Tracking datasets with DVC in MLOps - Practice Problems & Coding Challenges

Challenge - 5 Problems

💻 Command Output

intermediate

DVC add command output

You run the command dvc add data/raw_images in your project folder. What is the expected output?

MLOps

dvc add data/raw_images

A Warning: data/raw_images is empty, nothing added.

B

Adding data/raw_images to DVC.
Computing checksum...
Adding 'data/raw_images.dvc' file.

C Error: data/raw_images directory not found.

D dvc: command not found

💡 Hint

Think about what happens when you add a directory with files to DVC.

The 'dvc add' command tracks the specified directory or file by computing its checksum and creating a .dvc file to track it. It does not error if the directory exists and has files.

🧠 Conceptual

intermediate

Purpose of .dvc files

What is the main purpose of the .dvc files created when you track datasets with DVC?

A They are temporary cache files created during data processing.

B They store the actual dataset files inside the project folder.

C They are configuration files for DVC remote storage credentials.

D They contain metadata and checksums to track dataset versions without storing data directly.

💡 Hint

Think about how DVC tracks large files without putting them in Git.

DVC uses .dvc files to store metadata like checksums and paths to data, enabling version control without storing large files in Git.

🔀 Workflow

advanced

Correct sequence to track and push dataset with DVC

Arrange the steps in the correct order to track a new dataset folder data/images and push it to remote storage using DVC.

A 1,2,3,4

B 2,1,3,4

C 1,3,2,4

D 3,2,1,4

💡 Hint

Think about adding data, then staging files for Git, committing, then pushing data to remote.

First, you add the data to DVC, then stage the .dvc file and .gitignore for Git, commit the changes, and finally push the data to remote storage.

❓ Troubleshoot

advanced

DVC push error diagnosis

You run dvc push but get the error: ERROR: failed to upload data/images: Access denied to remote storage. What is the most likely cause?

A The .dvc file for data/images is corrupted.

B The dataset folder data/images does not exist locally.

C The remote storage credentials are missing or incorrect.

D Git repository is not initialized.

💡 Hint

Think about what 'Access denied' means when pushing data.

Access denied errors usually mean the remote storage credentials are missing, expired, or incorrect, preventing upload.

✅ Best Practice

expert

Best practice for dataset versioning with DVC

Which practice ensures reliable dataset versioning and collaboration when using DVC in a team?

A Track datasets with DVC, commit .dvc files to Git, and push data to a shared remote storage.

B Commit only the dataset files directly to Git to keep everything in one place.

C Store datasets only on local machines and share via email to avoid remote storage issues.

D Use DVC without Git to simplify version control.

💡 Hint

Think about how to keep data versions consistent and shareable in a team.

The best practice is to track datasets with DVC, commit the small .dvc files to Git, and push large data files to a shared remote storage for collaboration.