Getting Started with Data Version Control (DVC)

Dheeraj Bhat 13 Jun, 2023 • 4 min read


If you are reading this blog, you might have been familiar with what Git is and how it has been an integral part of software development. Similarly, Data Version Control (DVC) is an open-source, Git-based version management for Machine Learning development that instills best practices across the teams. A system called data version control manages and tracks changes to data and machine learning models in a collaborative and reproducible manner. It draws inspiration from version control systems used in software development, such as Git, but tailors specifically to data science projects.

Learning Objectives

In this article you will develop basic understanding of:

  • What is Git?
  • What is Data Version Control?
  • Understand the basics of Data Version Control

This article was published as a part of the Data Science Blogathon.

Advantages of Data Version Control (DVC)

ML Project Version Control

DVC lets you connect with storage providers like AWS S3, Microsoft Azure Blob Storage, Google Drive, Google Cloud Storage, HDFS, etc., to store ML models and datasets.

ML Experiment Management

It helps in easy navigation for automatic metric tracking.

Deployment and Collaboration

DVC introduces pipelines that help in the easy bundling of ML models, data, and code into production, remote machines, or a colleague’s computer.

 Source: dvc.orgNaNSource:</figcaption>
<h2>Learning Objectives</h2>
<p>With this article, you will learn the following:</p>
<li>Understanding the basics of DVC</li>
<li>How DVC can help in variety of problems?</li>
<li>Installing and using DVC in a git repository</li>
<li>Configuring DVC for GDrive remote storage</li>
<li>How to use DVC Pipelines for reproducing workflows?</li>
<h2>Use cases of DVC</h2>
<figure class=
 Source: dvc.orgNaNSource:</figcaption>
<p>The use cases of DVC are as follows:</p>
<li><b>Versioning Data and Models:</b> We can track versions of data and ML models using git commits. A metafile with .dvc extension is created for the data/models that need to be tracked by dvc which contains the metadata information like md5 hash, size, number of files, and the path.</li>
<li><b>CI/CD for Machine Learning: </b>DVC helps in managing data/models and reproducible pipelines</li>
<li>Fast and Secure Data Caching Hub: DVC’s built-in data caching speeds up data transfers and lets us set up a shared DVC cache that prevents repetitive transfers by linking working files and directories</li>
<li><b>Experiment Tracking:</b> Running DVC Experiments in your workspace captures relevant changes automatically (input data, source code, hyperparameters, artifacts, etc.). This helps to iterate quickly on experiments, creating checkpoints, and comparing results.</li>
<li><b>Model Registry:</b> DVC enables us to catalog ML models and versions. This helps to organize model versions from different sources, sharing metadata, and deploying specific models on dev, test, and production environments.</li>
<li><b>Data Registry:</b> DVC enables cross-project reusability of data artifacts i.e. different projects can depend on different repositories.</li>
<p>You can install dvc from <a href=

PyPi repository using the following command line:

pip install dvc

Depending on the type of remote storage that will be used, we have to install optional dependencies: [s3], [gdrive], [gs], [azure], [ssh], [hdfs], [webdav], [oss]. Use [all] to include them all. In this blog, we will be using google drive as remote storage, so pip install dvc[gdrive] for installing gdrive dependencies.

Learn More: Tracking ML Experiments With Data Version Control

Getting Started

In this blog, we will see how to use dvc for tracking data and ml models with gdrive as remote storage. Imagine the Git repository which contains the following structure:

 Folder StructureNaNFolder Structure</figcaption>
<p>The data and models folder will be very huge when it's compared with the source code of the repository. This is where DVC comes into the picture which helps to track data and models folder. Go to the root of the Git repository (a repository that includes data, ml models folders) and initialize dvc using the command:</p>
<pre><code>dvc init</code></pre>
<p>To start tracking data and models directory, run the following command:</p>
<pre><code>dvc add data
dvc add models</code></pre>
<p>Now, this creates a special file with a .dvc extension (data.dvc and models.dvc). This .dvc file contains metadata information like md5 hash, size, number of files, and the path. These .dvc files are versioned with source code with Git. The dvc add command will also add data and models folder to the .gitignore file. Then, we need to commit the changes to git using the following command:</p>
<pre><code>git add -A
git commit -m

Gdrive Remote Configuration

Now, we need to configure gdrive remote storage. Go to your google drive and create a folder called dvc_storage in it. Open the folder dvc_storage. Get the folder-id of the dvc_storage folder from the URL:

# example:

Now, use the following command to use the dvc_storage folder created in the google drive as remote storage:

dvc remote add myremote gdrive://folder-id

# example: dvc remote add myremote gdrive://0AIac4JZqHhKmUk9PDA

Now, we need to commit the changes to git repository by using the command:

git add -A
git commit -m "configure dvc remote storage"

To push the data to remote storage, we use the following command:

dvc push

Then, we push the changes to git using the command:

git push

To pull data from dvc, we can use the following command:

dvc pull

DVC Pipelines

We can make use of DVC pipelines to reproduce the workflows in our repository. The main advantage of this is that we can go back to a particular point in time and run the pipeline to reproduce the same result that we had achieved during the previous time. There are different stages in the DVC pipeline like prepare, train, and evaluate, with each of them performing different tasks. The DVC pipeline is nothing but a DAG (Directed Acyclic Graph). In this DAG graph, there are nodes and edges, with nodes representing the stages and edges representing the direct dependencies. The pipeline is defined in a YAML file (dvc.yaml). A simple dvc.yaml file is as follows:

    cmd: source src/
      - src/
      - data/raw
      - data/clean.csv
    cmd: python src/ data/model.csv
      - src/
      - data/clean.csv
      - data/predict.dat
    cmd: python src/ data/predict.dat
      - src/
      - data/predict.dat

Use the prepare stage to run the data cleaning and pre-processing steps. Use the train stage to train the machine learning model using the data from the prepare stage. The evaluate stage uses the trained model and predictions to provide different plots and metrics.


This blog helps you with the basics of Data Version Control and set up dvc using google drive as remote storage. For advanced uses (like CI/CD etc.), we need to set up DVC remote configuration using the Google Cloud project (click here). There are also other storage types supported like AWS S3, Microsoft Azure Blob Storage, self-hosted SSH servers, HDFS, HTTP, etc. DVC has most of the commands analogous to git (like dvc fetch, dvc checkout, and dvc status, etc, and a lot more). It also has Visual Studio Extension which makes things easier for developers using VS Code. Check out their GitHub repository to learn more about DVC and everything it offers.

Key Takeaways:

  • Understanding the basics of DVC
  • Become acquainted with the use cases of DVC
  • Installation and use of DVC in a git repository
  • GDrive Remote configuration in DVC


The media shown in this article is not owned by Analytics Vidhya and is used at the Author’s discretion.

Frequently Asked Questions

Q1. What is the DVC command?

A. The DVC command is a command-line tool that provides various functionalities for interacting with DVC projects. It includes commands for initializing a DVC project, tracking data files, managing data pipelines, running experiments, and collaborating with other team members. It serves as the primary interface for interacting with DVC’s features.

Q2. How does DVC work?

A. DVC (Data Version Control) provides a layer of version control specifically for data and machine learning models. It tracks changes to data files, dependencies, and experiments while storing them separately from the codebase, allowing for reproducibility and efficient collaboration.

Q3. What is DVC used for?

A. DVC is used for managing and versioning large datasets, machine learning models, and experiments. It helps streamline the data pipeline, enables reproducibility, and facilitates collaboration among data scientists and machine learning engineers.

Q4. Why use DVC instead of Git?

A. DVC complements Git by focusing on versioning and managing data and machine learning models, while Git primarily handles source code. DVC’s dedicated functionality for data and models includes handling large files efficiently, storing data separately, and enabling reproducibility, which are essential for machine learning projects.

Dheeraj Bhat 13 Jun 2023

Frequently Asked Questions

Lorem ipsum dolor sit amet, consectetur adipiscing elit,

Responses From Readers

Related Courses