{"id":22058621,"url":"https://github.com/v2ai/selfsup","last_synced_at":"2025-05-12T19:14:23.374Z","repository":{"id":85711535,"uuid":"319226994","full_name":"V2AI/SelfSup","owner":"V2AI","description":"Collections of self-supervised methods, based on cvpods.","archived":false,"fork":false,"pushed_at":"2021-08-21T05:01:14.000Z","size":144,"stargazers_count":58,"open_issues_count":0,"forks_count":8,"subscribers_count":5,"default_branch":"master","last_synced_at":"2025-05-12T19:14:14.464Z","etag":null,"topics":["barlow-twins","byol","cifar-10","classification","contrastive-learning","cvpods","det-con","eqco","imagenet","moco-v2","point-contrast","scrl","self-supervised","selfsup","simclr","simo","simsiam","swav"],"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/V2AI.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,"governance":null,"roadmap":null,"authors":null,"dei":null,"publiccode":null,"codemeta":null}},"created_at":"2020-12-07T06:38:31.000Z","updated_at":"2025-01-14T07:24:43.000Z","dependencies_parsed_at":null,"dependency_job_id":"b14c42e9-5014-44e8-98a6-0738bb3977be","html_url":"https://github.com/V2AI/SelfSup","commit_stats":null,"previous_names":["v2ai/selfsup"],"tags_count":0,"template":false,"template_full_name":null,"repository_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/V2AI%2FSelfSup","tags_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/V2AI%2FSelfSup/tags","releases_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/V2AI%2FSelfSup/releases","manifests_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/V2AI%2FSelfSup/manifests","owner_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners/V2AI","download_url":"https://codeload.github.com/V2AI/SelfSup/tar.gz/refs/heads/master","host":{"name":"GitHub","url":"https://github.com","kind":"github","repositories_count":253805844,"owners_count":21967054,"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":["barlow-twins","byol","cifar-10","classification","contrastive-learning","cvpods","det-con","eqco","imagenet","moco-v2","point-contrast","scrl","self-supervised","selfsup","simclr","simo","simsiam","swav"],"created_at":"2024-11-30T17:20:10.093Z","updated_at":"2025-05-12T19:14:23.357Z","avatar_url":"https://github.com/V2AI.png","language":"Python","readme":"# SelfSup\n\nCollections of self-supervised methods (MoCo series, SimCLR, **SiMo**, BYOL, SimSiam, SwAV, PointContrast, etc.). \n\n\n## Get Started\n\n### Install cvpods following the instructions.\n\nInstall cvpods from https://github.com/Megvii-BaseDetection/cvpods.git .\n\n### Prepare Datasets\n\n```shell\ncd cvpods\nln -s /path/to/your/ImageNet datasets/imagenet\n```\n\n### Train your own models\n\n```\ncd /path/to/your/SelfSup/examples/simclr/simclr.res50.scratch.imagenet.224size.256bs.200e\n# pre-train\npods_train --num-gpus 8\n# convert to weights\npython convert.py simclr.res50.scratch.imagenet.224size.256bs.200e/log/model_final.pth weights.pkl\n# downstream evaluation\ncd /path/to/your/simclr.res50.scratch.imagenet.224size.256bs.200e.lin_cls\npods_train --num-gpus 8 MODEL.WEIGHTS /path/to/your/weights.pkl\n\n```\n\n## Model Zoo\n\n### Supervised Classification \n\n#### ImageNet\n| Methods | Training Schedule | Top 1  Acc |\n| ------- | ------ | ------------------ |\n| Res50   | 100e    | 76.4               |\n\n#### CIFAR 10\n| Methods | Training Schedule | Top 1  Acc |\n| ------- | ------ | ------------------ |\n| Res50   | 200e    | 95.4              |\n\n#### STL 10\n| Methods | Training Schedule | Top 1  Acc |\n| ------- | ------ | ------------------ |\n| Res50   | 150e    | 86.1              |\n\n\n### Self-Supervised Learning - Classification\n\n\u003e All results in the below table are trained using resnet-50 and reported on the ILSVRC2012 dataset.\n\n| Methods       | Training Schedule | Batch Size | Our Acc@1 | Official Acc@1 |\n| -------       | ------            | ---------- | --------- | -------------- |\n| MoCo          | 200e              |     256    | 60.5      | 60.5           |\n| MoCov2        | 200e              |     256    | **67.6**      | 67.5           |\n| SimCLR        | 200e              |     256    | **63.2**      | 61.9           |\n| **SimCLR***       | 200e              |     256    | **67.3**      | **Ours**           |\n| **SiMo**          | 200e              |     256    | **68.1**      | **Ours**           |\n| SimSiam       | 100e              |     256    | 67.6      | 67.7           |\n| SwAV          | 200e              |     256    | **73.0**      | 72.7           |\n| BYOL          | 100e              |     2048   | **69.8**      | 66.5 (bs4096 from SimSiam paper) |\n| BarlowTwins   | 300e              |     1024   | Comming Soon| 71.7         |\n\n### Self-Supervised Learning - Detection (2D)\n\n\u003e All the results reported below are trained on ILSVRC2012 and evaluated on MS COCO using Faster-RCNN-FPN and resnet-50.\n\n| Methods | Training Schedule | Batch Size | Box AP | \n| ------- | ------ | ---------- | ------------------ |\n| SCRL    | 200    |     4096    | 39.9 ( official: 40.5 with bs 8192)   | \n| DetCon    | 200    |     256    | Comming Soon.      |\n\n### Self-Supervised Learning - 3D Scene Understanding\n\n| Methods       | Training Schedule | Downstream task |\n| ------------- | ----- | --------------- |\n| PointContrast | -     | Comming Soon.   |\n\n\n## Citation\n\nSelfSup is a part of [cvpods](https://github.com/Megvii-BaseDetection/cvpods), so if you find this repo useful in your research, or if you want to refer the implementations in this repo, please consider cite:\n\n```BibTeX\n\n@article{zhu2020eqco,\n  title={EqCo: Equivalent Rules for Self-supervised Contrastive Learning},\n  author={Zhu, Benjin and Huang, Junqiang and Li, Zeming and Zhang, Xiangyu and Sun, Jian},\n  journal={arXiv preprint arXiv:2010.01929},\n  year={2020}\n}\n\n@misc{zhu2020cvpods,\n  title={cvpods: All-in-one Toolbox for Computer Vision Research},\n  author={Zhu*, Benjin and Wang*, Feng and Wang, Jianfeng and Yang, Siwei and Chen, Jianhu and Li, Zeming},\n  year={2020}\n}\n```\n","funding_links":[],"categories":[],"sub_categories":[],"project_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Fv2ai%2Fselfsup","html_url":"https://awesome.ecosyste.ms/projects/github.com%2Fv2ai%2Fselfsup","lists_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Fv2ai%2Fselfsup/lists"}