{"id":18447404,"url":"https://github.com/paperswithcode/model-index","last_synced_at":"2025-04-08T00:32:07.450Z","repository":{"id":49868850,"uuid":"337802449","full_name":"paperswithcode/model-index","owner":"paperswithcode","description":"Create a source of truth for ML model results and browse it on Papers with Code","archived":false,"fork":false,"pushed_at":"2021-06-09T13:27:50.000Z","size":214,"stargazers_count":26,"open_issues_count":1,"forks_count":3,"subscribers_count":6,"default_branch":"main","last_synced_at":"2025-03-23T03:32:18.067Z","etag":null,"topics":[],"latest_commit_sha":null,"homepage":"","language":"Python","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/paperswithcode.png","metadata":{"files":{"readme":"README.md","changelog":null,"contributing":null,"funding":null,"license":"LICENSE","code_of_conduct":null,"threat_model":null,"audit":null,"citation":null,"codeowners":null,"security":null,"support":null}},"created_at":"2021-02-10T17:38:12.000Z","updated_at":"2024-09-10T02:05:51.000Z","dependencies_parsed_at":"2022-09-01T17:31:58.758Z","dependency_job_id":null,"html_url":"https://github.com/paperswithcode/model-index","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/paperswithcode%2Fmodel-index","tags_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/paperswithcode%2Fmodel-index/tags","releases_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/paperswithcode%2Fmodel-index/releases","manifests_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/paperswithcode%2Fmodel-index/manifests","owner_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners/paperswithcode","download_url":"https://codeload.github.com/paperswithcode/model-index/tar.gz/refs/heads/main","host":{"name":"GitHub","url":"https://github.com","kind":"github","repositories_count":247755440,"owners_count":20990617,"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":"2024-11-06T07:13:16.333Z","updated_at":"2025-04-08T00:32:04.744Z","avatar_url":"https://github.com/paperswithcode.png","language":"Python","funding_links":[],"categories":[],"sub_categories":[],"readme":"# model-index: maintain a source of truth for ML models\n\n\u003cp align=\"center\"\u003e\n\n\u003ca href=\"https://app.circleci.com/pipelines/github/paperswithcode/model-index\"\u003e\n  \u003cimg alt=\"Tests\" src=\"https://img.shields.io/circleci/build/github/paperswithcode/model-index/main\"\u003e\n\u003c/a\u003e\n\n\u003ca href=\"https://pypi.org/project/model-index/\"\u003e\n  \u003cimg alt=\"PyPI\" src=\"https://img.shields.io/pypi/v/model-index\"\u003e\n\u003c/a\u003e\n\n\u003ca href=\"https://model-index.readthedocs.io/en/latest/\"\u003e\n  \u003cimg alt=\"Docs\" src=\"https://img.shields.io/readthedocs/model-index\"\u003e\n\u003c/a\u003e\n\n\u003c/p\u003e\n\n`model-index` has two goals:\n- Make it easy to maintain a source-of-truth index of Machine Learning model metadata \n- Enable the community browse this model metadata on [Papers with Code](https://paperswithcode.com/)\n\nThe main design principle of `model-index` is **flexibility**. You can store your model metadata however is the\nmost convenient for you - as JSONs, YAMLs or as annotations inside markdown. `model-index` provides a convenient\nway to collect all this metadata into a single file that's browsable, searchable and comparable.\n\nYou can use this library locally or choose to upload the metadata to [Papers with Code](https://paperswithcode.com)\nto have your library featured on the website. \n\n## How it works\n\nThere is a root file for the model index: `model-index.yml` that links to (or contains) metadata. \n\n```yaml\nModels:\n  - Name: Inception v3\n    Metadata:\n      FLOPs: 5731284192\n      Parameters: 23834568\n      Training Data: ImageNet  \n      Training Resources: 8x V100 GPUs\n    Results:\n      - Task: Image Classification\n        Dataset: ImageNet\n        Metrics:\n          Top 1 Accuracy: 74.67%\n          Top 5 Accuracy: 92.1%\n    Paper: https://arxiv.org/abs/1512.00567v3\n    Code: https://github.com/rwightman/pytorch-image-models/blob/timm/models/inception_v3.py#L442\n    Weights: https://download.pytorch.org/models/inception_v3_google-1a9a5a14.pth \n    README: docs/inception-v3-readme.md\n```\n\nAll fields except for `Name` are **optional**. You can add any fields you like, but the ones above have a \nstandard meaning across different models and libraries. \n\nWe recommend putting the `model-index.yml` file in the root of your repository (so that relative links such as \n`docs/inception-v3-readme.md` are easier to write), but you can also put it anywhere else in the repository (e.g.\nin your `docs/` or `models/` folder). \n\n### Storing metadata in markdown files\n\nMetadata can also be directly stored in a model's README file. For example in this `docs/rexnet.md` file:\n\n```markdown\n\u003c!--\nType: model-index\nName: RexNet\nMetadata: \n  Epochs: 400\n  Batch Size: 512\nPaper: https://arxiv.org/abs/2007.00992v1\n--\u003e\n\n# Summary\n\nRank Expansion Networks (ReXNets) follow a set of new design \nprinciples for designing bottlenecks in image classification models.\n\n## Usage\n\nimport timm\nm = timm.create_model('rexnet_100', pretrained=True)\nm.eval()\n```\n\nIn this case, you just need to include this markdown file into the global `model-index.yml` file:\n\n```yaml\nModels:\n  - docs/rexnet.md\n```\n\n## Get started\n\nCheck out our [official documentation](https://model-index.readthedocs.io/en/latest/) on how to get started. \n\n## Uploading to Papers with Code\n\nTo feature your library on Papers with Code, get in touch with `hello@paperswithcode.com` and the model index\nof your library will be automatically included into Papers with Code. \n\n\n\n\n\n\n\n\n\n","project_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Fpaperswithcode%2Fmodel-index","html_url":"https://awesome.ecosyste.ms/projects/github.com%2Fpaperswithcode%2Fmodel-index","lists_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Fpaperswithcode%2Fmodel-index/lists"}