Data Science Version Control or DVC is an open-source tool for data science and machine learning projects. With a simple and flexible Git-like architecture and interface it helps data scientists:
- manage machine learning models - versioning, including data sets and transformations (scripts) that were used to generate models;
- make projects reproducible;
- make projects shareable;
- manage experiments with branching and metrics tracking;
It aims to replace tools like Excel and Docs that are being commonly used as a knowledge repo and a ledger for the team, ad-hoc scripts to track and move deploy different model versions, ad-hoc data file suffixes and prefixes.
How DVC works
DVC is compatible with Git for storing code and the dependency graph (DAG), but not data files cache. To store and share data files cache DVC supports remotes - any cloud (S3, Azure, Google Cloud, etc) or any on-premise network storage (via SSH, for example).
Please read Get Started for the full version. Common workflow commands include:
|Track code and data together||
|Connect code and data by commands||
|Make changes and reproduce||
|Share data and ML models||
There are three options to install DVC:
pip, Homebrew, or an OS-specific package:
pip install dvc
pip install git+git://github.com/iterative/dvc
brew install iterative/homebrew-dvc/dvc
brew cask install iterative/homebrew-dvc/dvc
Self-contained packages for Windows, Linux, Mac are available. The latest version of the packages can be found at GitHub releases page.
Ubuntu / Debian (apt)
sudo wget https://dvc.org/deb/dvc.list -O /etc/apt/sources.list.d/dvc.list sudo apt-get update sudo apt-get install dvc
Fedora / CentOS (rpm)
sudo wget https://dvc.org/rpm/dvc.repo -O /etc/yum.repos.d/dvc.repo sudo yum update sudo yum install dvc
Arch linux (AUR)
Unofficial package, any inquiries regarding the AUR package, refer to the maintainer.
yay -S dvc
- Git-annex - DVC uses the idea of storing the content of large files (that you don't want to see in your Git repository) in a local key-value store and uses file hardlinks/symlinks instead of the copying actual files.
- Git-LFS - DVC is compatible with any remote storage (S3, Google Cloud, Azure, SSH, etc). DVC utilizes reflinks or hardlinks to avoid copy operation on checkouts which makes much more efficient for large data files.
- Makefile (and its analogues). DVC tracks dependencies (DAG).
- Workflow Management Systems. DVC is a workflow management system designed specifically to manage machine learning experiments. DVC is built on top of Git.
- DAGsHub Is a Github equivalent for DVC - pushing your Git+DVC based repo to DAGsHub will give you a high level dashboard of your project, including DVC pipeline and metrics visualizations, as well as links to DVC managed files if they are in cloud storage.
Contributions are welcome! Please see our Contributing Guide for more details.
Want to stay up to date? Want to help improve DVC by participating in our occasional polls? Subscribe to our mailing list. No spam, really low traffic.
This project is distributed under the Apache license version 2.0 (see the LICENSE file in the project root).
By submitting a pull request for this project, you agree to license your contribution under the Apache license version 2.0 to this project.