DVC

DVC is an open source version control system for data science and machine learning projects. It works like Git, but it tracks large data files and models that Git cannot store well.

DVC stores pointers to data and model files in your Git repository, while the files themselves live in remote storage. This lets a team roll back to an earlier dataset or model with the same commands it uses for code. It also helps organize experiments, so you can compare results across runs. A VS Code extension brings these steps into the editor. For larger data infrastructure, the vendor points to a separate product called lakeFS.

• Version large data files alongside Git code
• Track experiments and compare results
• VS Code extension for DVC
• Git-like workflow for data and models
• Larger data infrastructure through lakeFS

DVC suits data scientists and ML engineers who work on small to medium projects and want their data versioned like code. Enterprises with petabyte-scale data may look at the separate lakeFS product instead.

DVC and its VS Code extension are free and open source, and the project is licensed under Apache-2.0. You still pay for the storage where your data lives. Setting up remote storage and team access takes some planning, so start with a small test project before you move a large dataset.