Data

Tools

Indexes

Applications

Experiments

Open Source Science

https://github.com/cviu-csu/maskcc

Mask-Consistent Contrastive Learning

https://github.com/cviu-csu/maskcc

Science Score: 13.0%

This score indicates how likely this project is to be science-related based on various indicators:

  • ○
    CITATION.cff file
  • ✓
    codemeta.json file
    Found codemeta.json file
  • ○
    .zenodo.json file
  • ○
    DOI references
  • ○
    Academic publication links
  • ○
    Academic email domains
  • ○
    Institutional organization owner
  • ○
    JOSS paper metadata
  • ○
    Scientific vocabulary similarity
    Low similarity (6.6%) to scientific vocabulary
Last synced: 9 months ago · JSON representation

Repository

Mask-Consistent Contrastive Learning

Basic Info
  • Host: GitHub
  • Owner: CVIU-CSU
  • Language: Python
  • Default Branch: main
  • Size: 1.91 MB
Statistics
  • Stars: 0
  • Watchers: 3
  • Forks: 0
  • Open Issues: 0
  • Releases: 0
Created about 2 years ago · Last pushed about 2 years ago
Metadata Files
Readme

README.md

Mask-Consistent Contrastive Learning

Introduction

The repository contains official Pytorch implementations of training and evaluation codes for Mask-Consistent Contrastive Learning.

Installation

  1. Follow tutorial to install pytorch1.11.0(or newer version) and cudatookit.
  2. Install mmcv, mmengine, mmcls and mmdet using MIM. shell pip install -U openmim mim install mmengine mim install "mmcv==2.0.0rc3" mim install "mmcls==1.0.0rc4" mim install "mmdet==3.0.0rc4"
  3. Install mmsegv1.0.0rc2 from source. shell git clone https://github.com/CVIU-CSU/MaskCC.git cd MaskCC pip install -v -e .
  4. Prepare datasets following tutorial. ## Training Cityscapes shell CUDA_VISIBLE_DEVICES=0,1,2,3,4,5,6,7 PORT=29513 bash tools/dist_train.sh configs/maskcc/maskcc_r50_8xb2-90k_cityscapes-512x1024.py 8 ADE20K: shell CUDA_VISIBLE_DEVICES=0,1,2,3 PORT=29513 bash tools/dist_train.sh configs/maskcc/maskcc_r50_4xb4-160k_ade20k-512x512.py 4 PASCAL-Context shell CUDA_VISIBLE_DEVICES=0,1,2,3 PORT=29513 bash tools/dist_train.sh configs/maskcc/maskcc_r50_4xb4-60k_pascal-context_480x480.py 4 COCO-Stuff10k shell CUDA_VISIBLE_DEVICES=0,1,2,3 PORT=29513 bash tools/dist_train.sh configs/maskcc/maskcc_r50_4xb4-60k_cocostuff_480x480.py 4 ## Evalution For single-scale test: shell CUDA_VISIBLE_DEVICES=0,1,2,3,4,5,6,7 PORT=29513 bash tools/dist_test.sh configs/maskcc/maskcc_r50_8xb2-90k_cityscapes-512x1024.py 8 For multi-scale test: The multi-scale test for mask2former is not supported in mmsegv1.0.0rc2.

Results

| Method | Backbone | Train Set | Eval Set | Batch | Iters | mIoU | Config | | :-------------: | :--------: | :-------: | :------: | :---: | :---: | :---: | ------------------------------------------------------------ | | Mask2Former | ResNet50 | Cityscapes train | Cityscapes val | 8x2 | 90K | 79.4 | config | | Mask2Former+MaskCC | ResNet50 | Cityscapes train | Cityscapes val | 8x2 | 90K | 80.9 | config | | Mask2Former | ResNet50 | ADE20K train | ADE20K val | 4x4 | 160K | 47.2 | config | | Mask2Former+MaskCC | ResNet50 | ADE20K train | ADE20K val | 4x4 | 160K | 48.4 | config | | Mask2Former | ResNet50 | PASCAL-Context train | PASCAL-Context val | 4x4 | 60K | 54.8 | config | | Mask2Former+MaskCC | ResNet50 | PASCAL-Context train | PASCAL-Context val | 4x4 | 60K | 55.3 | config | | Mask2Former | ResNet50 | COCO-Stuff10k train | COCO-Stuff10k val | 4x4 | 60K | 40.0 | config | | Mask2Former+MaskCC | ResNet50 | COCO-Stuff10k train | COCO-Stuff10k val | 4x4 | 60K | 41.2 | config |

Todo

  • [x] MaskCC code
  • [x] Configs
  • [x] Detailed readme
  • [ ] Citation

Owner

  • Name: CVIU-CSU
  • Login: CVIU-CSU
  • Kind: organization

GitHub Events

Total
Last Year

Dependencies

requirements/docs.txt pypi
  • docutils ==0.16.0
  • myst-parser *
  • sphinx ==4.0.2
  • sphinx_copybutton *
  • sphinx_markdown_tables *
requirements/mminstall.txt pypi
  • mmcls >=1.0.0rc0
  • mmcv >=2.0.0rc3,<2.1.0
  • mmdet >=3.0.0rc4
  • mmengine >=0.1.0,<1.0.0
requirements/optional.txt pypi
  • cityscapesscripts *
  • nibabel *
requirements/readthedocs.txt pypi
  • mmcv >=2.0.0rc0
  • mmengine *
  • prettytable *
  • torch *
  • torchvision *
requirements/runtime.txt pypi
  • matplotlib *
  • numpy *
  • packaging *
  • prettytable *
  • scipy *
requirements/tests.txt pypi
  • codecov * test
  • flake8 * test
  • interrogate * test
  • pytest * test
  • xdoctest >=0.10.0 test
  • yapf * test
requirements.txt pypi
setup.py pypi