{"id":16865739,"url":"https://github.com/zwwwayne/densesiam","last_synced_at":"2025-07-25T17:33:53.529Z","repository":{"id":47351991,"uuid":"515954881","full_name":"ZwwWayne/DenseSiam","owner":"ZwwWayne","description":"[ECCV2022] Dense Siamese Network for Dense Unsupervised Learning","archived":false,"fork":false,"pushed_at":"2022-07-21T00:49:10.000Z","size":77,"stargazers_count":28,"open_issues_count":0,"forks_count":1,"subscribers_count":1,"default_branch":"master","last_synced_at":"2025-03-25T06:41:42.939Z","etag":null,"topics":["pytorch","representation-learning","unsupervised-learning","unsupervised-semantic-segmentation"],"latest_commit_sha":null,"homepage":"","language":"Python","has_issues":true,"has_wiki":null,"has_pages":null,"mirror_url":null,"source_name":null,"license":"apache-2.0","status":null,"scm":"git","pull_requests_enabled":true,"icon_url":"https://github.com/ZwwWayne.png","metadata":{"files":{"readme":"README.md","changelog":null,"contributing":null,"funding":null,"license":"LICENSE","code_of_conduct":null,"threat_model":null,"audit":null,"citation":null,"codeowners":null,"security":null,"support":null}},"created_at":"2022-07-20T11:31:30.000Z","updated_at":"2024-03-26T02:06:00.000Z","dependencies_parsed_at":"2022-08-25T09:01:42.587Z","dependency_job_id":null,"html_url":"https://github.com/ZwwWayne/DenseSiam","commit_stats":null,"previous_names":[],"tags_count":0,"template":false,"template_full_name":null,"repository_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/ZwwWayne%2FDenseSiam","tags_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/ZwwWayne%2FDenseSiam/tags","releases_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/ZwwWayne%2FDenseSiam/releases","manifests_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/ZwwWayne%2FDenseSiam/manifests","owner_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners/ZwwWayne","download_url":"https://codeload.github.com/ZwwWayne/DenseSiam/tar.gz/refs/heads/master","host":{"name":"GitHub","url":"https://github.com","kind":"github","repositories_count":248370620,"owners_count":21092848,"icon_url":"https://github.com/github.png","version":null,"created_at":"2022-05-30T11:31:42.601Z","updated_at":"2022-07-04T15:15:14.044Z","host_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub","repositories_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories","repository_names_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repository_names","owners_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners"}},"keywords":["pytorch","representation-learning","unsupervised-learning","unsupervised-semantic-segmentation"],"created_at":"2024-10-13T14:48:20.576Z","updated_at":"2025-04-11T09:50:36.660Z","avatar_url":"https://github.com/ZwwWayne.png","language":"Python","funding_links":[],"categories":[],"sub_categories":[],"readme":"# Dense Siamese Network for Dense Unsupervised Learning\n\n## Introduction\n\nThis is an official release of the paper **Dense Siamese Network for Dense Unsupervised Learning**.\n\n\u003e [**Dense Siamese Network for Dense Unsupervised Learning**](https://arxiv.org/abs/2203.11075),  \n\u003e Wenwei Zhang, Jiangmiao Pang, Kai Chen, Chen Change Loy  \n\u003e In: Proc. European Conference on Computer Vision (ECCV), 2022  \n\u003e [[arXiv](https://arxiv.org/abs/2203.11075)][[project page](https://www.mmlab-ntu.com/project/densesiam/index.html)][[Bibetex](https://github.com/ZwwWayne/DenseSiam#citation)]\n\n## Results\n\n### Semantic segmentation on curated COCO stuff-thing dataset\n\nThe results of DenseSiam and their corresponding configs on unsupervised semantic segmentation task are shown as below.\nWe also re-implemented PiCIE based on the [official code release](https://github.com/janghyuncho/PiCIE).\n\n| Backbone | Method | Lr Schd | mIoU | Config | Download |\n| :---: | :---: | :---: | :---: | :---: | :---: |\n| R-18  | PiCIE | 10e       | 14.4 |[config](configs/picie/picie_r18_fpn_10e_coco_curated.py) | [model]() \u0026#124;  [log]() |\n| R-18  | DenseSiam | 10e     | 16.4 |[config](configs/densesiam/densesiam_r18_fpn_aux_seg-rebalance_4x64_sgd-fix-10e_coco-curated.py) | [model]() \u0026#124;  [log]() |\n\n### Unsupervised representation learning\n\n| Backbone | Method | Lr Schd | COCO Mask mAP| Config | Pre-train Download |\n| :---: | :---: | :---: | :---: | :---: | :---: |\n| R-50  | DenseSiam | 1x        | 36.8 |[config](configs/) | [model]() \u0026#124;  [log]() |\n\n## Installation\n\nIt requires the following OpenMMLab packages:\n\n- MIM \u003e= 0.1.5\n- MMCV-full \u003e= v1.3.14\n- MMDetection\n- MMSegmentation\n- MMSelfSup\n\n```bash\npip install openmim mmdet mmsegmentation mmselfsup\nmim install mmcv-full\n```\n\n## Usage\n\n### Data preparation\n\n- Download the [training set](http://images.cocodataset.org/zips/train2017.zip) and the [validdation set](http://images.cocodataset.org/zips/val2017.zip) of COCO dataset as well as the [stuffthing map](http://calvin.inf.ed.ac.uk/wp-content/uploads/data/cocostuffdataset/stuffthingmaps_trainval2017.zip).\n- Unzip these data and place them as the following structure\n- The `curated` directory copies the data split for unsupervised segmentation from [PiCIE](https://github.com/janghyuncho/PiCIE).\n\n```text\ndata/\n├── curated\n│   ├── train2017\n│   │   ├── Coco164kFull_Stuff_Coarse_7.txt\n│   ├── val2017\n│   │   ├── Coco164kFull_Stuff_Coarse_7.txt\n├── coco\n│   ├── annotations\n│   │   ├── train2017\n│   │   │   ├── xxxxxxxxx.png\n│   │   ├── val2017\n│   │   │   ├── xxxxxxxxx.png\n│   ├── train2017\n│   │   ├── xxxxxxxxx.jpeg\n│   ├── val2017\n│   │   ├── xxxxxxxxx.jpeg\n```\n\n### Training and testing\n\nFor training and testing, you can directly use mim to train and test the model\n\n```bash\n# train instance/panoptic segmentation models\nsh ./tools/slurm_train.sh $PARTITION $JOBNAME $CONFIG $WORK_DIR\n\n# test semantic segmentation models\nsh ./tools/slurm_test.sh $PARTITION $JOBNAME $CONFIG $CHECKPOINT --eval mIoU\n```\n\n- PARTITION: the slurm partition you are using\n- WORK_DIR: the working directory to save configs, logs, and checkpoints\n- CONFIG: the config files under the directory `configs/`\n- JOBNAME: the name of the job that are necessary for slurm\n\n## Acknowledgement\n\nThis codebase is based on [MMCV](https://github.com/open-mmlab/mmcv) and it benefits a lot from [PiCIE](https://github.com/janghyuncho/PiCIE) [MMSelfSup](https://github.com/open-mmlab/mmselfsup), and [Detectron2](https://github.com/facebookresearch/detectron2).\n\n## License\n\nThis project is released under the [Apache 2.0 license](LICENSE).\n\n## Citation\n\n```bibtex\n@inproceedings{zhang2022densesiam,\nauthor = {Wenwei Zhang and Jiangmiao Pang and Kai Chen and Chen Change Loy},\ntitle = {Dense Siamese Network for Dense Unsupervised Learning},\nyear = {2022},\nbooktitle = {ECCV},\n}\n```\n","project_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Fzwwwayne%2Fdensesiam","html_url":"https://awesome.ecosyste.ms/projects/github.com%2Fzwwwayne%2Fdensesiam","lists_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Fzwwwayne%2Fdensesiam/lists"}