This repository provides a standardized JSON format for describing Large Language Model (LLM) files, enabling consistent metadata across different models. It targets developers and researchers working with LLMs, offering a structured way to manage and discover model details like size, architecture, and file formats.

How It Works

The project defines a JSON schema for individual model descriptors, capturing essential information such as model size (e.g., 7B, 13B), architecture (e.g., Llama, MPT), and file formats including quantization details (e.g., ggml, q4_K_M). A GitHub Action automatically merges these individual JSON files into a central catalog.json, with another action ensuring schema validation for all contributions.

Quick Start & Requirements

• To contribute, fork the repository, create a new model JSON file in the models/ directory, and validate it using the provided validate.py script or createCatalog.py.
• Requires Python 3 for validation scripts.
• Contribution guide: https://github.com/lmstudio-ai/model-catalog/blob/main/CONTRIBUTING.md

Highlighted Details

• Standardized JSON schema for LLM file metadata.
• Supports detailed information on model size, architecture, and quantization.
• Automated validation and catalog generation via GitHub Actions.
• Community-driven contribution model.

Maintenance & Community

• Open to community contributions for cataloging models and improving the format.
• Contribution process outlined in CONTRIBUTING.md.

Licensing & Compatibility

• License not specified in the README.

Limitations & Caveats

The README does not specify the license, which may impact commercial use or integration into closed-source projects.