{"id":13455951,"url":"https://github.com/facebookresearch/dinov2","last_synced_at":"2025-05-14T04:07:36.821Z","repository":{"id":153376787,"uuid":"620911108","full_name":"facebookresearch/dinov2","owner":"facebookresearch","description":"PyTorch code and models for the DINOv2 self-supervised learning method.","archived":false,"fork":false,"pushed_at":"2024-08-07T13:44:50.000Z","size":1137,"stargazers_count":9441,"open_issues_count":250,"forks_count":845,"subscribers_count":95,"default_branch":"main","last_synced_at":"2024-12-17T01:37:21.051Z","etag":null,"topics":[],"latest_commit_sha":null,"homepage":null,"language":"Jupyter Notebook","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/facebookresearch.png","metadata":{"files":{"readme":"README.md","changelog":null,"contributing":"CONTRIBUTING.md","funding":null,"license":"LICENSE","code_of_conduct":"CODE_OF_CONDUCT.md","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":"2023-03-29T16:00:37.000Z","updated_at":"2024-12-16T19:18:43.000Z","dependencies_parsed_at":"2023-09-27T18:25:16.014Z","dependency_job_id":"d307c07c-5b1e-493d-9356-9065cd52767a","html_url":"https://github.com/facebookresearch/dinov2","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/facebookresearch%2Fdinov2","tags_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/facebookresearch%2Fdinov2/tags","releases_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/facebookresearch%2Fdinov2/releases","manifests_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/facebookresearch%2Fdinov2/manifests","owner_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners/facebookresearch","download_url":"https://codeload.github.com/facebookresearch/dinov2/tar.gz/refs/heads/main","host":{"name":"GitHub","url":"https://github.com","kind":"github","repositories_count":254069217,"owners_count":22009511,"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":[],"created_at":"2024-07-31T08:01:13.922Z","updated_at":"2025-05-14T04:07:36.761Z","avatar_url":"https://github.com/facebookresearch.png","language":"Jupyter Notebook","funding_links":[],"categories":["Jupyter Notebook","🧠 SOTA 2024-2025: Mobile LLMs \u0026 Multimodal","Paper List","Image \u0026 Vision","Resources","其他_机器视觉","Applications","largemodel","Repos","Contrastive \u0026 Self-Supervised Learning","Data \u0026 Analytics","Self-Supervision"],"sub_categories":["🎨 Vision Models \u0026 Feature Extraction","Seminal Papers","[TVCG24] NIS-SLAM: Neural Implicit Semantic RGB-D SLAM for 3D Consistent Scene Understanding","网络服务_其他","提示语（魔法）","Data Science \u0026 Visualization"],"readme":":new: [2023-10-26] *Added DINOv2 backbones with registers, following [Vision Transformers Need Registers](https://arxiv.org/abs/2309.16588).*\n\n# DINOv2: Learning Robust Visual Features without Supervision\n\n**[Meta AI Research, FAIR](https://ai.facebook.com/research/)**\n\nMaxime Oquab,\nTimothée Darcet,\nThéo Moutakanni,\nHuy V. Vo,\nMarc Szafraniec,\nVasil Khalidov,\nPatrick Labatut,\nArmand Joulin,\nPiotr Bojanowski\n\n[[`Paper #1`](https://arxiv.org/abs/2304.07193)] [`Paper #2`](https://arxiv.org/abs/2309.16588)] [[`Blog`](https://ai.facebook.com/blog/dino-v2-computer-vision-self-supervised-learning/)] [[`Demo`](https://dinov2.metademolab.com)] [[`BibTeX`](#citing-dinov2)]\n\nPyTorch implementation and pretrained models for DINOv2. For details, see the papers: **[DINOv2: Learning Robust Visual Features without Supervision](https://arxiv.org/abs/2304.07193)** and **[Vision Transformers Need Registers](https://arxiv.org/abs/2309.16588)**.\n\nDINOv2 models produce high-performance visual features that can be directly employed with classifiers as simple as linear layers on a variety of computer vision tasks; these visual features are robust and perform well across domains without any requirement for fine-tuning. The models were pretrained on a dataset of 142 M images without using any labels or annotations.\n\nhttps://github.com/facebookresearch/dinov2/assets/60359573/f168823e-7922-415a-b429-578badf5c356\n\n\u003cdiv align=\"center\"\u003e\n  Visualization of the three first principal components of the patch features of all frames, mapped to RGB values.\n\u003c/div\u003e\n\n## Pretrained models\n\n\u003ctable style=\"margin: auto\"\u003e\n  \u003cthead\u003e\n    \u003ctr\u003e\n      \u003cth\u003emodel\u003c/th\u003e\n      \u003cth\u003e# of\u003cbr /\u003eparams\u003c/th\u003e\n      \u003cth\u003ewith\u003cbr /\u003eregisters\u003c/th\u003e\n      \u003cth\u003eImageNet\u003cbr /\u003ek-NN\u003c/th\u003e\n      \u003cth\u003eImageNet\u003cbr /\u003elinear\u003c/th\u003e\n      \u003cth\u003edownload\u003c/th\u003e\n    \u003c/tr\u003e\n  \u003c/thead\u003e\n  \u003ctbody\u003e\n    \u003ctr\u003e\n      \u003ctd\u003eViT-S/14 distilled\u003c/td\u003e\n      \u003ctd align=\"right\"\u003e21 M\u003c/td\u003e\n      \u003ctd align=\"center\"\u003e:x:\u003c/td\u003e\n      \u003ctd align=\"right\"\u003e79.0%\u003c/td\u003e\n      \u003ctd align=\"right\"\u003e81.1%\u003c/td\u003e\n      \u003ctd\u003e\u003ca href=\"https://dl.fbaipublicfiles.com/dinov2/dinov2_vits14/dinov2_vits14_pretrain.pth\"\u003ebackbone only\u003c/a\u003e\u003c/td\u003e\n    \u003c/tr\u003e\n    \u003ctr\u003e\n      \u003ctd\u003eViT-S/14 distilled\u003c/td\u003e\n      \u003ctd align=\"right\"\u003e21 M\u003c/td\u003e\n      \u003ctd align=\"center\"\u003e:white_check_mark:\u003c/td\u003e\n      \u003ctd align=\"right\"\u003e79.1%\u003c/td\u003e\n      \u003ctd align=\"right\"\u003e80.9%\u003c/td\u003e\n      \u003ctd\u003e\u003ca href=\"https://dl.fbaipublicfiles.com/dinov2/dinov2_vits14/dinov2_vits14_reg4_pretrain.pth\"\u003ebackbone only\u003c/a\u003e\u003c/td\u003e\n    \u003c/tr\u003e\n    \u003ctr\u003e\n      \u003ctd\u003eViT-B/14 distilled\u003c/td\u003e\n      \u003ctd align=\"right\"\u003e86 M\u003c/td\u003e\n      \u003ctd align=\"center\"\u003e:x:\u003c/td\u003e\n      \u003ctd align=\"right\"\u003e82.1%\u003c/td\u003e\n      \u003ctd align=\"right\"\u003e84.5%\u003c/td\u003e\n      \u003ctd\u003e\u003ca href=\"https://dl.fbaipublicfiles.com/dinov2/dinov2_vitb14/dinov2_vitb14_pretrain.pth\"\u003ebackbone only\u003c/a\u003e\u003c/td\u003e\n    \u003c/tr\u003e\n    \u003ctr\u003e\n      \u003ctd\u003eViT-B/14 distilled\u003c/td\u003e\n      \u003ctd align=\"right\"\u003e86 M\u003c/td\u003e\n      \u003ctd align=\"center\"\u003e:white_check_mark:\u003c/td\u003e\n      \u003ctd align=\"right\"\u003e82.0%\u003c/td\u003e\n      \u003ctd align=\"right\"\u003e84.6%\u003c/td\u003e\n      \u003ctd\u003e\u003ca href=\"https://dl.fbaipublicfiles.com/dinov2/dinov2_vitb14/dinov2_vitb14_reg4_pretrain.pth\"\u003ebackbone only\u003c/a\u003e\u003c/td\u003e\n    \u003c/tr\u003e\n    \u003ctr\u003e\n      \u003ctd\u003eViT-L/14 distilled\u003c/td\u003e\n      \u003ctd align=\"right\"\u003e300 M\u003c/td\u003e\n      \u003ctd align=\"center\"\u003e:x:\u003c/td\u003e\n      \u003ctd align=\"right\"\u003e83.5%\u003c/td\u003e\n      \u003ctd align=\"right\"\u003e86.3%\u003c/td\u003e\n      \u003ctd\u003e\u003ca href=\"https://dl.fbaipublicfiles.com/dinov2/dinov2_vitl14/dinov2_vitl14_pretrain.pth\"\u003ebackbone only\u003c/a\u003e\u003c/td\u003e\n    \u003c/tr\u003e\n    \u003ctr\u003e\n      \u003ctd\u003eViT-L/14 distilled\u003c/td\u003e\n      \u003ctd align=\"right\"\u003e300 M\u003c/td\u003e\n      \u003ctd align=\"center\"\u003e:white_check_mark:\u003c/td\u003e\n      \u003ctd align=\"right\"\u003e83.8%\u003c/td\u003e\n      \u003ctd align=\"right\"\u003e86.7%\u003c/td\u003e\n      \u003ctd\u003e\u003ca href=\"https://dl.fbaipublicfiles.com/dinov2/dinov2_vitl14/dinov2_vitl14_reg4_pretrain.pth\"\u003ebackbone only\u003c/a\u003e\u003c/td\u003e\n    \u003c/tr\u003e\n    \u003ctr\u003e\n      \u003ctd\u003eViT-g/14\u003c/td\u003e\n      \u003ctd align=\"right\"\u003e1,100 M\u003c/td\u003e\n      \u003ctd align=\"center\"\u003e:x:\u003c/td\u003e\n      \u003ctd align=\"right\"\u003e83.5%\u003c/td\u003e\n      \u003ctd align=\"right\"\u003e86.5%\u003c/td\u003e\n      \u003ctd\u003e\u003ca href=\"https://dl.fbaipublicfiles.com/dinov2/dinov2_vitg14/dinov2_vitg14_pretrain.pth\"\u003ebackbone only\u003c/a\u003e\u003c/td\u003e\n    \u003c/tr\u003e\n    \u003ctr\u003e\n      \u003ctd\u003eViT-g/14\u003c/td\u003e\n      \u003ctd align=\"right\"\u003e1,100 M\u003c/td\u003e\n      \u003ctd align=\"center\"\u003e:white_check_mark:\u003c/td\u003e\n      \u003ctd align=\"right\"\u003e83.7%\u003c/td\u003e\n      \u003ctd align=\"right\"\u003e87.1%\u003c/td\u003e\n      \u003ctd\u003e\u003ca href=\"https://dl.fbaipublicfiles.com/dinov2/dinov2_vitg14/dinov2_vitg14_reg4_pretrain.pth\"\u003ebackbone only\u003c/a\u003e\u003c/td\u003e\n    \u003c/tr\u003e\n  \u003c/tbody\u003e\n\u003c/table\u003e\n\n### Pretrained backbones (via PyTorch Hub)\n\nPlease follow the instructions [here](https://pytorch.org/get-started/locally/) to install PyTorch (the only required dependency for loading the model). Installing PyTorch with CUDA support is strongly recommended.\n\nA corresponding [model card](MODEL_CARD.md) is included in the repository.\n\n```python\nimport torch\n\n# DINOv2\ndinov2_vits14 = torch.hub.load('facebookresearch/dinov2', 'dinov2_vits14')\ndinov2_vitb14 = torch.hub.load('facebookresearch/dinov2', 'dinov2_vitb14')\ndinov2_vitl14 = torch.hub.load('facebookresearch/dinov2', 'dinov2_vitl14')\ndinov2_vitg14 = torch.hub.load('facebookresearch/dinov2', 'dinov2_vitg14')\n\n# DINOv2 with registers\ndinov2_vits14_reg = torch.hub.load('facebookresearch/dinov2', 'dinov2_vits14_reg')\ndinov2_vitb14_reg = torch.hub.load('facebookresearch/dinov2', 'dinov2_vitb14_reg')\ndinov2_vitl14_reg = torch.hub.load('facebookresearch/dinov2', 'dinov2_vitl14_reg')\ndinov2_vitg14_reg = torch.hub.load('facebookresearch/dinov2', 'dinov2_vitg14_reg')\n```\n\n### Pretrained heads - Image classification\n\n\u003ctable style=\"margin: auto\"\u003e\n  \u003cthead\u003e\n    \u003ctr\u003e\n      \u003cth rowspan=\"2\"\u003ebackbone\u003c/th\u003e\n      \u003cth rowspan=\"2\"\u003ewith\u003cbr /\u003eregisters\u003c/th\u003e\n      \u003cth\u003edownload\u003c/th\u003e\n    \u003c/tr\u003e\n    \u003ctr\u003e\n      \u003cth\u003eImageNet\u003c/th\u003e\n    \u003c/tr\u003e\n  \u003c/thead\u003e\n  \u003ctbody\u003e\n    \u003ctr\u003e\n      \u003ctd\u003eViT-S/14 distilled\u003c/td\u003e\n      \u003ctd align=\"center\"\u003e:x:\u003c/td\u003e\n      \u003ctd\u003e\n        linear head (\u003ca href=\"https://dl.fbaipublicfiles.com/dinov2/dinov2_vits14/dinov2_vits14_linear_head.pth\"\u003e1 layer\u003c/a\u003e,\n        \u003ca href=\"https://dl.fbaipublicfiles.com/dinov2/dinov2_vits14/dinov2_vits14_linear4_head.pth\"\u003e4 layers\u003c/a\u003e)\n      \u003c/td\u003e\n    \u003c/tr\u003e\n    \u003ctr\u003e\n      \u003ctd\u003eViT-S/14 distilled\u003c/td\u003e\n      \u003ctd align=\"center\"\u003e:white_check_mark:\u003c/td\u003e\n      \u003ctd\u003e\n        linear head (\u003ca href=\"https://dl.fbaipublicfiles.com/dinov2/dinov2_vits14/dinov2_vits14_reg4_linear_head.pth\"\u003e1 layer\u003c/a\u003e,\n        \u003ca href=\"https://dl.fbaipublicfiles.com/dinov2/dinov2_vits14/dinov2_vits14_reg4_linear4_head.pth\"\u003e4 layers\u003c/a\u003e)\n      \u003c/td\u003e\n    \u003c/tr\u003e\n    \u003ctr\u003e\n      \u003ctd\u003eViT-B/14 distilled\u003c/td\u003e\n      \u003ctd align=\"center\"\u003e:x:\u003c/td\u003e\n      \u003ctd\u003e\n        linear head (\u003ca href=\"https://dl.fbaipublicfiles.com/dinov2/dinov2_vitb14/dinov2_vitb14_linear_head.pth\"\u003e1 layer\u003c/a\u003e,\n        \u003ca href=\"https://dl.fbaipublicfiles.com/dinov2/dinov2_vitb14/dinov2_vitb14_linear4_head.pth\"\u003e4 layers\u003c/a\u003e)\n    \u003c/tr\u003e\n    \u003ctr\u003e\n      \u003ctd\u003eViT-B/14 distilled\u003c/td\u003e\n      \u003ctd align=\"center\"\u003e:white_check_mark:\u003c/td\u003e\n      \u003ctd\u003e\n        linear head (\u003ca href=\"https://dl.fbaipublicfiles.com/dinov2/dinov2_vitb14/dinov2_vitb14_reg4_linear_head.pth\"\u003e1 layer\u003c/a\u003e,\n        \u003ca href=\"https://dl.fbaipublicfiles.com/dinov2/dinov2_vitb14/dinov2_vitb14_reg4_linear4_head.pth\"\u003e4 layers\u003c/a\u003e)\n    \u003c/tr\u003e\n    \u003ctr\u003e\n      \u003ctd\u003eViT-L/14 distilled\u003c/td\u003e\n      \u003ctd align=\"center\"\u003e:x:\u003c/td\u003e\n      \u003ctd\u003e\n        linear head (\u003ca href=\"https://dl.fbaipublicfiles.com/dinov2/dinov2_vitl14/dinov2_vitl14_linear_head.pth\"\u003e1 layer\u003c/a\u003e,\n        \u003ca href=\"https://dl.fbaipublicfiles.com/dinov2/dinov2_vitl14/dinov2_vitl14_linear4_head.pth\"\u003e4 layers\u003c/a\u003e)\n    \u003c/tr\u003e\n    \u003ctr\u003e\n      \u003ctd\u003eViT-L/14 distilled\u003c/td\u003e\n      \u003ctd align=\"center\"\u003e:white_check_mark:\u003c/td\u003e\n      \u003ctd\u003e\n        linear head (\u003ca href=\"https://dl.fbaipublicfiles.com/dinov2/dinov2_vitl14/dinov2_vitl14_reg4_linear_head.pth\"\u003e1 layer\u003c/a\u003e,\n        \u003ca href=\"https://dl.fbaipublicfiles.com/dinov2/dinov2_vitl14/dinov2_vitl14_reg4_linear4_head.pth\"\u003e4 layers\u003c/a\u003e)\n    \u003c/tr\u003e\n    \u003ctr\u003e\n      \u003ctd\u003eViT-g/14\u003c/td\u003e\n      \u003ctd align=\"center\"\u003e:x:\u003c/td\u003e\n      \u003ctd\u003e\n        linear head (\u003ca href=\"https://dl.fbaipublicfiles.com/dinov2/dinov2_vitg14/dinov2_vitg14_linear_head.pth\"\u003e1 layer\u003c/a\u003e,\n        \u003ca href=\"https://dl.fbaipublicfiles.com/dinov2/dinov2_vitg14/dinov2_vitg14_linear4_head.pth\"\u003e4 layers\u003c/a\u003e)\n    \u003c/tr\u003e\n    \u003ctr\u003e\n      \u003ctd\u003eViT-g/14\u003c/td\u003e\n      \u003ctd align=\"center\"\u003e:white_check_mark:\u003c/td\u003e\n      \u003ctd\u003e\n        linear head (\u003ca href=\"https://dl.fbaipublicfiles.com/dinov2/dinov2_vitg14/dinov2_vitg14_lreg4_inear_head.pth\"\u003e1 layer\u003c/a\u003e,\n        \u003ca href=\"https://dl.fbaipublicfiles.com/dinov2/dinov2_vitg14/dinov2_vitg14_reg4_linear4_head.pth\"\u003e4 layers\u003c/a\u003e)\n    \u003c/tr\u003e\n  \u003c/tbody\u003e\n\u003c/table\u003e\n\nThe (full) classifier models can be loaded via PyTorch Hub:\n\n```python\nimport torch\n\n# DINOv2\ndinov2_vits14_lc = torch.hub.load('facebookresearch/dinov2', 'dinov2_vits14_lc')\ndinov2_vitb14_lc = torch.hub.load('facebookresearch/dinov2', 'dinov2_vitb14_lc')\ndinov2_vitl14_lc = torch.hub.load('facebookresearch/dinov2', 'dinov2_vitl14_lc')\ndinov2_vitg14_lc = torch.hub.load('facebookresearch/dinov2', 'dinov2_vitg14_lc')\n\n# DINOv2 with registers\ndinov2_vits14_reg_lc = torch.hub.load('facebookresearch/dinov2', 'dinov2_vits14_reg_lc')\ndinov2_vitb14_reg_lc = torch.hub.load('facebookresearch/dinov2', 'dinov2_vitb14_reg_lc')\ndinov2_vitl14_reg_lc = torch.hub.load('facebookresearch/dinov2', 'dinov2_vitl14_reg_lc')\ndinov2_vitg14_reg_lc = torch.hub.load('facebookresearch/dinov2', 'dinov2_vitg14_reg_lc')\n```\n\n### Pretrained heads - Depth estimation\n\n\u003ctable style=\"margin: auto\"\u003e\n  \u003cthead\u003e\n    \u003ctr\u003e\n      \u003cth rowspan=\"2\"\u003ebackbone\u003c/th\u003e\n      \u003cth colspan=\"2\"\u003edownload head\u003c/th\u003e\n    \u003c/tr\u003e\n    \u003ctr\u003e\n      \u003cth\u003eNYUd\u003c/th\u003e\n      \u003cth\u003eKITTI\u003c/th\u003e\n    \u003c/tr\u003e\n  \u003c/thead\u003e\n  \u003ctbody\u003e\n    \u003ctr\u003e\n      \u003ctd\u003eViT-S/14 distilled\u003c/td\u003e\n      \u003ctd\u003e\n        linear (\u003ca href=\"https://dl.fbaipublicfiles.com/dinov2/dinov2_vits14/dinov2_vits14_nyu_linear_head.pth\"\u003e1 layer\u003c/a\u003e,\n        \u003ca href=\"https://dl.fbaipublicfiles.com/dinov2/dinov2_vits14/dinov2_vits14_nyu_linear4_head.pth\"\u003e4 layers\u003c/a\u003e),\n        \u003ca href=\"https://dl.fbaipublicfiles.com/dinov2/dinov2_vits14/dinov2_vits14_nyu_dpt_head.pth\"\u003eDPT\u003c/a\u003e\n      \u003c/td\u003e\n      \u003ctd\u003e\n        linear (\u003ca href=\"https://dl.fbaipublicfiles.com/dinov2/dinov2_vits14/dinov2_vits14_kitti_linear_head.pth\"\u003e1 layer\u003c/a\u003e,\n        \u003ca href=\"https://dl.fbaipublicfiles.com/dinov2/dinov2_vits14/dinov2_vits14_kitti_linear4_head.pth\"\u003e4 layers\u003c/a\u003e),\n        \u003ca href=\"https://dl.fbaipublicfiles.com/dinov2/dinov2_vits14/dinov2_vits14_kitti_dpt_head.pth\"\u003eDPT\u003c/a\u003e\n      \u003c/td\u003e\n    \u003c/tr\u003e\n    \u003ctr\u003e\n      \u003ctd\u003eViT-B/14 distilled\u003c/td\u003e\n      \u003ctd\u003e\n        linear (\u003ca href=\"https://dl.fbaipublicfiles.com/dinov2/dinov2_vitb14/dinov2_vitb14_linear_head.pth\"\u003e1 layer\u003c/a\u003e,\n        \u003ca href=\"https://dl.fbaipublicfiles.com/dinov2/dinov2_vitb14/dinov2_vitb14_nyu_linear4_head.pth\"\u003e4 layers\u003c/a\u003e),\n        \u003ca href=\"https://dl.fbaipublicfiles.com/dinov2/dinov2_vitb14/dinov2_vitb14_nyu_dpt_head.pth\"\u003eDPT\u003c/a\u003e\n      \u003c/td\u003e\n      \u003ctd\u003e\n        linear (\u003ca href=\"https://dl.fbaipublicfiles.com/dinov2/dinov2_vitb14/dinov2_vitb14_kitti_linear_head.pth\"\u003e1 layer\u003c/a\u003e,\n        \u003ca href=\"https://dl.fbaipublicfiles.com/dinov2/dinov2_vitb14/dinov2_vitb14_kitti_linear4_head.pth\"\u003e4 layers\u003c/a\u003e),\n        \u003ca href=\"https://dl.fbaipublicfiles.com/dinov2/dinov2_vitb14/dinov2_vitb14_kitti_dpt_head.pth\"\u003eDPT\u003c/a\u003e\n      \u003c/td\u003e\n    \u003c/tr\u003e\n    \u003ctr\u003e\n      \u003ctd\u003eViT-L/14 distilled\u003c/td\u003e\n      \u003ctd\u003e\n        linear (\u003ca href=\"https://dl.fbaipublicfiles.com/dinov2/dinov2_vitl14/dinov2_vitl14_linear_head.pth\"\u003e1 layer\u003c/a\u003e,\n        \u003ca href=\"https://dl.fbaipublicfiles.com/dinov2/dinov2_vitl14/dinov2_vitl14_nyu_linear4_head.pth\"\u003e4 layers\u003c/a\u003e),\n        \u003ca href=\"https://dl.fbaipublicfiles.com/dinov2/dinov2_vitl14/dinov2_vitl14_nyu_dpt_head.pth\"\u003eDPT\u003c/a\u003e\n      \u003c/td\u003e\n      \u003ctd\u003e\n        linear (\u003ca href=\"https://dl.fbaipublicfiles.com/dinov2/dinov2_vitl14/dinov2_vitl14_kitti_linear_head.pth\"\u003e1 layer\u003c/a\u003e,\n        \u003ca href=\"https://dl.fbaipublicfiles.com/dinov2/dinov2_vitl14/dinov2_vitl14_kitti_linear4_head.pth\"\u003e4 layers\u003c/a\u003e),\n        \u003ca href=\"https://dl.fbaipublicfiles.com/dinov2/dinov2_vitl14/dinov2_vitl14_kitti_dpt_head.pth\"\u003eDPT\u003c/a\u003e\n      \u003c/td\u003e\n    \u003c/tr\u003e\n    \u003ctr\u003e\n      \u003ctd\u003eViT-g/14\u003c/td\u003e\n      \u003ctd\u003e\n        linear (\u003ca href=\"https://dl.fbaipublicfiles.com/dinov2/dinov2_vitg14/dinov2_vitg14_linear_head.pth\"\u003e1 layer\u003c/a\u003e,\n        \u003ca href=\"https://dl.fbaipublicfiles.com/dinov2/dinov2_vitg14/dinov2_vitg14_nyu_linear4_head.pth\"\u003e4 layers\u003c/a\u003e),\n        \u003ca href=\"https://dl.fbaipublicfiles.com/dinov2/dinov2_vitg14/dinov2_vitg14_nyu_dpt_head.pth\"\u003eDPT\u003c/a\u003e\n      \u003c/td\u003e\n      \u003ctd\u003e\n        linear (\u003ca href=\"https://dl.fbaipublicfiles.com/dinov2/dinov2_vitg14/dinov2_vitg14_kitti_linear_head.pth\"\u003e1 layer\u003c/a\u003e,\n        \u003ca href=\"https://dl.fbaipublicfiles.com/dinov2/dinov2_vitg14/dinov2_vitg14_kitti_linear4_head.pth\"\u003e4 layers\u003c/a\u003e),\n        \u003ca href=\"https://dl.fbaipublicfiles.com/dinov2/dinov2_vitg14/dinov2_vitg14_kitti_dpt_head.pth\"\u003eDPT\u003c/a\u003e\n      \u003c/td\u003e\n    \u003c/tr\u003e\n  \u003c/tbody\u003e\n\u003c/table\u003e\n\n### Pretrained heads - Semantic segmentation\n\n\u003ctable style=\"margin: auto\"\u003e\n  \u003cthead\u003e\n    \u003ctr\u003e\n      \u003cth rowspan=\"2\"\u003ebackbone\u003c/th\u003e\n      \u003cth\u003edownload model\u003c/th\u003e\n      \u003cth colspan=\"2\"\u003edownload head\u003c/th\u003e\n    \u003c/tr\u003e\n    \u003ctr\u003e\n      \u003cth\u003eADE20K\u003c/th\u003e\n      \u003cth\u003eADE20K\u003c/th\u003e\n      \u003cth\u003eVOC2012\u003c/th\u003e\n    \u003c/tr\u003e\n  \u003c/thead\u003e\n  \u003ctbody\u003e\n    \u003ctr\u003e\n      \u003ctd\u003eViT-S/14 distilled\u003c/td\u003e\n      \u003ctd\u003e\u003c/td\u003e\n      \u003ctd\u003e\n        \u003ca href=\"https://dl.fbaipublicfiles.com/dinov2/dinov2_vits14/dinov2_vits14_ade20k_linear_head.pth\"\u003elinear\u003c/a\u003e,\n        \u003ca href=\"https://dl.fbaipublicfiles.com/dinov2/dinov2_vits14/dinov2_vits14_ade20k_ms_head.pth\"\u003emulti-scale\u003c/a\u003e\n      \u003c/td\u003e\n      \u003ctd\u003e\n        \u003ca href=\"https://dl.fbaipublicfiles.com/dinov2/dinov2_vits14/dinov2_vits14_voc2012_linear_head.pth\"\u003elinear\u003c/a\u003e,\n        \u003ca href=\"https://dl.fbaipublicfiles.com/dinov2/dinov2_vits14/dinov2_vits14_voc2012_ms_head.pth\"\u003emulti-scale\u003c/a\u003e\n      \u003c/td\u003e\n    \u003c/tr\u003e\n    \u003ctr\u003e\n      \u003ctd\u003eViT-B/14 distilled\u003c/td\u003e\n      \u003ctd\u003e\u003c/td\u003e\n      \u003ctd\u003e\n        \u003ca href=\"https://dl.fbaipublicfiles.com/dinov2/dinov2_vitb14/dinov2_vitb14_ade20k_linear_head.pth\"\u003elinear\u003c/a\u003e,\n        \u003ca href=\"https://dl.fbaipublicfiles.com/dinov2/dinov2_vitb14/dinov2_vitb14_ade20k_ms_head.pth\"\u003emulti-scale\u003c/a\u003e\n      \u003c/td\u003e\n      \u003ctd\u003e\n        \u003ca href=\"https://dl.fbaipublicfiles.com/dinov2/dinov2_vitb14/dinov2_vitb14_voc2012_linear_head.pth\"\u003elinear\u003c/a\u003e,\n        \u003ca href=\"https://dl.fbaipublicfiles.com/dinov2/dinov2_vitb14/dinov2_vitb14_voc2012_ms_head.pth\"\u003emulti-scale\u003c/a\u003e\n      \u003c/td\u003e\n    \u003c/tr\u003e\n    \u003ctr\u003e\n      \u003ctd\u003eViT-L/14 distilled\u003c/td\u003e\n      \u003ctd\u003e\u003c/td\u003e\n      \u003ctd\u003e\n        \u003ca href=\"https://dl.fbaipublicfiles.com/dinov2/dinov2_vitl14/dinov2_vitl14_ade20k_linear_head.pth\"\u003elinear\u003c/a\u003e,\n        \u003ca href=\"https://dl.fbaipublicfiles.com/dinov2/dinov2_vitl14/dinov2_vitl14_ade20k_ms_head.pth\"\u003emulti-scale\u003c/a\u003e\n      \u003c/td\u003e\n      \u003ctd\u003e\n        \u003ca href=\"https://dl.fbaipublicfiles.com/dinov2/dinov2_vitl14/dinov2_vitl14_voc2012_linear_head.pth\"\u003elinear\u003c/a\u003e,\n        \u003ca href=\"https://dl.fbaipublicfiles.com/dinov2/dinov2_vitl14/dinov2_vitl14_voc2012_ms_head.pth\"\u003emulti-scale\u003c/a\u003e\n      \u003c/td\u003e\n    \u003c/tr\u003e\n    \u003ctr\u003e\n      \u003ctd\u003eViT-g/14\u003c/td\u003e\n      \u003ctd\u003e\n        \u003ca href=\"https://dl.fbaipublicfiles.com/dinov2/dinov2_vitg14/dinov2_vitg14_ade20k_m2f.pth\"\u003eMask2Former\u003c/a\u003e\n      \u003c/td\u003e\n      \u003ctd\u003e\n        \u003ca href=\"https://dl.fbaipublicfiles.com/dinov2/dinov2_vitg14/dinov2_vitg14_ade20k_linear_head.pth\"\u003elinear\u003c/a\u003e,\n        \u003ca href=\"https://dl.fbaipublicfiles.com/dinov2/dinov2_vitg14/dinov2_vitg14_ade20k_ms_head.pth\"\u003emulti-scale\u003c/a\u003e\n      \u003c/td\u003e\n      \u003ctd\u003e\n        \u003ca href=\"https://dl.fbaipublicfiles.com/dinov2/dinov2_vitg14/dinov2_vitg14_voc2012_linear_head.pth\"\u003elinear\u003c/a\u003e,\n        \u003ca href=\"https://dl.fbaipublicfiles.com/dinov2/dinov2_vitg14/dinov2_vitg14_voc2012_ms_head.pth\"\u003emulti-scale\u003c/a\u003e\n      \u003c/td\u003e\n    \u003c/tr\u003e\n  \u003c/tbody\u003e\n\u003c/table\u003e\n\n## Installation\n\nThe training and evaluation code requires PyTorch 2.0 and [xFormers](https://github.com/facebookresearch/xformers) 0.0.18 as well as a number of other 3rd party packages. Note that the code has only been tested with the specified versions and also expects a Linux environment. To setup all the required dependencies for training and evaluation, please follow the instructions below:\n\n*[conda](https://docs.conda.io/projects/conda/en/latest/user-guide/getting-started.html)* **(Recommended)** - Clone the repository and then create and activate a `dinov2` conda environment using the provided environment definition:\n\n```shell\nconda env create -f conda.yaml\nconda activate dinov2\n```\n\n*[pip](https://pip.pypa.io/en/stable/getting-started/)* - Clone the repository and then use the provided `requirements.txt` to install the dependencies:\n\n```shell\npip install -r requirements.txt\n```\n\nFor dense tasks (depth estimation and semantic segmentation), there are additional dependencies (specific versions of `mmcv` and `mmsegmentation`) which are captured in the `extras` dependency specifications:\n\n*[conda](https://docs.conda.io/projects/conda/en/latest/user-guide/getting-started.html)* **(Recommended)**:\n\n```shell\nconda env create -f conda-extras.yaml\nconda activate dinov2-extras\n```\n\n*[pip](https://pip.pypa.io/en/stable/getting-started/)*:\n\n```shell\npip install -r requirements.txt -r requirements-extras.txt\n```\n\n## Data preparation\n\n### ImageNet-1k\n\nThe root directory of the dataset should hold the following contents:\n\n- `\u003cROOT\u003e/test/ILSVRC2012_test_00000001.JPEG`\n- `\u003cROOT\u003e/test/[..]`\n- `\u003cROOT\u003e/test/ILSVRC2012_test_00100000.JPEG`\n- `\u003cROOT\u003e/train/n01440764/n01440764_10026.JPEG`\n- `\u003cROOT\u003e/train/[...]`\n- `\u003cROOT\u003e/train/n15075141/n15075141_9993.JPEG`\n- `\u003cROOT\u003e/val/n01440764/ILSVRC2012_val_00000293.JPEG`\n- `\u003cROOT\u003e/val/[...]`\n- `\u003cROOT\u003e/val/n15075141/ILSVRC2012_val_00049174.JPEG`\n- `\u003cROOT\u003e/labels.txt`\n\nThe provided dataset implementation expects a few additional metadata files to be present under the extra directory:\n\n- `\u003cEXTRA\u003e/class-ids-TRAIN.npy`\n- `\u003cEXTRA\u003e/class-ids-VAL.npy`\n- `\u003cEXTRA\u003e/class-names-TRAIN.npy`\n- `\u003cEXTRA\u003e/class-names-VAL.npy`\n- `\u003cEXTRA\u003e/entries-TEST.npy`\n- `\u003cEXTRA\u003e/entries-TRAIN.npy`\n- `\u003cEXTRA\u003e/entries-VAL.npy`\n\nThese metadata files can be generated (once) with the following lines of Python code:\n\n```python\nfrom dinov2.data.datasets import ImageNet\n\nfor split in ImageNet.Split:\n    dataset = ImageNet(split=split, root=\"\u003cROOT\u003e\", extra=\"\u003cEXTRA\u003e\")\n    dataset.dump_extra()\n```\n\nNote that the root and extra directories do not have to be distinct directories.\n\n### ImageNet-22k\n\nPlease adapt the [dataset class](dinov2/data/datasets/image_net_22k.py) to match your local setup.\n\n\u003cbr /\u003e\n\n:warning: To execute the commands provided in the next sections for training and evaluation, the `dinov2` package should be included in the Python module search path, i.e. simply prefix the command to run with `PYTHONPATH=.`.\n\n## Training\n\n### Fast setup: training DINOv2 ViT-L/16 on ImageNet-1k\n\nRun DINOv2 training on 4 A100-80GB nodes (32 GPUs) in a SLURM cluster environment with submitit:\n\n```shell\npython dinov2/run/train/train.py \\\n    --nodes 4 \\\n    --config-file dinov2/configs/train/vitl16_short.yaml \\\n    --output-dir \u003cPATH/TO/OUTPUT/DIR\u003e \\\n    train.dataset_path=ImageNet:split=TRAIN:root=\u003cPATH/TO/DATASET\u003e:extra=\u003cPATH/TO/DATASET\u003e\n```\n\nTraining time is approximately 1 day and the resulting checkpoint should reach 81.6% on k-NN eval and 82.9% on linear eval.\n\nThe training code saves the weights of the teacher in the `eval` folder every 12500 iterations for evaluation.\n\n### Long setup: training DINOv2 ViT-L/14 on ImageNet-22k\n\nRun DINOv2 training on 12 A100-80GB nodes (96 GPUs) in a SLURM cluster environment with submitit:\n\n```shell\npython dinov2/run/train/train.py \\\n    --nodes 12 \\\n    --config-file dinov2/configs/train/vitl14.yaml \\\n    --output-dir \u003cPATH/TO/OUTPUT/DIR\u003e \\\n    train.dataset_path=ImageNet22k:root=\u003cPATH/TO/DATASET\u003e:extra=\u003cPATH/TO/DATASET\u003e\n```\n\nTraining time is approximately 3.3 days and the resulting checkpoint should reach 82.0% on k-NN eval and 84.5% on linear eval.\n\nThe training code saves the weights of the teacher in the `eval` folder every 12500 iterations for evaluation.\n\n\n## Evaluation\n\nThe training code regularly saves the teacher weights. In order to evaluate the model, run the following evaluation on a single node:\n\n### k-NN classification on ImageNet-1k\n\n```shell\npython dinov2/run/eval/knn.py \\\n    --config-file \u003cPATH/TO/OUTPUT/DIR\u003e/config.yaml \\\n    --pretrained-weights \u003cPATH/TO/OUTPUT/DIR\u003e/eval/training_24999/teacher_checkpoint.pth \\\n    --output-dir \u003cPATH/TO/OUTPUT/DIR\u003e/eval/training_24999/knn \\\n    --train-dataset ImageNet:split=TRAIN:root=\u003cPATH/TO/DATASET\u003e:extra=\u003cPATH/TO/DATASET\u003e \\\n    --val-dataset ImageNet:split=VAL:root=\u003cPATH/TO/DATASET\u003e:extra=\u003cPATH/TO/DATASET\u003e\n```\n\n### Logistic regression classification on ImageNet-1k\n\n```shell\npython dinov2/run/eval/log_regression.py \\\n    --config-file \u003cPATH/TO/OUTPUT/DIR\u003e/config.yaml \\\n    --pretrained-weights \u003cPATH/TO/OUTPUT/DIR\u003e/eval/training_24999/teacher_checkpoint.pth \\\n    --output-dir \u003cPATH/TO/OUTPUT/DIR\u003e/eval/training_24999/logreg \\\n    --train-dataset ImageNet:split=TRAIN:root=\u003cPATH/TO/DATASET\u003e:extra=\u003cPATH/TO/DATASET\u003e \\\n    --val-dataset ImageNet:split=VAL:root=\u003cPATH/TO/DATASET\u003e:extra=\u003cPATH/TO/DATASET\u003e\n```\n\n### Linear classification with data augmentation on ImageNet-1k\n\n```shell\npython dinov2/run/eval/linear.py \\\n    --config-file \u003cPATH/TO/OUTPUT/DIR\u003e/config.yaml \\\n    --pretrained-weights \u003cPATH/TO/OUTPUT/DIR\u003e/eval/training_24999/teacher_checkpoint.pth \\\n    --output-dir \u003cPATH/TO/OUTPUT/DIR\u003e/eval/training_24999/linear \\\n    --train-dataset ImageNet:split=TRAIN:root=\u003cPATH/TO/DATASET\u003e:extra=\u003cPATH/TO/DATASET\u003e \\\n    --val-dataset ImageNet:split=VAL:root=\u003cPATH/TO/DATASET\u003e:extra=\u003cPATH/TO/DATASET\u003e\n```\n\nWe release the weights from evaluating the different models:\n\n\u003ctable style=\"margin: auto\"\u003e\n  \u003ctr\u003e\n    \u003cth\u003emodel\u003c/th\u003e\n    \u003cth\u003ewith\u003cbr /\u003eregisters\u003c/th\u003e\n    \u003cth\u003eImageNet\u003cbr /\u003etop-1\u003c/th\u003e\n    \u003cth\u003elinear evaluation\u003c/th\u003e\n  \u003c/tr\u003e\n  \u003ctr\u003e\n    \u003ctd\u003eViT-S/14 distilled\u003c/td\u003e\n    \u003ctd align=\"center\"\u003e:x:\u003c/td\u003e\n    \u003ctd align=\"right\"\u003e81.1%\u003c/td\u003e\n    \u003ctd\u003e\u003ca href=\"https://dl.fbaipublicfiles.com/dinov2/dinov2_vits14/dinov2_vits14_linear_head.pth\"\u003elinear head weights\u003c/a\u003e\u003c/td\u003e\n  \u003c/tr\u003e\n  \u003ctr\u003e\n    \u003ctd\u003eViT-S/14 distilled\u003c/td\u003e\n    \u003ctd align=\"center\"\u003e:white_check_mark:\u003c/td\u003e\n    \u003ctd align=\"right\"\u003e80.8%\u003c/td\u003e\n    \u003ctd\u003e\u003ca href=\"https://dl.fbaipublicfiles.com/dinov2/dinov2_vits14/dinov2_vits14_reg4_linear_head.pth\"\u003elinear head weights\u003c/a\u003e\u003c/td\u003e\n  \u003c/tr\u003e\n  \u003ctr\u003e\n    \u003ctd\u003eViT-B/14 distilled\u003c/td\u003e\n    \u003ctd align=\"center\"\u003e:x:\u003c/td\u003e\n    \u003ctd align=\"right\"\u003e84.5%\u003c/td\u003e\n    \u003ctd\u003e\u003ca href=\"https://dl.fbaipublicfiles.com/dinov2/dinov2_vitb14/dinov2_vitb14_linear_head.pth\"\u003elinear head weights\u003c/a\u003e\u003c/td\u003e\n  \u003c/tr\u003e\n  \u003ctr\u003e\n    \u003ctd\u003eViT-B/14 distilled\u003c/td\u003e\n    \u003ctd align=\"center\"\u003e:white_check_mark:\u003c/td\u003e\n    \u003ctd align=\"right\"\u003e84.4%\u003c/td\u003e\n    \u003ctd\u003e\u003ca href=\"https://dl.fbaipublicfiles.com/dinov2/dinov2_vitb14/dinov2_vitb14_reg4_linear_head.pth\"\u003elinear head weights\u003c/a\u003e\u003c/td\u003e\n  \u003c/tr\u003e\n  \u003ctr\u003e\n    \u003ctd\u003eViT-L/14 distilled\u003c/td\u003e\n    \u003ctd align=\"center\"\u003e:x:\u003c/td\u003e\n    \u003ctd align=\"right\"\u003e86.3%\u003c/td\u003e\n    \u003ctd\u003e\u003ca href=\"https://dl.fbaipublicfiles.com/dinov2/dinov2_vitl14/dinov2_vitl14_linear_head.pth\"\u003elinear head weights\u003c/a\u003e\u003c/td\u003e\n  \u003c/tr\u003e\n  \u003ctr\u003e\n    \u003ctd\u003eViT-L/14 distilled\u003c/td\u003e\n    \u003ctd align=\"center\"\u003e:white_check_mark:\u003c/td\u003e\n    \u003ctd align=\"right\"\u003e86.5%\u003c/td\u003e\n    \u003ctd\u003e\u003ca href=\"https://dl.fbaipublicfiles.com/dinov2/dinov2_vitl14/dinov2_vitl14_reg4_linear_head.pth\"\u003elinear head weights\u003c/a\u003e\u003c/td\u003e\n  \u003c/tr\u003e\n  \u003ctr\u003e\n    \u003ctd\u003eViT-g/14\u003c/td\u003e\n    \u003ctd align=\"center\"\u003e:x:\u003c/td\u003e\n    \u003ctd align=\"right\"\u003e86.5%\u003c/td\u003e\n    \u003ctd\u003e\u003ca href=\"https://dl.fbaipublicfiles.com/dinov2/dinov2_vitg14/dinov2_vitg14_linear_head.pth\"\u003elinear head weights\u003c/a\u003e\u003c/td\u003e\n  \u003c/tr\u003e\n  \u003ctr\u003e\n    \u003ctd\u003eViT-g/14\u003c/td\u003e\n    \u003ctd align=\"center\"\u003e:white_check_mark:\u003c/td\u003e\n    \u003ctd align=\"right\"\u003e87.0%\u003c/td\u003e\n    \u003ctd\u003e\u003ca href=\"https://dl.fbaipublicfiles.com/dinov2/dinov2_vitg14/dinov2_vitg14_reg4_linear_head.pth\"\u003elinear head weights\u003c/a\u003e\u003c/td\u003e\n  \u003c/tr\u003e\n\u003c/table\u003e\n\nThe performance of the provided pretrained model weights can be evaluated as follows on ImageNet-1k:\n\n```shell\npython dinov2/run/eval/linear.py \\\n    --config-file dinov2/configs/eval/vitg14_pretrain.yaml \\\n    --pretrained-weights https://dl.fbaipublicfiles.com/dinov2/dinov2_vitg14/dinov2_vitg14_pretrain.pth \\\n    --train-dataset ImageNet:split=TRAIN:root=\u003cPATH/TO/DATASET\u003e:extra=\u003cPATH/TO/DATASET\u003e \\\n    --val-dataset ImageNet:split=VAL:root=\u003cPATH/TO/DATASET\u003e:extra=\u003cPATH/TO/DATASET\u003e\n```\n\n## Notebooks\n\nA few notebooks are provided to help the community leverage the models and code:\n\n\u003cul\u003e\n  \u003cli\u003e\u003ca href=\"https://github.com/facebookresearch/dinov2/blob/main/notebooks/depth_estimation.ipynb\"\u003eDepth estimation\u003c/a\u003e - How to load and use the depth heads in combination with a matching backbone via mmcv\u003c/li\u003e\n  \u003cli\u003e\u003ca href=\"https://github.com/facebookresearch/dinov2/blob/main/notebooks/semantic_segmentation.ipynb\"\u003eSemantic segmentation\u003c/a\u003e - How to load and use the segmentation heads in combination with a matching backbone via mmcv, and also how to load and use the Mask2Former-based segmentation model trained on ADE20K\u003c/li\u003e\n\u003c/ul\u003e\n\n## License\n\nDINOv2 code and model weights are released under the Apache License 2.0. See [LICENSE](LICENSE) for additional details.\n\n## Contributing\n\nSee [contributing](CONTRIBUTING.md) and the [code of conduct](CODE_OF_CONDUCT.md).\n\n## Citing DINOv2\n\nIf you find this repository useful, please consider giving a star :star: and citation :t-rex::\n\n```\n@misc{oquab2023dinov2,\n  title={DINOv2: Learning Robust Visual Features without Supervision},\n  author={Oquab, Maxime and Darcet, Timothée and Moutakanni, Theo and Vo, Huy V. and Szafraniec, Marc and Khalidov, Vasil and Fernandez, Pierre and Haziza, Daniel and Massa, Francisco and El-Nouby, Alaaeldin and Howes, Russell and Huang, Po-Yao and Xu, Hu and Sharma, Vasu and Li, Shang-Wen and Galuba, Wojciech and Rabbat, Mike and Assran, Mido and Ballas, Nicolas and Synnaeve, Gabriel and Misra, Ishan and Jegou, Herve and Mairal, Julien and Labatut, Patrick and Joulin, Armand and Bojanowski, Piotr},\n  journal={arXiv:2304.07193},\n  year={2023}\n}\n```\n\n```\n@misc{darcet2023vitneedreg,\n  title={Vision Transformers Need Registers},\n  author={Darcet, Timothée and Oquab, Maxime and Mairal, Julien and Bojanowski, Piotr},\n  journal={arXiv:2309.16588},\n  year={2023}\n}\n```\n","project_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Ffacebookresearch%2Fdinov2","html_url":"https://awesome.ecosyste.ms/projects/github.com%2Ffacebookresearch%2Fdinov2","lists_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Ffacebookresearch%2Fdinov2/lists"}