DVC

DVC is a data version control tool for individual data scientists and small machine learning projects. It links datasets and model artifacts to code history so you can recover the inputs behind an experiment and reproduce its results.

• Data and model versions: Keep large artifacts outside Git while recording their version information alongside project code.
• Local experiment tracking: Compare data, code, parameters, models and performance plots within your Git repository, without a separate tracking server.
• Reproducible pipelines: Record how code and data produce artifacts, version those definitions with Git, and rerun only the steps affected by a change.
• Experiment sharing: Share experiments through existing Git hosting, including GitHub and GitLab, and reproduce another person's experiment.

DVC is free and open source under Apache 2.0, with support for macOS, Windows and Linux. Data and model files can stay in a local cache or use remote storage such as Amazon S3, Azure, Google Cloud Storage, Google Drive or SSH storage.

The command-line tool handles versioning and pipelines. A VS Code extension provides experiment tracking and data management inside the editor and requires the core DVC tool.