---
id: model-index
version: "0.1.11"
license: MIT
license_treatment: permissive
maintenance: abandoned
---

# model-index — Create a source of truth for ML model results and browse it on Papers with Code

License: permissive · Maintenance: abandoned · Downloads: 234.8K/mo

## What it is and what it does

model-index is a tool for organizing and publishing machine learning model metadata in a standardized, browsable format. It lets you define model information—training parameters, results, dataset details, paper links, and code references—in a single YAML file or spread across markdown files, then optionally upload that metadata to Papers with Code for community discovery.

The package is designed around flexibility: you can store metadata however is convenient (YAML, JSON, or markdown annotations), and model-index collects it into a unified index. All fields except model name are optional, and you can add custom fields beyond the standard ones. The main use case is maintaining a source-of-truth registry for a machine learning library or research project.

Use it for:

- Publish a structured registry of trained models in a research repository or ML library for Papers with Code discovery
- Maintain model metadata (FLOPs, parameters, training resources, results) in version-controlled YAML alongside code
- Embed model documentation directly in markdown files and aggregate it into a searchable index
- Track model performance metrics and results across datasets in a standardized, comparable format
- Link model weights, papers, and code implementations in a single queryable source of truth

## Worth the install?

AI-flagged interpretation of the facts above — verify before relying.

Maintains a structured index of machine learning model metadata (parameters, results, training data, paper links) stored as YAML, JSON, or markdown annotations, with optional integration to Papers with Code.

No. The package is abandoned (no commits since June 2021, no releases since March 2021) and has received no maintenance or security updates in over four years. While install friction is low and the license is permissive, the lack of active maintenance means dependency compatibility cannot be guaranteed, and any bugs or breaking changes in its runtime dependencies will go unfixed. Use only if you have an existing project already relying on it and can maintain a local fork if needed.

## Install

pip install model-index
uv add model-index
poetry add model-index

## Installing model-index

Before you install: Low install friction with four straightforward runtime dependencies. However, the package is abandoned—last release was 0.1.11 in March 2021, with no commits since June 2021. No active maintenance or security updates should be expected.

License in practice: MIT license permits unrestricted use, modification, and distribution with minimal restrictions, making it suitable for both open and proprietary projects.

Quickstart:

pip install model-index

from model_index import ModelIndex
index = ModelIndex.from_file('model-index.yml')

Verify before relying:

- Whether the package still works with current versions of its runtime dependencies (pyyaml, markdown, ordered-set, click)
- Whether Papers with Code still accepts model-index submissions and whether the integration remains functional
- Python version compatibility—requires_python is unspecified in the fact sheet

## Package facts

- License: MIT (permissive)
- Python support: unspecified
- Install friction: low
- Maintenance: abandoned
- Downloads: 234.8K/month (top 15,000 on PyPI)
- Known vulnerabilities: none known

## Tags

ml model metadata index, machine learning model registry, model documentation yaml, papers with code integration, model results tracking, ml model source of truth, model metadata management, ml-metadata, model-registry, abandoned

[View on SkillFed](https://skillfed.io/packages/model-index) · [View on PyPI](https://pypi.org/project/model-index/)