{"id":37622901,"url":"https://github.com/georg-wolflein/pathology-foundation-models","last_synced_at":"2026-01-16T10:42:43.063Z","repository":{"id":243688856,"uuid":"813142401","full_name":"georg-wolflein/pathology-foundation-models","owner":"georg-wolflein","description":"List of pathology feature extractors and foundation models","archived":false,"fork":false,"pushed_at":"2025-06-09T17:06:55.000Z","size":43,"stargazers_count":130,"open_issues_count":3,"forks_count":10,"subscribers_count":5,"default_branch":"master","last_synced_at":"2025-06-09T18:23:37.849Z","etag":null,"topics":["computational-pathology","deep-learning","feature-extraction","foundation-models","pathology","self-supervised-learning"],"latest_commit_sha":null,"homepage":"","language":null,"has_issues":true,"has_wiki":null,"has_pages":null,"mirror_url":null,"source_name":null,"license":null,"status":null,"scm":"git","pull_requests_enabled":true,"icon_url":"https://github.com/georg-wolflein.png","metadata":{"files":{"readme":"README.md","changelog":null,"contributing":null,"funding":null,"license":null,"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,"zenodo":null}},"created_at":"2024-06-10T14:53:21.000Z","updated_at":"2025-06-09T17:06:58.000Z","dependencies_parsed_at":"2024-06-10T17:37:12.582Z","dependency_job_id":"599d742b-a145-4cad-9722-78471922f6d2","html_url":"https://github.com/georg-wolflein/pathology-foundation-models","commit_stats":null,"previous_names":["georg-wolflein/pathology-feature-extractors","georg-wolflein/pathology-foundation-models"],"tags_count":0,"template":false,"template_full_name":null,"purl":"pkg:github/georg-wolflein/pathology-foundation-models","repository_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/georg-wolflein%2Fpathology-foundation-models","tags_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/georg-wolflein%2Fpathology-foundation-models/tags","releases_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/georg-wolflein%2Fpathology-foundation-models/releases","manifests_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/georg-wolflein%2Fpathology-foundation-models/manifests","owner_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners/georg-wolflein","download_url":"https://codeload.github.com/georg-wolflein/pathology-foundation-models/tar.gz/refs/heads/master","sbom_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/georg-wolflein%2Fpathology-foundation-models/sbom","scorecard":null,"host":{"name":"GitHub","url":"https://github.com","kind":"github","repositories_count":286080680,"owners_count":28478050,"icon_url":"https://github.com/github.png","version":null,"created_at":"2022-05-30T11:31:42.601Z","updated_at":"2026-01-16T06:30:42.265Z","status":"ssl_error","status_checked_at":"2026-01-16T06:30:16.248Z","response_time":107,"last_error":"SSL_read: unexpected eof while reading","robots_txt_status":"success","robots_txt_updated_at":"2025-07-24T06:49:26.215Z","robots_txt_url":"https://github.com/robots.txt","online":false,"can_crawl_api":true,"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":["computational-pathology","deep-learning","feature-extraction","foundation-models","pathology","self-supervised-learning"],"created_at":"2026-01-16T10:42:42.479Z","updated_at":"2026-01-16T10:42:43.054Z","avatar_url":"https://github.com/georg-wolflein.png","language":null,"funding_links":[],"categories":[],"sub_categories":[],"readme":"\u003cdiv align=\"center\"\u003e\n\u003ch1\u003ePathology Feature Extractors and Foundation Models\u003c/h1\u003e\n\u003c/div\u003e\n\nWe are witnessing the emergence of many new feature extractors trained using self-supervised learning on large pathology datasets.\nThis repository aims to provide a comprehensive list of these models, alongside key information about them.\n\nI aim to update this list as new models are released, but please submit a pull request / issue for any models I have missed!\n\n## Patch-level models\n\n| Name                                                                                                                                              | Group                                                                                                           | Weights            | Released                                                                                              | SSL                                                                             | WSIs                            | Tiles    | Patients   | Batch size | Iterations | Architecture           | Parameters | Embed dim | Input size | Dataset                                          | Links                                                                                                                                                                                                                                                                                                                                        |\n| ------------------------------------------------------------------------------------------------------------------------------------------------- | --------------------------------------------------------------------------------------------------------------- | ------------------ | ----------------------------------------------------------------------------------------------------- | ------------------------------------------------------------------------------- | ------------------------------- | -------- | ---------- | ---------- | ---------- | ---------------------- | ---------- | --------- | ---------- | ------------------------------------------------ | -------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |\n| [CTransPath](https://www.sciencedirect.com/science/article/abs/pii/S1361841522002043)                                                             | Sichuan University / Tencent AI Lab                                                                             | :white_check_mark: | Dec 2021[\\*](https://github.com/Xiyue-Wang/TransPath/commit/4b1c67655dd38cb192567b0981b6c1e9ade59ecf) | [SRCL](https://www.sciencedirect.com/science/article/abs/pii/S1361841522002043) | 32K                             | 16M      |            |            |            | Swin-Transformer       | 28M        | 768       | 224        | TCGA, PAIP                                       | [\u003cimg src=\"https://raw.githubusercontent.com/FortAwesome/Font-Awesome/6.x/svgs/brands/github.svg\" width=\"20\"\u003e](https://github.com/Xiyue-Wang/TransPath)                                                                                                                                                                                      |\n| [RetCCL](https://www.sciencedirect.com/science/article/abs/pii/S1361841522002043)                                                                 | Sichuan University / Tencent AI Lab                                                                             | :white_check_mark: | Dec 2021[\\*](https://github.com/Xiyue-Wang/RetCCL/commit/e6faf0bd85c8e7e617882dd5d74e644d28eac771)    | [CCL](https://www.sciencedirect.com/science/article/abs/pii/S1361841522002043)  | 32K                             | 16M      |            |            |            | ResNet-50              | 26M        | 2048      | 224        | TCGA, PAIP                                       | [\u003cimg src=\"https://raw.githubusercontent.com/FortAwesome/Font-Awesome/6.x/svgs/brands/github.svg\" width=\"20\"\u003e](https://github.com/Xiyue-Wang/RetCCL)                                                                                                                                                                                         |\n| [REMEDIS](https://www.nature.com/articles/s41551-023-01049-7)                                                                                     | [Google Research](https://research.google)                                                                      | :white_check_mark: | May 2022[\\*](https://arxiv.org/abs/2205.09723v1)                                                      | SimCLR/BiT                                                                      | 29K                             | 50M      | 11K cases  | 4096       | 1.2M       | ResNet-50              | 26M        | 2048      | 224        | TCGA                                             | [\u003cimg src=\"https://raw.githubusercontent.com/FortAwesome/Font-Awesome/6.x/svgs/brands/github.svg\" width=\"20\"\u003e](https://github.com/google-research/medical-ai-research-foundations)                                                                                                                                                           |\n| [HIPT](https://ieeexplore.ieee.org/document/9880275)                                                                                              | [Mahmood Lab](https://faisal.ai)                                                                                | :white_check_mark: | Jun 2022[\\*](https://arxiv.org/abs/2206.02647v1)                                                      | DINOv1                                                                          | 11K                             | 100M     |            | 256        | 400K       | ViT-S                  | 22M        | 384       | 256        | TCGA                                             | [\u003cimg src=\"https://raw.githubusercontent.com/FortAwesome/Font-Awesome/6.x/svgs/brands/github.svg\" width=\"20\"\u003e](https://github.com/mahmoodlab/HIPT)                                                                                                                                                                                           |\n| [Lunit-DINO](https://arxiv.org/abs/2212.04690)                                                                                                    | [Lunit](https://www.lunit.io)                                                                                   | :white_check_mark: | Dec 2022[\\*](https://arxiv.org/abs/2212.04690v1)                                                      | DINOv1                                                                          | 21K                             |          |            |            |            | ViT-S                  | 22M        | 384       | 224        | TCGA                                             | [\u003cimg src=\"https://raw.githubusercontent.com/FortAwesome/Font-Awesome/6.x/svgs/brands/github.svg\" width=\"20\"\u003e](https://github.com/lunit-io/benchmark-ssl-pathology)                                                                                                                                                                          |\n| [Lunit-{BT,MoCoV2,SwAV}](https://arxiv.org/abs/2212.04690)                                                                                        | [Lunit](https://www.lunit.io)                                                                                   | :white_check_mark: | Dec 2022[\\*](https://arxiv.org/abs/2212.04690v1)                                                      | {BT,MoCoV2,SwAV}                                                                | 21K                             |          |            |            |            | ResNet-50              |            | 2048      | 224        | TCGA                                             | [\u003cimg src=\"https://raw.githubusercontent.com/FortAwesome/Font-Awesome/6.x/svgs/brands/github.svg\" width=\"20\"\u003e](https://github.com/lunit-io/benchmark-ssl-pathology)                                                                                                                                                                          |\n| [Phikon](https://www.medrxiv.org/content/10.1101/2023.07.21.23292757v2)                                                                           | [Owkin](https://www.owkin.com)                                                                                  | :white_check_mark: | Jul 2023[\\*](https://www.medrxiv.org/content/10.1101/2023.07.21.23292757v1)                           | iBOT                                                                            | 6.1K                            | 43M      | 5.6K       | 1440       | 155K       | ViT-B                  | 86M        | 768       | 224        | TCGA                                             | [\u003cimg src=\"https://raw.githubusercontent.com/FortAwesome/Font-Awesome/6.x/svgs/brands/github.svg\" width=\"20\"\u003e](https://github.com/owkin/HistoSSLscaling) [\u003cimg src=\"https://huggingface.co/datasets/huggingface/brand-assets/resolve/main/hf-logo.svg\" width=\"25\"\u003e](https://huggingface.co/owkin/phikon)                                     |\n| [CONCH](https://www.nature.com/articles/s41591-024-02856-4) (VL)                                                                                  | [Mahmood Lab](https://faisal.ai)                                                                                | :white_check_mark: | Jul 2023[\\*](https://arxiv.org/abs/2307.12914v1)                                                      | iBOT \u0026 vision-language pretraining                                              | 21K                             | 16M      |            | 1024       | 80 epochs  | ViT-B                  | 86M        | 768       | 224        | proprietary                                      | [\u003cimg src=\"https://raw.githubusercontent.com/FortAwesome/Font-Awesome/6.x/svgs/brands/github.svg\" width=\"20\"\u003e](https://github.com/mahmoodlab/CONCH) [\u003cimg src=\"https://huggingface.co/datasets/huggingface/brand-assets/resolve/main/hf-logo.svg\" width=\"25\"\u003e](https://huggingface.co/MahmoodLab/CONCH)                                      |\n| **[UNI](https://www.nature.com/articles/s41591-024-02857-3)**                                                                                     | [Mahmood Lab](https://faisal.ai)                                                                                | :white_check_mark: | Aug 2023[\\*](https://arxiv.org/abs/2308.15474v1)                                                      | DINOv2                                                                          | **100K**                        | 100M     |            |            |            | ViT-L                  |            | 1024      | 224        | proprietary (Mass-100K)                          | [\u003cimg src=\"https://raw.githubusercontent.com/FortAwesome/Font-Awesome/6.x/svgs/brands/github.svg\" width=\"20\"\u003e](https://github.com/mahmoodlab/UNI) [\u003cimg src=\"https://huggingface.co/datasets/huggingface/brand-assets/resolve/main/hf-logo.svg\" width=\"25\"\u003e](https://huggingface.co/MahmoodLab/UNI)                                          |\n| **[Virchow](https://www.nature.com/articles/s41591-024-03141-0)**                                                                                 | [Paige](https://paige.ai) / [Microsoft](https://www.microsoft.com/en-us/research/lab/microsoft-health-futures)  | :white_check_mark: | Sep 2023[\\*](https://arxiv.org/abs/2309.07778v1)                                                      | DINOv2                                                                          | **1.5M**                        |          | 120K       |            |            | ViT-H                  | 632M       | 2560      | 224        | proprietary (from MSKCC)                         | [\u003cimg src=\"https://huggingface.co/datasets/huggingface/brand-assets/resolve/main/hf-logo.svg\" width=\"25\"\u003e](https://huggingface.co/paige-ai/Virchow)                                                                                                                                                                                          |\n| **[Campanella _et al._](https://arxiv.org/abs/2310.07033)** (DINO)                                                                                | [Thomas Fuchs Lab](https://www.hpims.org/labs/thomas-fuchs-lab/)                                                | :white_check_mark: | Oct 2023[\\*](https://arxiv.org/abs/2310.07033v1)                                                      | DINOv1                                                                          | **420K**                        | 3.3B     | 77K        | 1080       | 1.3K INE   | ViT-S                  | 22M        | 384       | 224        | proprietary (MSHS)                               | [\u003cimg src=\"https://huggingface.co/datasets/huggingface/brand-assets/resolve/main/hf-logo.svg\" width=\"25\"\u003e](https://huggingface.co/MountSinaiCompPath/SP22M) ([\u003cimg src=\"https://huggingface.co/datasets/huggingface/brand-assets/resolve/main/hf-logo.svg\" width=\"25\"\u003e](https://huggingface.co/MountSinaiCompPath/SP85M))                    |\n| **[Campanella _et al._](https://arxiv.org/abs/2310.07033)** (MAE)                                                                                 | [Thomas Fuchs Lab](https://www.hpims.org/labs/thomas-fuchs-lab/)                                                | :x:                | Oct 2023[\\*](https://arxiv.org/abs/2310.07033v1)                                                      | MAE                                                                             | **420K**                        | 3.3B     | 77K        | 1440       | 2.5K INE   | ViT-L                  | 303M       | 1024      | 224        | proprietary (MSHS)                               |\n| [Path Foundation](https://arxiv.org/abs/2310.13259)                                                                                               | [Google](https://research.google)                                                                               | :white_check_mark: | Oct 2023[\\*](https://arxiv.org/abs/2310.13259v1)                                                      | SimCLR, MSN                                                                     | 6K                              | 60M      |            | 1024       |            | ViT-S                  | 22M        | 384       | 224        | TCGA                                             | [\u003cimg src=\"https://raw.githubusercontent.com/FortAwesome/Font-Awesome/6.x/svgs/brands/github.svg\" width=\"20\"\u003e](https://github.com/Google-Health/imaging-research/tree/master/path-foundation)                                                                                                                                                |\n| [PathoDuet](https://arxiv.org/abs/2312.09894)                                                                                                     | [Shanghai Jiao Tong University](https://life.sjtu.edu.cn/)                                                      | :white_check_mark: | Dec 2023[\\*](https://arxiv.org/abs/2312.09894v1)                                                      | inspired by MoCoV3                                                              | 11K                             | 13M      |            | 2048       | 100 epochs | ViT-B                  |            | 4096      | 224        | TCGA                                             | [\u003cimg src=\"https://raw.githubusercontent.com/FortAwesome/Font-Awesome/6.x/svgs/brands/github.svg\" width=\"20\"\u003e](https://github.com/openmedlab/PathoDuet)                                                                                                                                                                                      |\n| **[RudolfV](https://arxiv.org/abs/2401.04079)**                                                                                                   | [Aignostics](https://www.aignostics.com)                                                                        | :x:                | Jan 2024[\\*](https://arxiv.org/abs/2401.04079v1)                                                      | DINOv2                                                                          | **130K**                        | 750M     | 36K        |            |            | ViT-L                  | 300M       |           | 224        | proprietary (from EU \u0026 US), TCGA                 |\n| [kaiko](https://arxiv.org/abs/2404.15217)                                                                                                         | [kaiko.ai](https://www.kaiko.ai)                                                                                | :white_check_mark: | Mar 2024[\\*](https://arxiv.org/abs/2404.15217v1)                                                      | DINOv2                                                                          | 29K                             | 260M\\*\\* |            | 512        | 200 INE    | ViT-L                  |            | 1024      | 224        | TCGA                                             | [\u003cimg src=\"https://raw.githubusercontent.com/FortAwesome/Font-Awesome/6.x/svgs/brands/github.svg\" width=\"20\"\u003e](https://github.com/kaiko-ai/towards_large_pathology_fms)                                                                                                                                                                      |\n| **[PLUTO](https://arxiv.org/abs/2405.07905)**                                                                                                     | [PathAI](https://www.pathai.com)                                                                                | :x:                | May 2024[\\*](https://arxiv.org/abs/2405.07905v1)                                                      | DINOv2 (+ MAE and Fourier loss)                                                 | **160K**                        | 200M     |            |            |            | FlexiViT-S             | 22M        |           | 224        | proprietary (PathAI)                             |                                                                                                                                                                                                                                                                                                                                              |\n| [BEPH](https://www.biorxiv.org/content/10.1101/2024.05.16.594499)                                                                                 | [Shanghai Jiao Tong University](https://life.sjtu.edu.cn/)                                                      | :white_check_mark: | May 2024[\\*](https://www.biorxiv.org/content/10.1101/2024.05.16.594499v1)                             | BEiTv2                                                                          | 12K                             | 12M      |            | 1024       |            | ViT-B                  | 193M       | 1024      | 224        | TCGA                                             | [\u003cimg src=\"https://raw.githubusercontent.com/FortAwesome/Font-Awesome/6.x/svgs/brands/github.svg\" width=\"20\"\u003e](https://github.com/Zhcyoung/BEPH)                                                                                                                                                                                             |\n| **[Prov-GigaPath](https://www.nature.com/articles/s41586-024-07441-w)**                                                                           | [Microsoft](https://www.microsoft.com/en-us/research/) / [Providence](https://www.providence.org)               | :white_check_mark: | May 2024[\\*](https://www.nature.com/articles/s41586-024-07441-w)                                      | DINOv2                                                                          | **170K**                        | 1.4B     | 30K        | 384        |            | ViT                    |            | 1536      | 224        | proprietary (Providence)                         | [\u003cimg src=\"https://raw.githubusercontent.com/FortAwesome/Font-Awesome/6.x/svgs/brands/github.svg\" width=\"20\"\u003e](https://github.com/prov-gigapath/prov-gigapath) [\u003cimg src=\"https://huggingface.co/datasets/huggingface/brand-assets/resolve/main/hf-logo.svg\" width=\"25\"\u003e](https://huggingface.co/prov-gigapath/prov-gigapath)                |\n| **[Hibou-B](https://arxiv.org/abs/2406.05074)**                                                                                                   | [HistAI](https://www.hist.ai)                                                                                   | :white_check_mark: | Jun 2024[\\*](https://arxiv.org/abs/2406.05074v1)                                                      | DINOv2                                                                          | **1.1M**                        | 510M     | 310K cases | 1024       | 500K       | ViT-B                  | 86M        | 768       | 224        | proprietary                                      | [\u003cimg src=\"https://raw.githubusercontent.com/FortAwesome/Font-Awesome/6.x/svgs/brands/github.svg\" width=\"20\"\u003e](https://github.com/HistAI/hibou) [\u003cimg src=\"https://huggingface.co/datasets/huggingface/brand-assets/resolve/main/hf-logo.svg\" width=\"25\"\u003e](https://huggingface.co/histai/hibou-b)                                            |\n| **[Hibou-L](https://arxiv.org/abs/2406.05074)**                                                                                                   | [HistAI](https://www.hist.ai)                                                                                   | :white_check_mark: | Jun 2024[\\*](https://arxiv.org/abs/2406.05074v1)                                                      | DINOv2                                                                          | **1.1M**                        | 1.2B     | 310K cases | 1024       | 1.2M       | ViT-L                  | 304M       | 1024      | 224        | proprietary                                      | [\u003cimg src=\"https://raw.githubusercontent.com/FortAwesome/Font-Awesome/6.x/svgs/brands/github.svg\" width=\"20\"\u003e](https://github.com/HistAI/hibou) [\u003cimg src=\"https://huggingface.co/datasets/huggingface/brand-assets/resolve/main/hf-logo.svg\" width=\"25\"\u003e](https://huggingface.co/histai/hibou-L)                                            |\n| **[H-optimus-0](https://www.bioptimus.com/news/bioptimus-launches-h-optimus-0-the-worlds-largest-open-source-ai-foundation-model-for-pathology)** | [Bioptimus](https://www.bioptimus.com)                                                                          | :white_check_mark: | Jul 2024[\\*](https://github.com/bioptimus/releases/commit/f967dd8d6de387fc0926cbe29b35b3cc5abc5500)   | DINOv2/iBOT                                                                     | **500K** (across 4,000 clinics) | \u003e100M    | 200K       |            |            | ViT-G with 4 registers | 1.1B       | 1536      | 224        | proprietary                                      | [\u003cimg src=\"https://raw.githubusercontent.com/FortAwesome/Font-Awesome/6.x/svgs/brands/github.svg\" width=\"20\"\u003e](https://github.com/bioptimus/releases/tree/main/models/h-optimus/v0) [\u003cimg src=\"https://huggingface.co/datasets/huggingface/brand-assets/resolve/main/hf-logo.svg\" width=\"25\"\u003e](https://huggingface.co/bioptimus/H-optimus-0) |\n| [mSTAR](https://arxiv.org/abs/2407.15362) (VL)                                                                                                    | [Smart Lab](https://hkustsmartlab.github.io)                                                                    | :x:                | Jul 2024[\\*](https://arxiv.org/abs/2407.15362v1)                                                      | mSTAR (multimodal)                                                              | 10K                             |          | 10K        |            |            | ViT-L                  |            |           | 224        | TCGA                                             |                                                                                                                                                                                                                                                                                                                                              |\n| **[Virchow 2](https://arxiv.org/abs/2408.00738)**                                                                                                 | [Paige](https://paige.ai) / [Microsoft](https://www.microsoft.com/en-us/research/lab/microsoft-health-futures)  | :white_check_mark: | Aug 2024[\\*](https://arxiv.org/abs/2408.00738v1)                                                      | DINOv2 (+ ECT and KDE)                                                          | **3.1M**                        | 2B       | 230K       | 4096       |            | ViT-H with 4 registers | 632M       | 3584      | 224        | proprietary (from MSKCC and international sites) | [\u003cimg src=\"https://huggingface.co/datasets/huggingface/brand-assets/resolve/main/hf-logo.svg\" width=\"25\"\u003e](https://huggingface.co/paige-ai/Virchow2)                                                                                                                                                                                         |\n| **[Virchow 2G](https://arxiv.org/abs/2408.00738)**                                                                                                | [Paige](https://paige.ai) / [Microsoft](https://www.microsoft.com/en-us/research/lab/microsoft-health-futures)  | :x:                | Aug 2024[\\*](https://arxiv.org/abs/2408.00738v1)                                                      | DINOv2 (+ ECT and KDE)                                                          | **3.1M**                        | 2B       | 230K       | 3072       |            | ViT-G with 8 registers | 1.9B       | 3584      | 224        | proprietary (from MSKCC and international sites) |                                                                                                                                                                                                                                                                                                                                              |\n| [Phikon-v2](https://arxiv.org/abs/2409.09173)                                                                                                     | [Owkin](https://www.owkin.com)                                                                                  | :white_check_mark: | Sep 2024[\\*](https://arxiv.org/abs/2409.09173v1)                                                      | DINOv2                                                                          | 58.4K                           | 456M     |            | 4096       | 250K       | ViT-L                  | 307M       | 1024      | 224        | PANCAN-XL (TCGA, CPTAC, GTEx, proprietary)       | [\u003cimg src=\"https://huggingface.co/datasets/huggingface/brand-assets/resolve/main/hf-logo.svg\" width=\"25\"\u003e](https://huggingface.co/owkin/phikon-v2)                                                                                                                                                                                           |\n| [MUSK](https://www.nature.com/articles/s41586-024-08378-w)\u003csub\u003eV\u003c/sub\u003e (VL)                                                                       | [Li Lab (Stanford)](https://med.stanford.edu/lilab.html)                                                        | :white_check_mark: | Jan 2025[\\*](https://www.nature.com/articles/s41586-024-08378-w)                                      | Unified masked modeling (MLM, MIM) + contrastive learning                       | 33K                             | 50M      | 12K        | 2048       | 20 epochs  | BEiT3                  |            |           | 384        | TCGA                                             | [\u003cimg src=\"https://raw.githubusercontent.com/FortAwesome/Font-Awesome/6.x/svgs/brands/github.svg\" width=\"20\"\u003e](https://github.com/lilab-stanford/MUSK) [\u003cimg src=\"https://huggingface.co/datasets/huggingface/brand-assets/resolve/main/hf-logo.svg\" width=\"25\"\u003e](https://huggingface.co/xiangjx/musk)                                       |\n| **[RudolfV2](https://arxiv.org/abs/2501.05409)**                                                                                                  | [Mayo](https://www.mayoclinic.org), [Charité](https://www.charite.de), [Aignostics](https://www.aignostics.com) | :x:                | Jan 2025[\\*](https://arxiv.org/abs/2501.05409v1)                                                      |                                                                                 | **1.2M**                        | 3.4B     | 490K cases |            |            | ViT-H                  | 632M       |           |            |                                                  |                                                                                                                                                                                                                                                                                                                                              |\n| **[UNI](https://www.nature.com/articles/s41591-024-02857-3)2-h**                                                                                  | [Mahmood Lab](https://faisal.ai)                                                                                | :white_check_mark: | Jan 2025[\\*](https://github.com/mahmoodlab/UNI/commit/91e141ce3d8aca1be0de83a89d0caab45af6a470)       | DINOv2                                                                          | **350K**                        | 200M     |            |            |            | ViT-H with 8 registers | 681M       | 1536      | 224        | proprietary (Mass)                               | [\u003cimg src=\"https://raw.githubusercontent.com/FortAwesome/Font-Awesome/6.x/svgs/brands/github.svg\" width=\"20\"\u003e](https://github.com/mahmoodlab/UNI) [\u003cimg src=\"https://huggingface.co/datasets/huggingface/brand-assets/resolve/main/hf-logo.svg\" width=\"25\"\u003e](https://huggingface.co/MahmoodLab/UNI2-h)                                       |\n| **[UNI](https://www.nature.com/articles/s41591-024-02857-3)2-g-preview**                                                                          | [Mahmood Lab](https://faisal.ai)                                                                                | :x:                | Jan 2025[\\*](https://github.com/mahmoodlab/UNI/commit/91e141ce3d8aca1be0de83a89d0caab45af6a470)       | DINOv2                                                                          | **350K**                        | 200M     |            |            |            | ViT-G                  |            |           |            | proprietary (Mass)                               | [\u003cimg src=\"https://raw.githubusercontent.com/FortAwesome/Font-Awesome/6.x/svgs/brands/github.svg\" width=\"20\"\u003e](https://github.com/mahmoodlab/UNI)                                                                                                                                                                                            |\n\n\nNotes:\n\n- Models marked with VL indicate language-vision pretraining (others are vision-only)\n- Models trained on \u003e100K slides may be considered foundation models and are marked in **bold**\n- \\# of WSIs, tiles, and patients are reported to 2 significant figures\n- INE = ImageNet epochs\n- Order is chronological\n- Some of these feature extractors have been evaluated in a benchmarking study for whole slide classification [here](https://arxiv.org/abs/2311.11772).\n- \\*\\* means inferred from other numbers provided in the paper\n\n## Slide-level / patient-level models\nThis table includes models that produce slide-level or patient-level embeddings without supervision.\n| Name                                                                | Group                                                                                                          | Weights            | Released                                                         | SSL                                   | WSIs                         | Patients | Batch size | Iterations       | Architecture             | Parameters | Embed dim | Patch size | Dataset                                                                        | Links                                                                                                                                                                                                                                                                                                                         |\n| ------------------------------------------------------------------- | -------------------------------------------------------------------------------------------------------------- | ------------------ | ---------------------------------------------------------------- | ------------------------------------- | ---------------------------- | -------- | ---------- | ---------------- | ------------------------ | ---------- | --------- | ---------- | ------------------------------------------------------------------------------ | ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |\n| [GigaSSL](https://arxiv.org/abs/2212.03273)                         | [CBIO](https://cbio.mines-paristech.fr)                                                                        | :white_check_mark: | Dec 2022[\\*](https://arxiv.org/abs/2212.03273v1)                 | SimCLR                                | 12K                          |          |            | 1K epochs        | ResNet-18                |            | 256       | 256        | TCGA                                                                           | [\u003cimg src=\"https://raw.githubusercontent.com/FortAwesome/Font-Awesome/6.x/svgs/brands/github.svg\" width=\"20\"\u003e](https://github.com/trislaz/gigassl)                                                                                                                                                                            |\n| [PRISM](https://arxiv.org/abs/2405.10254) (VL)                      | [Paige](https://paige.ai) / [Microsoft](https://www.microsoft.com/en-us/research/lab/microsoft-health-futures) | :white_check_mark: | May 2024[\\*](https://arxiv.org/abs/2405.10254v1)                 | contrastive (with language)           | **590K** (190K text reports) | 190K     | 64 (x4)    | 75K (10 epochs)  | Perceiver + BioGPT       |            | 1280      | 224        | proprietary                                                                    | [\u003cimg src=\"https://huggingface.co/datasets/huggingface/brand-assets/resolve/main/hf-logo.svg\" width=\"25\"\u003e](https://huggingface.co/paige-ai/Prism)                                                                                                                                                                             |\n| [Prov-GigaPath](https://www.nature.com/articles/s41586-024-07441-w) | [Microsoft](https://www.microsoft.com/en-us/research/) / [Providence](https://www.providence.org)              | :white_check_mark: | May 2024[\\*](https://www.nature.com/articles/s41586-024-07441-w) | DINOv2                                | **170K**                     | 30K      |            |                  | LongNet                  | 86M        | 1536      | 224        | proprietary (Providence)                                                       | [\u003cimg src=\"https://raw.githubusercontent.com/FortAwesome/Font-Awesome/6.x/svgs/brands/github.svg\" width=\"20\"\u003e](https://github.com/prov-gigapath/prov-gigapath) [\u003cimg src=\"https://huggingface.co/datasets/huggingface/brand-assets/resolve/main/hf-logo.svg\" width=\"25\"\u003e](https://huggingface.co/prov-gigapath/prov-gigapath) |\n| [MADELEINE](https://arxiv.org/abs/2408.02859) (VL)                  | [Mahmood Lab](https://faisal.ai)                                                                               | :white_check_mark: | Aug 2024[\\*](https://arxiv.org/abs/2408.02859v1)                 | contrastive (InfoNCE \u0026 OT)            | 16K                          | 2K       | 120        | 90 epochs        | multi-head attention MIL |            | 512       | 256        | [ACROBAT](https://acrobat.grand-challenge.org/data/), BWH Kidney (proprietary) | [\u003cimg src=\"https://raw.githubusercontent.com/FortAwesome/Font-Awesome/6.x/svgs/brands/github.svg\" width=\"20\"\u003e](https://github.com/mahmoodlab/MADELEINE)                                                                                                                                                                       |\n| [CHIEF](https://www.nature.com/articles/s41586-024-07894-z) (VL)    | [Yu Lab](https://yulab.hms.harvard.edu)                                                                        | :white_check_mark: | Sep 2024[\\*](https://www.nature.com/articles/s41586-024-07894-z) |                                       |                              |          |            |                  |                          |            |           |            |                                                                                | [\u003cimg src=\"https://raw.githubusercontent.com/FortAwesome/Font-Awesome/6.x/svgs/brands/github.svg\" width=\"20\"\u003e](https://github.com/hms-dbmi/CHIEF)                                                                                                                                                                             |\n| [COBRA](https://arxiv.org/abs/2411.13623)                           | [Kather Lab](https://jnkather.github.io)                                                                       | :white_check_mark: | Nov 2024[\\*](https://arxiv.org/abs/2411.13623v1)                 | COBRA (MoCo-v3 in FM embedding space) | 3K                           | 2.8K     | 1024       | 2K epochs        | Mamba-2 + ABMIL          | 15M        | 768       | 224        | TCGA (BRCA, CRC, LUAD, LUSC, STAD)                                             | [\u003cimg src=\"https://raw.githubusercontent.com/FortAwesome/Font-Awesome/6.x/svgs/brands/github.svg\" width=\"20\"\u003e](https://github.com/KatherLab/COBRA) [\u003cimg src=\"https://huggingface.co/datasets/huggingface/brand-assets/resolve/main/hf-logo.svg\" width=\"25\"\u003e](https://huggingface.co/KatherLab/COBRA)                         |\n| [TITAN](https://arxiv.org/abs/2411.19666)\u003csub\u003eV\u003c/sub\u003e (VL)          | [Mahmood Lab](https://faisal.ai)                                                                               | :white_check_mark: | Dec 2024[\\*](https://arxiv.org/abs/2411.19666v1)                 | iBOT                                  | **340K**                     |          | 1024       | 91K (270 epochs) | ViT (smaller)            | 42M        |           | 224        | Mass-340K (proprietary)                                                        | [\u003cimg src=\"https://raw.githubusercontent.com/FortAwesome/Font-Awesome/6.x/svgs/brands/github.svg\" width=\"20\"\u003e](https://github.com/mahmoodlab/TITAN) [\u003cimg src=\"https://huggingface.co/datasets/huggingface/brand-assets/resolve/main/hf-logo.svg\" width=\"25\"\u003e](https://huggingface.co/MahmoodLab/TITAN)                       |\n| [THREADS](https://arxiv.org/abs/2501.16652) (WSI, RNA, DNA)         | [Mahmood Lab](https://faisal.ai)                                                                               | :x:                | Jan 2025[\\*](https://arxiv.org/abs/2501.16652v1)                 |                                       | 47K                          |          | 1200       | up to 101 epochs | ViT-L                    |            |           | 224        | MBTG-47k (MGH, BWH, TCGA, GTEx)                                                |                                                                                                                                                                                                                                                                                                                               |\n \n","project_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Fgeorg-wolflein%2Fpathology-foundation-models","html_url":"https://awesome.ecosyste.ms/projects/github.com%2Fgeorg-wolflein%2Fpathology-foundation-models","lists_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Fgeorg-wolflein%2Fpathology-foundation-models/lists"}