{"id":38545627,"url":"https://github.com/graphcore-research/minimol","last_synced_at":"2026-01-17T07:19:08.998Z","repository":{"id":248656568,"uuid":"826706460","full_name":"graphcore-research/minimol","owner":"graphcore-research","description":"MiniMol is a 10M-parameters molecular fingerprinting model pre-trained on \u003e3300 biological and quantum tasks ","archived":false,"fork":false,"pushed_at":"2025-05-29T14:21:46.000Z","size":6397,"stargazers_count":25,"open_issues_count":2,"forks_count":6,"subscribers_count":5,"default_branch":"main","last_synced_at":"2025-09-30T04:54:17.115Z","etag":null,"topics":[],"latest_commit_sha":null,"homepage":"","language":"Jupyter Notebook","has_issues":true,"has_wiki":null,"has_pages":null,"mirror_url":null,"source_name":null,"license":"mit","status":null,"scm":"git","pull_requests_enabled":true,"icon_url":"https://github.com/graphcore-research.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":"2024-07-10T08:12:53.000Z","updated_at":"2025-09-19T14:56:18.000Z","dependencies_parsed_at":"2024-08-22T11:46:11.065Z","dependency_job_id":"5dd2acc3-54bb-4c3b-885b-51c182e283e7","html_url":"https://github.com/graphcore-research/minimol","commit_stats":null,"previous_names":["graphcore-research/minimol"],"tags_count":0,"template":false,"template_full_name":null,"purl":"pkg:github/graphcore-research/minimol","repository_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/graphcore-research%2Fminimol","tags_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/graphcore-research%2Fminimol/tags","releases_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/graphcore-research%2Fminimol/releases","manifests_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/graphcore-research%2Fminimol/manifests","owner_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners/graphcore-research","download_url":"https://codeload.github.com/graphcore-research/minimol/tar.gz/refs/heads/main","sbom_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/graphcore-research%2Fminimol/sbom","scorecard":null,"host":{"name":"GitHub","url":"https://github.com","kind":"github","repositories_count":286080680,"owners_count":28503467,"icon_url":"https://github.com/github.png","version":null,"created_at":"2022-05-30T11:31:42.601Z","updated_at":"2026-01-17T06:57:29.758Z","status":"ssl_error","status_checked_at":"2026-01-17T06:56:03.931Z","response_time":85,"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":[],"created_at":"2026-01-17T07:19:08.935Z","updated_at":"2026-01-17T07:19:08.990Z","avatar_url":"https://github.com/graphcore-research.png","language":"Jupyter Notebook","funding_links":[],"categories":[],"sub_categories":[],"readme":"![Minimol architecture](figs/minimol-architecture.png)\n\nA parameter-efficient molecular featuriser that generalises well to biological tasks thanks to the effective pre-training on biological and quantum mechnical datasets.\n\nThe model has been introduced in the paper [𝙼𝚒𝚗𝚒𝙼𝚘𝚕: A Parameter-Efficient Foundation Model for Molecular Learning](https://arxiv.org/abs/2404.14986), published in the ICML workshop on *Accessible and Efficient Foundation Models for Biological Discovery* in 2024.\n\n## Usage\n\nEmbeddings can be generated in four lines of code:\n\n```\nfrom minimol import Minimol\nmodel = Minimol()\nsmiles = [\n    'COc1ccc2cc(C(=O)NC3(C(=O)N[C@H](Cc4ccccc4)C(=O)NCC4CCN(CC5CCOCC5)CC4)CCCC3)sc2c1',\n    'Nc1nc(=O)c2c([nH]1)NCC(CNc1ccc(C(=O)NC(CCC(=O)O)C(=O)O)cc1)N2C=O',\n    'O=C1CCCN1CCCCN1CCN(c2cc(C(F)(F)F)ccn2)CC1',\n    'c1ccc(-c2cccnc2)cc1',\n]\nmodel(smiles)\n\u003e\u003e A list of 4 tensors of (512,) shape\n```\n\nFor a Colab notebook showing how to use Minimol's fingerprints to achieve SoTA results on a downstream task, click here: [![Open In Colab](https://colab.research.google.com/assets/colab-badge.svg)](https://githubtocolab.com/graphcore-research/minimol/blob/main/notebooks/downstream_adaptation.ipynb)\n\n## Installation\n\n### Pip\nWhen used with cuda, use `nvcc --version` to see which version of the driver is installed on your machine, to select the wheel (cuXXX):\n```\npip install torch-sparse torch-cluster torch-scatter -f https://pytorch-geometric.com/whl/torch-2.3.0+cu124.html\npip install minimol\n```\n\n### Local\n``` \ngit clone git@github.com:graphcore-research/minimol.git \ncd minimol\nmamba env create -f env.yml -n minimol_venv\nmamba activate minimol\n```\n*To install mamba see [the official documentation](https://mamba.readthedocs.io/en/latest/installation/mamba-installation.html).*\n\n## Performance\n\nThe model has been evaluated on 22 benchmarks from the ADMET group of [Therapeutics Data Commons (TDC)](https://tdcommons.ai). These are the results when comparing to [MolE](https://arxiv.org/abs/2211.02657) and TOP5 models from the TDC leaderboard (as of June 2024):\n\n| TDC Dataset          |          |            | TDC Leaderboard | MolE           |          | MiniMol (GINE)|          |\n|----------------------|----------|------------|-----------------|----------------|----------|---------------|----------|\n| **Name**             | **Size** | **Metric** | **SoTA Result** | **Result**     | **Rank** | **Result**    | **Rank** |\n| **Absorption**       |          |            |                 |                |          |               |          |\n| Caco2 Wang           | 906      | MAE        | 0.276 ± 0.005   | 0.310 ± 0.010  | 6        | 0.350 ± 0.018 | 7        |\n| Bioavailability Ma   | 640      | AUROC      | 0.748 ± 0.033   | 0.654 ± 0.028  | 7        | 0.689 ± 0.020 | 5        |\n| Lipophilicity AZ     | 4,200    | MAE        | 0.467 ± 0.006   | 0.469 ± 0.009  | 3        | 0.456 ± 0.008 | 1        |\n| Solubility AqSolDB   | 9,982    | MAE        | 0.761 ± 0.025   | 0.792 ± 0.005  | 5        | 0.741 ± 0.013 | 1        |\n| HIA Hou              | 578      | AUROC      | 0.989 ± 0.001   | 0.963 ± 0.019  | 7        | 0.993 ± 0.005 | 1        |\n| Pgp Broccatelli      | 1,212    | AUROC      | 0.938 ± 0.002   | 0.915 ± 0.005  | 7        | 0.942 ± 0.002 | 1        |\n| **Distribution**     |          |            |                 |                |          |               |          |\n| BBB Martins          | 1,975    | AUROC      | 0.916 ± 0.001   | 0.903 ± 0.005  | 7        | 0.924 ± 0.003 | 1        |\n| PPBR AZ              | 1,797    | MAE        | 7.526 ± 0.106   | 8.073 ± 0.335  | 6        | 7.696 ± 0.125 | 4        |\n| VDss Lombardo        | 1,130    | Spearman   | 0.713 ± 0.007   | 0.654 ± 0.031  | 3        | 0.535 ± 0.027 | 7        |\n| **Metabolism**       |          |            |                 |                |          |               |          |\n| CYP2C9 Veith         | 12,092   | AUPRC      | 0.859 ± 0.001   | 0.801 ± 0.003  | 5        | 0.823 ± 0.006 | 4        |\n| CYP2D6 Veith         | 13,130   | AUPRC      | 0.790 ± 0.001   | 0.682 ± 0.008  | 6        | 0.719 ± 0.004 | 5        |\n| CYP3A4 Veith         | 12,328   | AUPRC      | 0.916 ± 0.000   | 0.867 ± 0.003  | 7        | 0.877 ± 0.001 | 4        |\n| CYP2C9 Substrate     | 666      | AUPRC      | 0.441 ± 0.033   | 0.446 ± 0.062  | 2        | 0.474 ± 0.025 | 1        |\n| CYP2D6 Substrate     | 664      | AUPRC      | 0.736 ± 0.024   | 0.699 ± 0.018  | 7        | 0.695 ± 0.032 | 6        |\n| CYP3A4 Substrate     | 667      | AUROC      | 0.662 ± 0.031   | 0.670 ± 0.018  | 1        | 0.663 ± 0.008 | 2        |\n| **Excretion**        |          |            |                 |                |          |               |          |\n| Half Life Obach      | 667      | Spearman   | 0.562 ± 0.008   | 0.549 ± 0.024  | 4        | 0.495 ± 0.042 | 6        |\n| Clearance Hepatocyte | 1,102    | Spearman   | 0.498 ± 0.009   | 0.381 ± 0.038  | 7        | 0.446 ± 0.029 | 3        |\n| Clearance Microsome  | 1,020    | Spearman   | 0.630 ± 0.010   | 0.607 ± 0.027  | 6        | 0.628 ± 0.005 | 2        |\n| **Toxicity**         |          |            |                 |                |          |               |          |\n| LD50 Zhu             | 7,385    | MAE        | 0.552 ± 0.009   | 0.823 ± 0.019  | 7        | 0.585 ± 0.005 | 2        |\n| hERG                 | 648      | AUROC      | 0.880 ± 0.002   | 0.813 ± 0.009  | 7        | 0.846 ± 0.016 | 4        |\n| Ames                 | 7,255    | AUROC      | 0.871 ± 0.002   | 0.883 ± 0.005  | 1        | 0.849 ± 0.004 | 5        |\n| DILI                 | 475      | AUROC      | 0.925 ± 0.005   | 0.577 ± 0.021  | 7        | 0.956 ± 0.006 | 1        |\n|                      |          |            |                 | **Mean Rank:** | 5.2      |               | 3.3      |\n\n## License\n\nCopyright (c) 2024 Graphcore Ltd. Licensed under the MIT License.\n\nThe included code is released under the MIT license (see [details of the license](LICENSE)).\n","project_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Fgraphcore-research%2Fminimol","html_url":"https://awesome.ecosyste.ms/projects/github.com%2Fgraphcore-research%2Fminimol","lists_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Fgraphcore-research%2Fminimol/lists"}