_fpn/fpn_r101_512x512_160k_ade20k/fpn_r101_512x512_160k_ade20k_20200718_131734-306b5004.pth

13
configs/setr/README.md
View File

@ -4,6 +4,17 @@
<!-- [ALGORITHM] -->
<a href="https://github.com/fudan-zvg/SETR">Official Repo</a>
<a href="https://github.com/open-mmlab/mmsegmentation/blob/v0.17.0/mmseg/models/decode_heads/setr_up_head.py#L11">Code Snippet</a>
```None
This head has two version head.
```
<details>
<summary align="right"><a href="https://arxiv.org/abs/2012.15840">SETR (CVPR'2021)</a></summary>
```latex
@article{zheng2020rethinking,
title={Rethinking Semantic Segmentation from a Sequence-to-Sequence Perspective with Transformers},
@ -13,6 +24,8 @@
}
```
</details>
## Results and models
### ADE20K

62
configs/setr/setr.yml
View File

@ -1,87 +1,97 @@
Collections:
- Metadata:
- Name: setr
Metadata:
Training Data:
- ADE20K
Name: setr
Paper:
URL: https://arxiv.org/abs/2012.15840
Title: Rethinking Semantic Segmentation from a Sequence-to-Sequence Perspective
with Transformers
README: configs/setr/README.md
Code:
URL: https://github.com/open-mmlab/mmsegmentation/blob/v0.17.0/mmseg/models/decode_heads/setr_up_head.py#L11
Version: v0.17.0
Converted From:
Code: https://github.com/fudan-zvg/SETR
Models:
- Config: configs/setr/setr_naive_512x512_160k_b16_ade20k.py
- Name: setr_naive_512x512_160k_b16_ade20k
In Collection: setr
Metadata:
backbone: ViT-L
crop size: (512,512)
lr schd: 160000
inference time (ms/im):
- backend: PyTorch
batch size: 1
- value: 211.86
hardware: V100
backend: PyTorch
batch size: 1
mode: FP32
resolution: (512,512)
value: 211.86
lr schd: 160000
memory (GB): 18.4
Name: setr_naive_512x512_160k_b16_ade20k
Results:
- Task: Semantic Segmentation
Dataset: ADE20K
Metrics:
mIoU: 48.28
mIoU(ms+flip): 49.56
Task: Semantic Segmentation
Config: configs/setr/setr_naive_512x512_160k_b16_ade20k.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/setr/setr_naive_512x512_160k_b16_ade20k/setr_naive_512x512_160k_b16_ade20k_20210619_191258-061f24f5.pth
- Config: configs/setr/setr_pup_512x512_160k_b16_ade20k.py
- Name: setr_pup_512x512_160k_b16_ade20k
In Collection: setr
Metadata:
backbone: ViT-L
crop size: (512,512)
lr schd: 160000
inference time (ms/im):
- backend: PyTorch
batch size: 1
- value: 222.22
hardware: V100
backend: PyTorch
batch size: 1
mode: FP32
resolution: (512,512)
value: 222.22
lr schd: 160000
memory (GB): 19.54
Name: setr_pup_512x512_160k_b16_ade20k
Results:
- Task: Semantic Segmentation
Dataset: ADE20K
Metrics:
mIoU: 48.24
mIoU(ms+flip): 49.99
Task: Semantic Segmentation
Config: configs/setr/setr_pup_512x512_160k_b16_ade20k.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/setr/setr_pup_512x512_160k_b16_ade20k/setr_pup_512x512_160k_b16_ade20k_20210619_191343-7e0ce826.pth
- Config: configs/setr/setr_mla_512x512_160k_b8_ade20k.py
- Name: setr_mla_512x512_160k_b8_ade20k
In Collection: setr
Metadata:
backbone: ViT-L
crop size: (512,512)
lr schd: 160000
memory (GB): 10.96
Name: setr_mla_512x512_160k_b8_ade20k
Results:
- Task: Semantic Segmentation
Dataset: ADE20K
Metrics:
mIoU: 47.34
mIoU(ms+flip): 49.05
Task: Semantic Segmentation
Config: configs/setr/setr_mla_512x512_160k_b8_ade20k.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/setr/setr_mla_512x512_160k_b8_ade20k/setr_mla_512x512_160k_b8_ade20k_20210619_191118-c6d21df0.pth
- Config: configs/setr/setr_mla_512x512_160k_b16_ade20k.py
- Name: setr_mla_512x512_160k_b16_ade20k
In Collection: setr
Metadata:
backbone: ViT-L
crop size: (512,512)
lr schd: 160000
inference time (ms/im):
- backend: PyTorch
batch size: 1
- value: 190.48
hardware: V100
backend: PyTorch
batch size: 1
mode: FP32
resolution: (512,512)
value: 190.48
lr schd: 160000
memory (GB): 17.3
Name: setr_mla_512x512_160k_b16_ade20k
Results:
- Task: Semantic Segmentation
Dataset: ADE20K
Metrics:
mIoU: 47.54
mIoU(ms+flip): 49.37
Task: Semantic Segmentation
Config: configs/setr/setr_mla_512x512_160k_b16_ade20k.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/setr/setr_mla_512x512_160k_b16_ade20k/setr_mla_512x512_160k_b16_ade20k_20210619_191057-f9741de7.pth

9
configs/swin/README.md
View File

@ -4,6 +4,13 @@
<!-- [ALGORITHM] -->
<a href="https://github.com/microsoft/Swin-Transformer">Official Repo</a>
<a href="https://github.com/open-mmlab/mmsegmentation/blob/v0.17.0/mmseg/models/backbones/swin.py#L524">Code Snippet</a>
<details>
<summary align="right"><a href="https://arxiv.org/abs/2103.14030">Swin Transformer (arXiv'2021)</a></summary>
```latex
@article{liu2021Swin,
title={Swin Transformer: Hierarchical Vision Transformer using Shifted Windows},
@ -13,6 +20,8 @@
}
```
</details>
## Usage
To use other repositories' pre-trained models, it is necessary to convert keys.

81
configs/swin/swin.yml
View File

@ -1,122 +1,131 @@
Collections:
- Metadata:
- Name: swin
Metadata:
Training Data:
- ADE20K
Name: swin
Paper:
URL: https://arxiv.org/abs/2103.14030
Title: 'Swin Transformer: Hierarchical Vision Transformer using Shifted Windows'
README: configs/swin/README.md
Code:
URL: https://github.com/open-mmlab/mmsegmentation/blob/v0.17.0/mmseg/models/backbones/swin.py#L524
Version: v0.17.0
Converted From:
Code: https://github.com/microsoft/Swin-Transformer
Models:
- Config: configs/swin/upernet_swin_tiny_patch4_window7_512x512_160k_ade20k_pretrain_224x224_1K.py
- Name: upernet_swin_tiny_patch4_window7_512x512_160k_ade20k_pretrain_224x224_1K
In Collection: swin
Metadata:
backbone: Swin-T
crop size: (512,512)
lr schd: 160000
inference time (ms/im):
- backend: PyTorch
batch size: 1
- value: 47.48
hardware: V100
backend: PyTorch
batch size: 1
mode: FP32
resolution: (512,512)
value: 47.48
lr schd: 160000
memory (GB): 5.02
Name: upernet_swin_tiny_patch4_window7_512x512_160k_ade20k_pretrain_224x224_1K
Results:
- Task: Semantic Segmentation
Dataset: ADE20K
Metrics:
mIoU: 44.41
mIoU(ms+flip): 45.79
Task: Semantic Segmentation
Config: configs/swin/upernet_swin_tiny_patch4_window7_512x512_160k_ade20k_pretrain_224x224_1K.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/swin/upernet_swin_tiny_patch4_window7_512x512_160k_ade20k_pretrain_224x224_1K/upernet_swin_tiny_patch4_window7_512x512_160k_ade20k_pretrain_224x224_1K_20210531_112542-e380ad3e.pth
- Config: configs/swin/upernet_swin_small_patch4_window7_512x512_160k_ade20k_pretrain_224x224_1K.py
- Name: upernet_swin_small_patch4_window7_512x512_160k_ade20k_pretrain_224x224_1K
In Collection: swin
Metadata:
backbone: Swin-S
crop size: (512,512)
lr schd: 160000
inference time (ms/im):
- backend: PyTorch
batch size: 1
- value: 67.93
hardware: V100
backend: PyTorch
batch size: 1
mode: FP32
resolution: (512,512)
value: 67.93
lr schd: 160000
memory (GB): 6.17
Name: upernet_swin_small_patch4_window7_512x512_160k_ade20k_pretrain_224x224_1K
Results:
- Task: Semantic Segmentation
Dataset: ADE20K
Metrics:
mIoU: 47.72
mIoU(ms+flip): 49.24
Task: Semantic Segmentation
Config: configs/swin/upernet_swin_small_patch4_window7_512x512_160k_ade20k_pretrain_224x224_1K.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/swin/upernet_swin_small_patch4_window7_512x512_160k_ade20k_pretrain_224x224_1K/upernet_swin_small_patch4_window7_512x512_160k_ade20k_pretrain_224x224_1K_20210526_192015-ee2fff1c.pth
- Config: configs/swin/upernet_swin_base_patch4_window7_512x512_160k_ade20k_pretrain_224x224_1K.py
- Name: upernet_swin_base_patch4_window7_512x512_160k_ade20k_pretrain_224x224_1K
In Collection: swin
Metadata:
backbone: Swin-B
crop size: (512,512)
lr schd: 160000
inference time (ms/im):
- backend: PyTorch
batch size: 1
- value: 79.05
hardware: V100
backend: PyTorch
batch size: 1
mode: FP32
resolution: (512,512)
value: 79.05
lr schd: 160000
memory (GB): 7.61
Name: upernet_swin_base_patch4_window7_512x512_160k_ade20k_pretrain_224x224_1K
Results:
- Task: Semantic Segmentation
Dataset: ADE20K
Metrics:
mIoU: 47.99
mIoU(ms+flip): 49.57
Task: Semantic Segmentation
Config: configs/swin/upernet_swin_base_patch4_window7_512x512_160k_ade20k_pretrain_224x224_1K.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/swin/upernet_swin_base_patch4_window7_512x512_160k_ade20k_pretrain_224x224_1K/upernet_swin_base_patch4_window7_512x512_160k_ade20k_pretrain_224x224_1K_20210526_192340-593b0e13.pth
- Config: configs/swin/upernet_swin_base_patch4_window7_512x512_160k_ade20k_pretrain_224x224_22K.py
- Name: upernet_swin_base_patch4_window7_512x512_160k_ade20k_pretrain_224x224_22K
In Collection: swin
Metadata:
backbone: Swin-B
crop size: (512,512)
lr schd: 160000
Name: upernet_swin_base_patch4_window7_512x512_160k_ade20k_pretrain_224x224_22K
Results:
- Task: Semantic Segmentation
Dataset: ADE20K
Metrics:
mIoU: 50.31
mIoU(ms+flip): 51.9
Task: Semantic Segmentation
Config: configs/swin/upernet_swin_base_patch4_window7_512x512_160k_ade20k_pretrain_224x224_22K.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/swin/upernet_swin_base_patch4_window7_512x512_160k_ade20k_pretrain_224x224_22K/upernet_swin_base_patch4_window7_512x512_160k_ade20k_pretrain_224x224_22K_20210526_211650-762e2178.pth
- Config: configs/swin/upernet_swin_base_patch4_window12_512x512_160k_ade20k_pretrain_384x384_1K.py
- Name: upernet_swin_base_patch4_window12_512x512_160k_ade20k_pretrain_384x384_1K
In Collection: swin
Metadata:
backbone: Swin-B
crop size: (512,512)
lr schd: 160000
inference time (ms/im):
- backend: PyTorch
batch size: 1
- value: 82.64
hardware: V100
backend: PyTorch
batch size: 1
mode: FP32
resolution: (512,512)
value: 82.64
lr schd: 160000
memory (GB): 8.52
Name: upernet_swin_base_patch4_window12_512x512_160k_ade20k_pretrain_384x384_1K
Results:
- Task: Semantic Segmentation
Dataset: ADE20K
Metrics:
mIoU: 48.35
mIoU(ms+flip): 49.65
Task: Semantic Segmentation
Config: configs/swin/upernet_swin_base_patch4_window12_512x512_160k_ade20k_pretrain_384x384_1K.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/swin/upernet_swin_base_patch4_window12_512x512_160k_ade20k_pretrain_384x384_1K/upernet_swin_base_patch4_window12_512x512_160k_ade20k_pretrain_384x384_1K_20210531_132020-05b22ea4.pth
- Config: configs/swin/upernet_swin_base_patch4_window12_512x512_160k_ade20k_pretrain_384x384_22K.py
- Name: upernet_swin_base_patch4_window12_512x512_160k_ade20k_pretrain_384x384_22K
In Collection: swin
Metadata:
backbone: Swin-B
crop size: (512,512)
lr schd: 160000
Name: upernet_swin_base_patch4_window12_512x512_160k_ade20k_pretrain_384x384_22K
Results:
- Task: Semantic Segmentation
Dataset: ADE20K
Metrics:
mIoU: 50.76
mIoU(ms+flip): 52.4
Task: Semantic Segmentation
Config: configs/swin/upernet_swin_base_patch4_window12_512x512_160k_ade20k_pretrain_384x384_22K.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/swin/upernet_swin_base_patch4_window12_512x512_160k_ade20k_pretrain_384x384_22K/upernet_swin_base_patch4_window12_512x512_160k_ade20k_pretrain_384x384_22K_20210531_125459-429057bf.pth

9
configs/unet/README.md
View File

@ -4,6 +4,13 @@
<!-- [ALGORITHM] -->
<a href="http://lmb.informatik.uni-freiburg.de/people/ronneber/u-net">Official Repo</a>
<a href="https://github.com/open-mmlab/mmsegmentation/blob/v0.17.0/mmseg/models/backbones/unet.py#L225">Code Snippet</a>
<details>
<summary align="right"><a href="https://arxiv.org/abs/1505.04597">UNet (MICCAI'2016/Nat. Methods'2019)</a></summary>
```latex
@inproceedings{ronneberger2015u,
title={U-net: Convolutional networks for biomedical image segmentation},
@ -15,6 +22,8 @@
}
```
</details>
## Results and models
### DRIVE

85
configs/unet/unet.yml
View File

@ -1,177 +1,186 @@
Collections:
- Metadata:
- Name: unet
Metadata:
Training Data:
- DRIVE
- STARE
- CHASE_DB1
- HRF
Name: unet
Paper:
URL: https://arxiv.org/abs/1505.04597
Title: 'U-Net: Convolutional Networks for Biomedical Image Segmentation'
README: configs/unet/README.md
Code:
URL: https://github.com/open-mmlab/mmsegmentation/blob/v0.17.0/mmseg/models/backbones/unet.py#L225
Version: v0.17.0
Converted From:
Code: http://lmb.informatik.uni-freiburg.de/people/ronneber/u-net
Models:
- Config: configs/unet/fcn_unet_s5-d16_64x64_40k_drive.py
- Name: fcn_unet_s5-d16_64x64_40k_drive
In Collection: unet
Metadata:
backbone: UNet-S5-D16
crop size: (64,64)
lr schd: 40000
memory (GB): 0.68
Name: fcn_unet_s5-d16_64x64_40k_drive
Results:
- Task: Semantic Segmentation
Dataset: DRIVE
Metrics:
mIoU: 78.67
Task: Semantic Segmentation
Config: configs/unet/fcn_unet_s5-d16_64x64_40k_drive.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/unet/fcn_unet_s5-d16_64x64_40k_drive/fcn_unet_s5-d16_64x64_40k_drive_20201223_191051-5daf6d3b.pth
- Config: configs/unet/pspnet_unet_s5-d16_64x64_40k_drive.py
- Name: pspnet_unet_s5-d16_64x64_40k_drive
In Collection: unet
Metadata:
backbone: UNet-S5-D16
crop size: (64,64)
lr schd: 40000
memory (GB): 0.599
Name: pspnet_unet_s5-d16_64x64_40k_drive
Results:
- Task: Semantic Segmentation
Dataset: DRIVE
Metrics:
mIoU: 78.62
Task: Semantic Segmentation
Config: configs/unet/pspnet_unet_s5-d16_64x64_40k_drive.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/unet/pspnet_unet_s5-d16_64x64_40k_drive/pspnet_unet_s5-d16_64x64_40k_drive_20201227_181818-aac73387.pth
- Config: configs/unet/deeplabv3_unet_s5-d16_64x64_40k_drive.py
- Name: deeplabv3_unet_s5-d16_64x64_40k_drive
In Collection: unet
Metadata:
backbone: UNet-S5-D16
crop size: (64,64)
lr schd: 40000
memory (GB): 0.596
Name: deeplabv3_unet_s5-d16_64x64_40k_drive
Results:
- Task: Semantic Segmentation
Dataset: DRIVE
Metrics:
mIoU: 78.69
Task: Semantic Segmentation
Config: configs/unet/deeplabv3_unet_s5-d16_64x64_40k_drive.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/unet/deeplabv3_unet_s5-d16_64x64_40k_drive/deeplabv3_unet_s5-d16_64x64_40k_drive_20201226_094047-0671ff20.pth
- Config: configs/unet/fcn_unet_s5-d16_128x128_40k_stare.py
- Name: fcn_unet_s5-d16_128x128_40k_stare
In Collection: unet
Metadata:
backbone: UNet-S5-D16
crop size: (128,128)
lr schd: 40000
memory (GB): 0.968
Name: fcn_unet_s5-d16_128x128_40k_stare
Results:
- Task: Semantic Segmentation
Dataset: STARE
Metrics:
mIoU: 81.02
Task: Semantic Segmentation
Config: configs/unet/fcn_unet_s5-d16_128x128_40k_stare.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/unet/fcn_unet_s5-d16_128x128_40k_stare/fcn_unet_s5-d16_128x128_40k_stare_20201223_191051-7d77e78b.pth
- Config: configs/unet/pspnet_unet_s5-d16_128x128_40k_stare.py
- Name: pspnet_unet_s5-d16_128x128_40k_stare
In Collection: unet
Metadata:
backbone: UNet-S5-D16
crop size: (128,128)
lr schd: 40000
memory (GB): 0.982
Name: pspnet_unet_s5-d16_128x128_40k_stare
Results:
- Task: Semantic Segmentation
Dataset: STARE
Metrics:
mIoU: 81.22
Task: Semantic Segmentation
Config: configs/unet/pspnet_unet_s5-d16_128x128_40k_stare.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/unet/pspnet_unet_s5-d16_128x128_40k_stare/pspnet_unet_s5-d16_128x128_40k_stare_20201227_181818-3c2923c4.pth
- Config: configs/unet/deeplabv3_unet_s5-d16_128x128_40k_stare.py
- Name: deeplabv3_unet_s5-d16_128x128_40k_stare
In Collection: unet
Metadata:
backbone: UNet-S5-D16
crop size: (128,128)
lr schd: 40000
memory (GB): 0.999
Name: deeplabv3_unet_s5-d16_128x128_40k_stare
Results:
- Task: Semantic Segmentation
Dataset: STARE
Metrics:
mIoU: 80.93
Task: Semantic Segmentation
Config: configs/unet/deeplabv3_unet_s5-d16_128x128_40k_stare.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/unet/deeplabv3_unet_s5-d16_128x128_40k_stare/deeplabv3_unet_s5-d16_128x128_40k_stare_20201226_094047-93dcb93c.pth
- Config: configs/unet/fcn_unet_s5-d16_128x128_40k_chase_db1.py
- Name: fcn_unet_s5-d16_128x128_40k_chase_db1
In Collection: unet
Metadata:
backbone: UNet-S5-D16
crop size: (128,128)
lr schd: 40000
memory (GB): 0.968
Name: fcn_unet_s5-d16_128x128_40k_chase_db1
Results:
- Task: Semantic Segmentation
Dataset: CHASE_DB1
Metrics:
mIoU: 80.24
Task: Semantic Segmentation
Config: configs/unet/fcn_unet_s5-d16_128x128_40k_chase_db1.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/unet/fcn_unet_s5-d16_128x128_40k_chase_db1/fcn_unet_s5-d16_128x128_40k_chase_db1_20201223_191051-11543527.pth
- Config: configs/unet/pspnet_unet_s5-d16_128x128_40k_chase_db1.py
- Name: pspnet_unet_s5-d16_128x128_40k_chase_db1
In Collection: unet
Metadata:
backbone: UNet-S5-D16
crop size: (128,128)
lr schd: 40000
memory (GB): 0.982
Name: pspnet_unet_s5-d16_128x128_40k_chase_db1
Results:
- Task: Semantic Segmentation
Dataset: CHASE_DB1
Metrics:
mIoU: 80.36
Task: Semantic Segmentation
Config: configs/unet/pspnet_unet_s5-d16_128x128_40k_chase_db1.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/unet/pspnet_unet_s5-d16_128x128_40k_chase_db1/pspnet_unet_s5-d16_128x128_40k_chase_db1_20201227_181818-68d4e609.pth
- Config: configs/unet/deeplabv3_unet_s5-d16_128x128_40k_chase_db1.py
- Name: deeplabv3_unet_s5-d16_128x128_40k_chase_db1
In Collection: unet
Metadata:
backbone: UNet-S5-D16
crop size: (128,128)
lr schd: 40000
memory (GB): 0.999
Name: deeplabv3_unet_s5-d16_128x128_40k_chase_db1
Results:
- Task: Semantic Segmentation
Dataset: CHASE_DB1
Metrics:
mIoU: 80.47
Task: Semantic Segmentation
Config: configs/unet/deeplabv3_unet_s5-d16_128x128_40k_chase_db1.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/unet/deeplabv3_unet_s5-d16_128x128_40k_chase_db1/deeplabv3_unet_s5-d16_128x128_40k_chase_db1_20201226_094047-4c5aefa3.pth
- Config: configs/unet/fcn_unet_s5-d16_256x256_40k_hrf.py
- Name: fcn_unet_s5-d16_256x256_40k_hrf
In Collection: unet
Metadata:
backbone: UNet-S5-D16
crop size: (256,256)
lr schd: 40000
memory (GB): 2.525
Name: fcn_unet_s5-d16_256x256_40k_hrf
Results:
- Task: Semantic Segmentation
Dataset: HRF
Metrics:
mIoU: 79.45
Task: Semantic Segmentation
Config: configs/unet/fcn_unet_s5-d16_256x256_40k_hrf.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/unet/fcn_unet_s5-d16_256x256_40k_hrf/fcn_unet_s5-d16_256x256_40k_hrf_20201223_173724-d89cf1ed.pth
- Config: configs/unet/pspnet_unet_s5-d16_256x256_40k_hrf.py
- Name: pspnet_unet_s5-d16_256x256_40k_hrf
In Collection: unet
Metadata:
backbone: UNet-S5-D16
crop size: (256,256)
lr schd: 40000
memory (GB): 2.588
Name: pspnet_unet_s5-d16_256x256_40k_hrf
Results:
- Task: Semantic Segmentation
Dataset: HRF
Metrics:
mIoU: 80.07
Task: Semantic Segmentation
Config: configs/unet/pspnet_unet_s5-d16_256x256_40k_hrf.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/unet/pspnet_unet_s5-d16_256x256_40k_hrf/pspnet_unet_s5-d16_256x256_40k_hrf_20201227_181818-fdb7e29b.pth
- Config: configs/unet/deeplabv3_unet_s5-d16_256x256_40k_hrf.py
- Name: deeplabv3_unet_s5-d16_256x256_40k_hrf
In Collection: unet
Metadata:
backbone: UNet-S5-D16
crop size: (256,256)
lr schd: 40000
memory (GB): 2.604
Name: deeplabv3_unet_s5-d16_256x256_40k_hrf
Results:
- Task: Semantic Segmentation
Dataset: HRF
Metrics:
mIoU: 80.21
Task: Semantic Segmentation
Config: configs/unet/deeplabv3_unet_s5-d16_256x256_40k_hrf.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/unet/deeplabv3_unet_s5-d16_256x256_40k_hrf/deeplabv3_unet_s5-d16_256x256_40k_hrf_20201226_094047-3a1fdf85.pth

9
configs/upernet/README.md
View File

@ -4,6 +4,13 @@
<!-- [ALGORITHM] -->
<a href="https://github.com/CSAILVision/unifiedparsing">Official Repo</a>
<a href="https://github.com/open-mmlab/mmsegmentation/blob/v0.17.0/mmseg/models/decode_heads/uper_head.py#L13">Code Snippet</a>
<details>
<summary align="right"><a href="https://arxiv.org/pdf/1807.10221.pdf">UPerNet (ECCV'2018)</a></summary>
```latex
@inproceedings{xiao2018unified,
title={Unified perceptual parsing for scene understanding},
@ -14,6 +21,8 @@
}
```
</details>
## Results and models
### Cityscapes

173
configs/upernet/upernet.yml
View File

@ -1,296 +1,305 @@
Collections:
- Metadata:
- Name: upernet
Metadata:
Training Data:
- Cityscapes
- ADE20K
- Pascal VOC 2012 + Aug
Name: upernet
Paper:
URL: https://arxiv.org/pdf/1807.10221.pdf
Title: Unified Perceptual Parsing for Scene Understanding
README: configs/upernet/README.md
Code:
URL: https://github.com/open-mmlab/mmsegmentation/blob/v0.17.0/mmseg/models/decode_heads/uper_head.py#L13
Version: v0.17.0
Converted From:
Code: https://github.com/CSAILVision/unifiedparsing
Models:
- Config: configs/upernet/upernet_r50_512x1024_40k_cityscapes.py
- Name: upernet_r50_512x1024_40k_cityscapes
In Collection: upernet
Metadata:
backbone: R-50
crop size: (512,1024)
lr schd: 40000
inference time (ms/im):
- backend: PyTorch
batch size: 1
- value: 235.29
hardware: V100
backend: PyTorch
batch size: 1
mode: FP32
resolution: (512,1024)
value: 235.29
lr schd: 40000
memory (GB): 6.4
Name: upernet_r50_512x1024_40k_cityscapes
Results:
- Task: Semantic Segmentation
Dataset: Cityscapes
Metrics:
mIoU: 77.1
mIoU(ms+flip): 78.37
Task: Semantic Segmentation
Config: configs/upernet/upernet_r50_512x1024_40k_cityscapes.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/upernet/upernet_r50_512x1024_40k_cityscapes/upernet_r50_512x1024_40k_cityscapes_20200605_094827-aa54cb54.pth
- Config: configs/upernet/upernet_r101_512x1024_40k_cityscapes.py
- Name: upernet_r101_512x1024_40k_cityscapes
In Collection: upernet
Metadata:
backbone: R-101
crop size: (512,1024)
lr schd: 40000
inference time (ms/im):
- backend: PyTorch
batch size: 1
- value: 263.85
hardware: V100
backend: PyTorch
batch size: 1
mode: FP32
resolution: (512,1024)
value: 263.85
lr schd: 40000
memory (GB): 7.4
Name: upernet_r101_512x1024_40k_cityscapes
Results:
- Task: Semantic Segmentation
Dataset: Cityscapes
Metrics:
mIoU: 78.69
mIoU(ms+flip): 80.11
Task: Semantic Segmentation
Config: configs/upernet/upernet_r101_512x1024_40k_cityscapes.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/upernet/upernet_r101_512x1024_40k_cityscapes/upernet_r101_512x1024_40k_cityscapes_20200605_094933-ebce3b10.pth
- Config: configs/upernet/upernet_r50_769x769_40k_cityscapes.py
- Name: upernet_r50_769x769_40k_cityscapes
In Collection: upernet
Metadata:
backbone: R-50
crop size: (769,769)
lr schd: 40000
inference time (ms/im):
- backend: PyTorch
batch size: 1
- value: 568.18
hardware: V100
backend: PyTorch
batch size: 1
mode: FP32
resolution: (769,769)
value: 568.18
lr schd: 40000
memory (GB): 7.2
Name: upernet_r50_769x769_40k_cityscapes
Results:
- Task: Semantic Segmentation
Dataset: Cityscapes
Metrics:
mIoU: 77.98
mIoU(ms+flip): 79.7
Task: Semantic Segmentation
Config: configs/upernet/upernet_r50_769x769_40k_cityscapes.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/upernet/upernet_r50_769x769_40k_cityscapes/upernet_r50_769x769_40k_cityscapes_20200530_033048-92d21539.pth
- Config: configs/upernet/upernet_r101_769x769_40k_cityscapes.py
- Name: upernet_r101_769x769_40k_cityscapes
In Collection: upernet
Metadata:
backbone: R-101
crop size: (769,769)
lr schd: 40000
inference time (ms/im):
- backend: PyTorch
batch size: 1
- value: 641.03
hardware: V100
backend: PyTorch
batch size: 1
mode: FP32
resolution: (769,769)
value: 641.03
lr schd: 40000
memory (GB): 8.4
Name: upernet_r101_769x769_40k_cityscapes
Results:
- Task: Semantic Segmentation
Dataset: Cityscapes
Metrics:
mIoU: 79.03
mIoU(ms+flip): 80.77
Task: Semantic Segmentation
Config: configs/upernet/upernet_r101_769x769_40k_cityscapes.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/upernet/upernet_r101_769x769_40k_cityscapes/upernet_r101_769x769_40k_cityscapes_20200530_040819-83c95d01.pth
- Config: configs/upernet/upernet_r50_512x1024_80k_cityscapes.py
- Name: upernet_r50_512x1024_80k_cityscapes
In Collection: upernet
Metadata:
backbone: R-50
crop size: (512,1024)
lr schd: 80000
Name: upernet_r50_512x1024_80k_cityscapes
Results:
- Task: Semantic Segmentation
Dataset: Cityscapes
Metrics:
mIoU: 78.19
mIoU(ms+flip): 79.19
Task: Semantic Segmentation
Config: configs/upernet/upernet_r50_512x1024_80k_cityscapes.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/upernet/upernet_r50_512x1024_80k_cityscapes/upernet_r50_512x1024_80k_cityscapes_20200607_052207-848beca8.pth
- Config: configs/upernet/upernet_r101_512x1024_80k_cityscapes.py
- Name: upernet_r101_512x1024_80k_cityscapes
In Collection: upernet
Metadata:
backbone: R-101
crop size: (512,1024)
lr schd: 80000
Name: upernet_r101_512x1024_80k_cityscapes
Results:
- Task: Semantic Segmentation
Dataset: Cityscapes
Metrics:
mIoU: 79.4
mIoU(ms+flip): 80.46
Task: Semantic Segmentation
Config: configs/upernet/upernet_r101_512x1024_80k_cityscapes.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/upernet/upernet_r101_512x1024_80k_cityscapes/upernet_r101_512x1024_80k_cityscapes_20200607_002403-f05f2345.pth
- Config: configs/upernet/upernet_r50_769x769_80k_cityscapes.py
- Name: upernet_r50_769x769_80k_cityscapes
In Collection: upernet
Metadata:
backbone: R-50
crop size: (769,769)
lr schd: 80000
Name: upernet_r50_769x769_80k_cityscapes
Results:
- Task: Semantic Segmentation
Dataset: Cityscapes
Metrics:
mIoU: 79.39
mIoU(ms+flip): 80.92
Task: Semantic Segmentation
Config: configs/upernet/upernet_r50_769x769_80k_cityscapes.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/upernet/upernet_r50_769x769_80k_cityscapes/upernet_r50_769x769_80k_cityscapes_20200607_005107-82ae7d15.pth
- Config: configs/upernet/upernet_r101_769x769_80k_cityscapes.py
- Name: upernet_r101_769x769_80k_cityscapes
In Collection: upernet
Metadata:
backbone: R-101
crop size: (769,769)
lr schd: 80000
Name: upernet_r101_769x769_80k_cityscapes
Results:
- Task: Semantic Segmentation
Dataset: Cityscapes
Metrics:
mIoU: 80.1
mIoU(ms+flip): 81.49
Task: Semantic Segmentation
Config: configs/upernet/upernet_r101_769x769_80k_cityscapes.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/upernet/upernet_r101_769x769_80k_cityscapes/upernet_r101_769x769_80k_cityscapes_20200607_001014-082fc334.pth
- Config: configs/upernet/upernet_r50_512x512_80k_ade20k.py
- Name: upernet_r50_512x512_80k_ade20k
In Collection: upernet
Metadata:
backbone: R-50
crop size: (512,512)
lr schd: 80000
inference time (ms/im):
- backend: PyTorch
batch size: 1
- value: 42.74
hardware: V100
backend: PyTorch
batch size: 1
mode: FP32
resolution: (512,512)
value: 42.74
lr schd: 80000
memory (GB): 8.1
Name: upernet_r50_512x512_80k_ade20k
Results:
- Task: Semantic Segmentation
Dataset: ADE20K
Metrics:
mIoU: 40.7
mIoU(ms+flip): 41.81
Task: Semantic Segmentation
Config: configs/upernet/upernet_r50_512x512_80k_ade20k.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/upernet/upernet_r50_512x512_80k_ade20k/upernet_r50_512x512_80k_ade20k_20200614_144127-ecc8377b.pth
- Config: configs/upernet/upernet_r101_512x512_80k_ade20k.py
- Name: upernet_r101_512x512_80k_ade20k
In Collection: upernet
Metadata:
backbone: R-101
crop size: (512,512)
lr schd: 80000
inference time (ms/im):
- backend: PyTorch
batch size: 1
- value: 49.16
hardware: V100
backend: PyTorch
batch size: 1
mode: FP32
resolution: (512,512)
value: 49.16
lr schd: 80000
memory (GB): 9.1
Name: upernet_r101_512x512_80k_ade20k
Results:
- Task: Semantic Segmentation
Dataset: ADE20K
Metrics:
mIoU: 42.91
mIoU(ms+flip): 43.96
Task: Semantic Segmentation
Config: configs/upernet/upernet_r101_512x512_80k_ade20k.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/upernet/upernet_r101_512x512_80k_ade20k/upernet_r101_512x512_80k_ade20k_20200614_185117-32e4db94.pth
- Config: configs/upernet/upernet_r50_512x512_160k_ade20k.py
- Name: upernet_r50_512x512_160k_ade20k
In Collection: upernet
Metadata:
backbone: R-50
crop size: (512,512)
lr schd: 160000
Name: upernet_r50_512x512_160k_ade20k
Results:
- Task: Semantic Segmentation
Dataset: ADE20K
Metrics:
mIoU: 42.05
mIoU(ms+flip): 42.78
Task: Semantic Segmentation
Config: configs/upernet/upernet_r50_512x512_160k_ade20k.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/upernet/upernet_r50_512x512_160k_ade20k/upernet_r50_512x512_160k_ade20k_20200615_184328-8534de8d.pth
- Config: configs/upernet/upernet_r101_512x512_160k_ade20k.py
- Name: upernet_r101_512x512_160k_ade20k
In Collection: upernet
Metadata:
backbone: R-101
crop size: (512,512)
lr schd: 160000
Name: upernet_r101_512x512_160k_ade20k
Results:
- Task: Semantic Segmentation
Dataset: ADE20K
Metrics:
mIoU: 43.82
mIoU(ms+flip): 44.85
Task: Semantic Segmentation
Config: configs/upernet/upernet_r101_512x512_160k_ade20k.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/upernet/upernet_r101_512x512_160k_ade20k/upernet_r101_512x512_160k_ade20k_20200615_161951-91b32684.pth
- Config: configs/upernet/upernet_r50_512x512_20k_voc12aug.py
- Name: upernet_r50_512x512_20k_voc12aug
In Collection: upernet
Metadata:
backbone: R-50
crop size: (512,512)
lr schd: 20000
inference time (ms/im):
- backend: PyTorch
batch size: 1
- value: 43.16
hardware: V100
backend: PyTorch
batch size: 1
mode: FP32
resolution: (512,512)
value: 43.16
lr schd: 20000
memory (GB): 6.4
Name: upernet_r50_512x512_20k_voc12aug
Results:
- Task: Semantic Segmentation
Dataset: Pascal VOC 2012 + Aug
Metrics:
mIoU: 74.82
mIoU(ms+flip): 76.35
Task: Semantic Segmentation
Config: configs/upernet/upernet_r50_512x512_20k_voc12aug.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/upernet/upernet_r50_512x512_20k_voc12aug/upernet_r50_512x512_20k_voc12aug_20200617_165330-5b5890a7.pth
- Config: configs/upernet/upernet_r101_512x512_20k_voc12aug.py
- Name: upernet_r101_512x512_20k_voc12aug
In Collection: upernet
Metadata:
backbone: R-101
crop size: (512,512)
lr schd: 20000
inference time (ms/im):
- backend: PyTorch
batch size: 1
- value: 50.05
hardware: V100
backend: PyTorch
batch size: 1
mode: FP32
resolution: (512,512)
value: 50.05
lr schd: 20000
memory (GB): 7.5
Name: upernet_r101_512x512_20k_voc12aug
Results:
- Task: Semantic Segmentation
Dataset: Pascal VOC 2012 + Aug
Metrics:
mIoU: 77.1
mIoU(ms+flip): 78.29
Task: Semantic Segmentation
Config: configs/upernet/upernet_r101_512x512_20k_voc12aug.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/upernet/upernet_r101_512x512_20k_voc12aug/upernet_r101_512x512_20k_voc12aug_20200617_165629-f14e7f27.pth
- Config: configs/upernet/upernet_r50_512x512_40k_voc12aug.py
- Name: upernet_r50_512x512_40k_voc12aug
In Collection: upernet
Metadata:
backbone: R-50
crop size: (512,512)
lr schd: 40000
Name: upernet_r50_512x512_40k_voc12aug
Results:
- Task: Semantic Segmentation
Dataset: Pascal VOC 2012 + Aug
Metrics:
mIoU: 75.92
mIoU(ms+flip): 77.44
Task: Semantic Segmentation
Config: configs/upernet/upernet_r50_512x512_40k_voc12aug.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/upernet/upernet_r50_512x512_40k_voc12aug/upernet_r50_512x512_40k_voc12aug_20200613_162257-ca9bcc6b.pth
- Config: configs/upernet/upernet_r101_512x512_40k_voc12aug.py
- Name: upernet_r101_512x512_40k_voc12aug
In Collection: upernet
Metadata:
backbone: R-101
crop size: (512,512)
lr schd: 40000
Name: upernet_r101_512x512_40k_voc12aug
Results:
- Task: Semantic Segmentation
Dataset: Pascal VOC 2012 + Aug
Metrics:
mIoU: 77.43
mIoU(ms+flip): 78.56
Task: Semantic Segmentation
Config: configs/upernet/upernet_r101_512x512_40k_voc12aug.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/upernet/upernet_r101_512x512_40k_voc12aug/upernet_r101_512x512_40k_voc12aug_20200613_163549-e26476ac.pth

9
configs/vit/README.md
View File

@ -4,6 +4,13 @@
<!-- [ALGORITHM] -->
<a href="https://github.com/google-research/vision_transformer">Official Repo</a>
<a href="https://github.com/open-mmlab/mmsegmentation/blob/v0.17.0/mmseg/models/backbones/vit.py#L98">Code Snippet</a>
<details>
<summary align="right"><a href="https://arxiv.org/pdf/2010.11929.pdf">Vision Transformer (ICLR'2021)</a></summary>
```latex
@article{dosoViTskiy2020,
title={An Image is Worth 16x16 Words: Transformers for Image Recognition at Scale},
@ -13,6 +20,8 @@
}
```
</details>
## Usage
To use other repositories' pre-trained models, it is necessary to convert keys.

167
configs/vit/vit.yml
View File

@ -1,248 +1,257 @@
Collections:
- Metadata:
- Name: vit
Metadata:
Training Data:
- ADE20K
Name: vit
Paper:
URL: https://arxiv.org/pdf/2010.11929.pdf
Title: Vision Transformer
README: configs/vit/README.md
Code:
URL: https://github.com/open-mmlab/mmsegmentation/blob/v0.17.0/mmseg/models/backbones/vit.py#L98
Version: v0.17.0
Converted From:
Code: https://github.com/google-research/vision_transformer
Models:
- Config: configs/vit/upernet_vit-b16_mln_512x512_80k_ade20k.py
- Name: upernet_vit-b16_mln_512x512_80k_ade20k
In Collection: vit
Metadata:
backbone: ViT-B + MLN
crop size: (512,512)
lr schd: 80000
inference time (ms/im):
- backend: PyTorch
batch size: 1
- value: 144.09
hardware: V100
backend: PyTorch
batch size: 1
mode: FP32
resolution: (512,512)
value: 144.09
lr schd: 80000
memory (GB): 9.2
Name: upernet_vit-b16_mln_512x512_80k_ade20k
Results:
- Task: Semantic Segmentation
Dataset: ADE20K
Metrics:
mIoU: 47.71
mIoU(ms+flip): 49.51
Task: Semantic Segmentation
Config: configs/vit/upernet_vit-b16_mln_512x512_80k_ade20k.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/vit/upernet_vit-b16_mln_512x512_80k_ade20k/upernet_vit-b16_mln_512x512_80k_ade20k_20210624_130547-0403cee1.pth
- Config: configs/vit/upernet_vit-b16_mln_512x512_160k_ade20k.py
- Name: upernet_vit-b16_mln_512x512_160k_ade20k
In Collection: vit
Metadata:
backbone: ViT-B + MLN
crop size: (512,512)
lr schd: 160000
inference time (ms/im):
- backend: PyTorch
batch size: 1
- value: 131.93
hardware: V100
backend: PyTorch
batch size: 1
mode: FP32
resolution: (512,512)
value: 131.93
lr schd: 160000
memory (GB): 9.2
Name: upernet_vit-b16_mln_512x512_160k_ade20k
Results:
- Task: Semantic Segmentation
Dataset: ADE20K
Metrics:
mIoU: 46.75
mIoU(ms+flip): 48.46
Task: Semantic Segmentation
Config: configs/vit/upernet_vit-b16_mln_512x512_160k_ade20k.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/vit/upernet_vit-b16_mln_512x512_160k_ade20k/upernet_vit-b16_mln_512x512_160k_ade20k_20210624_130547-852fa768.pth
- Config: configs/vit/upernet_vit-b16_ln_mln_512x512_160k_ade20k.py
- Name: upernet_vit-b16_ln_mln_512x512_160k_ade20k
In Collection: vit
Metadata:
backbone: ViT-B + LN + MLN
crop size: (512,512)
lr schd: 160000
inference time (ms/im):
- backend: PyTorch
batch size: 1
- value: 146.63
hardware: V100
backend: PyTorch
batch size: 1
mode: FP32
resolution: (512,512)
value: 146.63
lr schd: 160000
memory (GB): 9.21
Name: upernet_vit-b16_ln_mln_512x512_160k_ade20k
Results:
- Task: Semantic Segmentation
Dataset: ADE20K
Metrics:
mIoU: 47.73
mIoU(ms+flip): 49.95
Task: Semantic Segmentation
Config: configs/vit/upernet_vit-b16_ln_mln_512x512_160k_ade20k.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/vit/upernet_vit-b16_ln_mln_512x512_160k_ade20k/upernet_vit-b16_ln_mln_512x512_160k_ade20k_20210621_172828-f444c077.pth
- Config: configs/vit/upernet_deit-s16_512x512_80k_ade20k.py
- Name: upernet_deit-s16_512x512_80k_ade20k
In Collection: vit
Metadata:
backbone: DeiT-S
crop size: (512,512)
lr schd: 80000
inference time (ms/im):
- backend: PyTorch
batch size: 1
- value: 33.5
hardware: V100
backend: PyTorch
batch size: 1
mode: FP32
resolution: (512,512)
value: 33.5
lr schd: 80000
memory (GB): 4.68
Name: upernet_deit-s16_512x512_80k_ade20k
Results:
- Task: Semantic Segmentation
Dataset: ADE20K
Metrics:
mIoU: 42.96
mIoU(ms+flip): 43.79
Task: Semantic Segmentation
Config: configs/vit/upernet_deit-s16_512x512_80k_ade20k.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/vit/upernet_deit-s16_512x512_80k_ade20k/upernet_deit-s16_512x512_80k_ade20k_20210624_095228-afc93ec2.pth
- Config: configs/vit/upernet_deit-s16_512x512_160k_ade20k.py
- Name: upernet_deit-s16_512x512_160k_ade20k
In Collection: vit
Metadata:
backbone: DeiT-S
crop size: (512,512)
lr schd: 160000
inference time (ms/im):
- backend: PyTorch
batch size: 1
- value: 34.26
hardware: V100
backend: PyTorch
batch size: 1
mode: FP32
resolution: (512,512)
value: 34.26
lr schd: 160000
memory (GB): 4.68
Name: upernet_deit-s16_512x512_160k_ade20k
Results:
- Task: Semantic Segmentation
Dataset: ADE20K
Metrics:
mIoU: 42.87
mIoU(ms+flip): 43.79
Task: Semantic Segmentation
Config: configs/vit/upernet_deit-s16_512x512_160k_ade20k.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/vit/upernet_deit-s16_512x512_160k_ade20k/upernet_deit-s16_512x512_160k_ade20k_20210621_160903-5110d916.pth
- Config: configs/vit/upernet_deit-s16_mln_512x512_160k_ade20k.py
- Name: upernet_deit-s16_mln_512x512_160k_ade20k
In Collection: vit
Metadata:
backbone: DeiT-S + MLN
crop size: (512,512)
lr schd: 160000
inference time (ms/im):
- backend: PyTorch
batch size: 1
- value: 89.45
hardware: V100
backend: PyTorch
batch size: 1
mode: FP32
resolution: (512,512)
value: 89.45
lr schd: 160000
memory (GB): 5.69
Name: upernet_deit-s16_mln_512x512_160k_ade20k
Results:
- Task: Semantic Segmentation
Dataset: ADE20K
Metrics:
mIoU: 43.82
mIoU(ms+flip): 45.07
Task: Semantic Segmentation
Config: configs/vit/upernet_deit-s16_mln_512x512_160k_ade20k.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/vit/upernet_deit-s16_mln_512x512_160k_ade20k/upernet_deit-s16_mln_512x512_160k_ade20k_20210621_161021-fb9a5dfb.pth
- Config: configs/vit/upernet_deit-s16_ln_mln_512x512_160k_ade20k.py
- Name: upernet_deit-s16_ln_mln_512x512_160k_ade20k
In Collection: vit
Metadata:
backbone: DeiT-S + LN + MLN
crop size: (512,512)
lr schd: 160000
inference time (ms/im):
- backend: PyTorch
batch size: 1
- value: 80.71
hardware: V100
backend: PyTorch
batch size: 1
mode: FP32
resolution: (512,512)
value: 80.71
lr schd: 160000
memory (GB): 5.69
Name: upernet_deit-s16_ln_mln_512x512_160k_ade20k
Results:
- Task: Semantic Segmentation
Dataset: ADE20K
Metrics:
mIoU: 43.52
mIoU(ms+flip): 45.01
Task: Semantic Segmentation
Config: configs/vit/upernet_deit-s16_ln_mln_512x512_160k_ade20k.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/vit/upernet_deit-s16_ln_mln_512x512_160k_ade20k/upernet_deit-s16_ln_mln_512x512_160k_ade20k_20210621_161021-c0cd652f.pth
- Config: configs/vit/upernet_deit-b16_512x512_80k_ade20k.py
- Name: upernet_deit-b16_512x512_80k_ade20k
In Collection: vit
Metadata:
backbone: DeiT-B
crop size: (512,512)
lr schd: 80000
inference time (ms/im):
- backend: PyTorch
batch size: 1
- value: 103.2
hardware: V100
backend: PyTorch
batch size: 1
mode: FP32
resolution: (512,512)
value: 103.2
lr schd: 80000
memory (GB): 7.75
Name: upernet_deit-b16_512x512_80k_ade20k
Results:
- Task: Semantic Segmentation
Dataset: ADE20K
Metrics:
mIoU: 45.24
mIoU(ms+flip): 46.73
Task: Semantic Segmentation
Config: configs/vit/upernet_deit-b16_512x512_80k_ade20k.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/vit/upernet_deit-b16_512x512_80k_ade20k/upernet_deit-b16_512x512_80k_ade20k_20210624_130529-1e090789.pth
- Config: configs/vit/upernet_deit-b16_512x512_160k_ade20k.py
- Name: upernet_deit-b16_512x512_160k_ade20k
In Collection: vit
Metadata:
backbone: DeiT-B
crop size: (512,512)
lr schd: 160000
inference time (ms/im):
- backend: PyTorch
batch size: 1
- value: 96.25
hardware: V100
backend: PyTorch
batch size: 1
mode: FP32
resolution: (512,512)
value: 96.25
lr schd: 160000
memory (GB): 7.75
Name: upernet_deit-b16_512x512_160k_ade20k
Results:
- Task: Semantic Segmentation
Dataset: ADE20K
Metrics:
mIoU: 45.36
mIoU(ms+flip): 47.16
Task: Semantic Segmentation
Config: configs/vit/upernet_deit-b16_512x512_160k_ade20k.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/vit/upernet_deit-b16_512x512_160k_ade20k/upernet_deit-b16_512x512_160k_ade20k_20210621_180100-828705d7.pth
- Config: configs/vit/upernet_deit-b16_mln_512x512_160k_ade20k.py
- Name: upernet_deit-b16_mln_512x512_160k_ade20k
In Collection: vit
Metadata:
backbone: DeiT-B + MLN
crop size: (512,512)
lr schd: 160000
inference time (ms/im):
- backend: PyTorch
batch size: 1
- value: 128.53
hardware: V100
backend: PyTorch
batch size: 1
mode: FP32
resolution: (512,512)
value: 128.53
lr schd: 160000
memory (GB): 9.21
Name: upernet_deit-b16_mln_512x512_160k_ade20k
Results:
- Task: Semantic Segmentation
Dataset: ADE20K
Metrics:
mIoU: 45.46
mIoU(ms+flip): 47.16
Task: Semantic Segmentation
Config: configs/vit/upernet_deit-b16_mln_512x512_160k_ade20k.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/vit/upernet_deit-b16_mln_512x512_160k_ade20k/upernet_deit-b16_mln_512x512_160k_ade20k_20210621_191949-4e1450f3.pth
- Config: configs/vit/upernet_deit-b16_ln_mln_512x512_160k_ade20k.py
- Name: upernet_deit-b16_ln_mln_512x512_160k_ade20k
In Collection: vit
Metadata:
backbone: DeiT-B + LN + MLN
crop size: (512,512)
lr schd: 160000
inference time (ms/im):
- backend: PyTorch
batch size: 1
- value: 129.03
hardware: V100
backend: PyTorch
batch size: 1
mode: FP32
resolution: (512,512)
value: 129.03
lr schd: 160000
memory (GB): 9.21
Name: upernet_deit-b16_ln_mln_512x512_160k_ade20k
Results:
- Task: Semantic Segmentation
Dataset: ADE20K
Metrics:
mIoU: 45.37
mIoU(ms+flip): 47.23
Task: Semantic Segmentation
Config: configs/vit/upernet_deit-b16_ln_mln_512x512_160k_ade20k.py
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/vit/upernet_deit-b16_ln_mln_512x512_160k_ade20k/upernet_deit-b16_ln_mln_512x512_160k_ade20k_20210623_153535-8a959c14.pth