Tracking datasets with DVC in MLOps - Cheat Sheet & Quick Revision

Recall & Review

beginner

What is DVC in the context of dataset tracking?

DVC (Data Version Control) is a tool that helps track, version, and manage datasets and machine learning models, similar to how Git manages code.

intermediate

How does DVC track large datasets without storing them directly in Git?

DVC stores dataset metadata and pointers in Git, while the actual large data files are stored separately in remote storage like cloud or local cache.

beginner

Which command initializes DVC in a project?

The command is dvc init. It sets up DVC configuration files and folders in your project.

beginner

What is the purpose of the dvc add command?

It tells DVC to start tracking a dataset or file. It creates a .dvc file that tracks the file's version and location.

intermediate

How do you share datasets tracked by DVC with your team?

You push the dataset files to a remote storage using dvc push and share the Git repository with the .dvc files. Team members then use dvc pull to download the data.

What does the dvc add command do?

A Initializes a new Git repository

B Uploads data to remote storage

C Removes a dataset from tracking

D Starts tracking a file or dataset with DVC

dvc add tells DVC to track a file or dataset by creating a metadata file.

Where does DVC store large dataset files by default?

A Inside the Git repository

B In the .gitignore file

C In a separate cache or remote storage

D In the project’s source code folder

DVC keeps large files outside Git, in a cache or remote storage, to avoid bloating the Git repo.

Which command uploads tracked data files to remote storage?

A dvc push

B dvc add

C dvc pull

D dvc init

dvc push uploads data files to the configured remote storage.

What file does DVC create to track a dataset after dvc add?

A A .dvc file

B A .gitignore file

C A README.md file

D A config.yaml file

DVC creates a .dvc file that stores metadata about the tracked dataset.

How do team members get the dataset tracked by DVC after cloning the repo?

A Run <code>git pull</code> only

B Run <code>dvc pull</code> to download data files

C Copy files manually

D Run <code>dvc add</code> again

After cloning, team members run dvc pull to fetch the actual data files from remote storage.

Explain how DVC helps manage datasets in machine learning projects.

Think about how DVC separates data from code and helps teams work together.

Describe the steps to start tracking a dataset with DVC and share it with your team.

Focus on commands and the flow from local tracking to sharing.