vlmrun-hub  by vlm-run

Pydantic schema hub for VLM-driven structured data extraction

created 8 months ago
528 stars

Top 60.7% on sourcepulse

GitHubView on GitHub
Project Summary

This project provides a curated catalog of Pydantic schemas designed for extracting structured data from visual inputs using Vision Language Models (VLMs). It targets developers and researchers needing to automate visual data extraction, offering pre-validated, industry-specific schemas to streamline integration and ensure data quality.

How It Works

The hub leverages Pydantic's robust data validation and type-hinting capabilities to define schemas for various visual data types like invoices, driver's licenses, and product information. These schemas act as structured output targets for VLMs, enabling them to return data in a predictable, machine-readable format. This approach simplifies complex parsing and validation, making VLM outputs directly usable in downstream applications and workflows.

Quick Start & Requirements

  • Install via pip: pip install vlmrun-hub
  • Requires Python 3.7+
  • Examples demonstrate usage with OpenAI's API, Instructor, and Ollama.
  • Schema Showcase Notebook: [Link to Notebook]

Highlighted Details

  • Extensive catalog of industry-specific schemas (document processing, finance, healthcare, retail, media, etc.).
  • Model-agnostic design, allowing use with multiple VLM providers.
  • Built-in data validation via Pydantic ensures output accuracy.
  • Optimized for Visual ETL (Extract, Transform, Load) workflows.

Maintenance & Community

  • Community contributions are encouraged via CONTRIBUTING and SCHEMA-GUIDELINES.md.
  • Support and community engagement available via Discord and email (support@vlm.run).
  • Active presence on Twitter and LinkedIn.

Licensing & Compatibility

  • The project appears to be under a permissive license, but specific license details are not explicitly stated in the README. Compatibility for commercial use is likely, but requires verification of the actual license file.

Limitations & Caveats

The README does not explicitly state the license type, which could be a blocker for some users. While it mentions qualitative results and benchmarks, direct performance metrics for the schemas themselves are not provided.

Health Check
Last commit

2 months ago

Responsiveness

1 week

Pull Requests (30d)
0
Issues (30d)
0
Star History
25 stars in the last 90 days

Explore Similar Projects

Feedback? Help us improve.