IDSIA/sacred

Health v2 · maintenance only

Adoption not part of the score

4374 stars · 393 forks observed · 2026-08-28

What it is AI-extracted, prompt v1, taxonomy v1, 2026-08-29, confidence not recorded

Sacred is a Python library for configuring, organizing, logging, and reproducing machine learning and scientific experiments. It provides config scopes, dependency injection, a per-experiment CLI, observers that save run metadata to databases like MongoDB, and automatic random seeding.

Use cases

• track parameters of machine learning experiments
• reproduce experiment results with fixed seeds
• log experiment runs to a MongoDB database
• run experiments with different settings from the command line
• record dependencies and machine info for each run
• organize and compare results of many experiment variants

When to choose

• you run many Python experiments and need reproducibility and run tracking
• you want automatic config logging and CLI parameter overrides with minimal boilerplate
• you prefer a lightweight experiment tracker without a hosted service

When to avoid

• you need a full-featured experiment dashboard with visualizations like W&B or MLflow
• your project is not in Python
• you need distributed training orchestration rather than experiment bookkeeping

Facets

library · maturity maintenance

configuration-management logging monitoring cli developer-tools machine-learning data-science developer-tools python cli cross-platform experiment-tracking reproducibility machine-learning-infrastructure observers mongodb automatic-seeding automation

Member repositories

For agents

markdown · JSON · MCP: product_card(name="IDSIA/sacred")