Hugging Face has released Trackio, a lightweight open-source experiment tracking library designed to be easy to integrate and transparent. Built as a drop-in replacement for Weights & Biases (wandb), Trackio offers local dashboards by default and seamless syncing with Hugging Face Spaces for sharing and collaboration.

Background and Context

Experiment tracking is a crucial part of machine learning workflows, but existing tools often come with drawbacks such as complex setup, proprietary APIs, paywalls, or heavy overhead. Trackio aims to address these issues by providing a lightweight, open-source, and free experiment tracking Python library.

The Hugging Face Science team switched to Trackio after finding that existing solutions didn't fully align with their needs. Key benefits they've experienced include effortless sharing and embedding of experiment dashboards via Hugging Face Spaces or embed plots in blogs and documentation—no logins or complex dashboards required. Additionally, Trackio allows for direct tracking of GPU energy usage via nvidia-smi integration, making it easier to quantify environmental impacts and include sustainability metrics in model cards.

Trackio's design makes it easy to customize and experiment with logging features without compromising training performance. The library is under 1,000 lines of code, making it hackable and extensible. Logs persist locally in SQLite and are automatically backed up to Parquet datasets on Hugging Face every five minutes when synced.

Why It Matters to the Industry

Trackio's emphasis on reproducibility and accessibility provides researchers with a straightforward way to log and share experiments without relying on proprietary services. This is particularly important in the adult industry, where experiment tracking can be critical for understanding performance and iterating effectively.

The library's local-first design allows logs and dashboards to run and persist locally by default, with the option to host on Hugging Face Spaces. This makes it easy to share progress seamlessly with teammates or embed plots in documentation without requiring complex setup or logins.

Trackio also integrates with Hugging Face libraries such as transformers and accelerate, enabling minimal-setup logging for training runs. This integration is particularly useful for adult industry platforms and operators who require seamless tracking of metrics and parameters during training.

Getting Started with Trackio

Installing Trackio is as simple as running a pip install command:

  • pip install trackio
  • uv pip install trackio

Since Trackio is API-compatible with WandB, you can start using it instantly by swapping imports:

  • - import wandb
  • + import trackio as wandb

What Comes Next?

Hugging Face has released Trackio in beta, and the library is currently being tested by researchers and developers. The team encourages feedback and contributions to make Trackio an even more effective experiment tracking tool.

Key Facts

  • Trackio is a lightweight open-source experiment tracking library designed for ease of integration and transparency.
  • Built as a drop-in replacement for Weights & Biases (wandb), Trackio offers local dashboards by default and seamless syncing with Hugging Face Spaces.
  • Trackio's design makes it easy to customize and experiment with logging features without compromising training performance.
  • The library is under 1,000 lines of code, making it hackable and extensible.
  • Logs persist locally in SQLite and are automatically backed up to Parquet datasets on Hugging Face every five minutes when synced.