Trackio Tutorial: Hugging Face's new, FREE experiment tracking library

2026-03-27 · Source: HuggingFace · Field: Technology & Digital — Artificial Intelligence & Machine Learning, Software Development & Engineering · Depth: Intermediate, medium

Summary

Track You is an open-source Python package designed for local-first machine learning experiment tracking, offering an API largely identical to Weights & Biases for seamless integration. It allows users to initialize projects, log real-time metrics, images, audio, and tables, and visualize results via a local SQLite-backed dashboard launched with `trackyou show`. A key feature is its alert system, which can trigger notifications in the UI, CLI, Slack, or Discord when metrics deviate. While primarily local, Track You supports remote logging by streaming data to Hugging Face Spaces, enabling sharing and embedding. The tool also provides extensive CLI capabilities for programmatic access to metrics and snapshots, which can be integrated with AI agents via a dedicated "skill" to autonomously analyze training runs, identify issues like gradient explosions, and suggest hyperparameter adjustments for auto-research workflows.

Key takeaway

For AI Engineers or ML Scientists managing numerous experiments, Track You offers a robust, open-source, and local-first solution for tracking and analysis. Its Weights & Biases-compatible API simplifies adoption, while the AI agent integration allows for autonomous debugging and hyperparameter optimization. Consider integrating Track You skills into your auto-research agents to automate the identification of training issues like gradient explosions and receive actionable recommendations, significantly streamlining your experimental workflows.

Key insights

Track You offers local-first, open-source ML experiment tracking with W&B-compatible API and AI agent integration.

Principles

• Local-first logging ensures speed and privacy.
• API compatibility simplifies migration and adoption.
• AI agent integration enables autonomous experiment analysis.

Method

Initialize a Track You project, log metrics during training using a W&B-like API, and visualize results with `trackyou show`. For remote logging, specify a Hugging Face Space ID. Integrate with AI agents by installing the Track You skill.

In practice

• Use `trackyou.init` to start a new experiment run.
• Employ `trackyou.log` to record metrics in real-time.
• Install the Track You skill for AI agent-driven analysis.

Topics

• Experiment Tracking
• Hugging Face Spaces
• AI Agents
• Weights & Biases API
• CLI Tools

Best for: Machine Learning Engineer, AI Engineer, AI Scientist

Related on AIssential

• ML Experiment Tracking in H2O Driverless AI | Part 10 — Effective experiment tracking is crucial for data science, enabling parallel runs, reproducibility, real-time monitoring, and comparison.
• What is Hugging Face? - Models, Datasets & Spaces — Hugging Face provides a unified platform for ML models, datasets, and application deployment.
• W&B Models: Track all ML experiments — Weights & Biases automates ML experiment tracking and visualization across diverse frameworks.
• Migrating Your GitHub CI to Hugging Face Jobs — Migrating GitHub CI to Hugging Face Jobs cuts CPU test times by 30% and enables cost-effective GPU testing via custom runners and hardware.
• [P] We made GoodSeed, a pleasant ML experiment tracker — GoodSeed is an open-source ML experiment tracker with a clean UI and Neptune migration capabilities.
• MLflow. Stop losing your best experiments: companion notebook — The notebook provides a complete MLflow workflow for experiment tracking, model registration, and deployment.

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by HuggingFace.