gradio-app/trackio
A lightweight experiment tracking library from Hugging Face for logging ML training runs and metrics.
Not currently ranked — collecting fresh signals.
star history
Trackio is a local-first experiment tracking library built by Hugging Face, designed for both human researchers and AI agents running autonomous experiments. It stores logs in SQLite databases with optional Parquet export, supports high-throughput parallel experiment tracking, and provides a CLI interface for querying data including direct SQL access. The library includes a Gradio-inspired dashboard for visualizing metrics, media, tables, and alerts.
Frequently asked
- What is gradio-app/trackio?
- A lightweight experiment tracking library from Hugging Face for logging ML training runs and metrics.
- Is trackio open source?
- Yes — gradio-app/trackio is open source, released under the MIT license.
- What language is trackio written in?
- gradio-app/trackio is primarily written in Python.
- How popular is trackio?
- gradio-app/trackio has 1.6k stars on GitHub.
- Where can I find trackio?
- gradio-app/trackio is on GitHub at https://github.com/gradio-app/trackio.