Show HN: Reproducible experiments without committing every tweak

Category: cli-tool

Tags: reproducibility, data-versioning, experiment-tracking

Score: 6.0/10 (Innovation: 5, Technical: 6, Documentation: 7, Utility: 6)

Rair is a CLI tool that tracks code diffs, git commits, inputs, outputs, and metadata for reproducible experiments without cluttering git history. It is interesting because it solves the common pain point of tweaking parameters and tracking results in research workflows with automatic file discovery and minimal configuration.

Target audience: data scientists, ML researchers, computational scientists

Repository: https://github.com/DLR-Institute-of-Future-Fuels/rair · Python · MIT · 2 stars

View on Hacker News