DVC @ Summer of Code

About DVCLive

We have a new project DVCLive, not DVC so that might be better for students. DVCLive is newer (easy to contribute) and more ML specific (it might be a better fit for the students). 

DVCLive is an open-source library for monitoring the progress of metrics during training of machine learning models. It's built with Git and MLOps principles in mind:

• Codification of data 
  Tracked metrics are stored in readable text files that can be versioned by Git or other version control tools.
• Distributed 
  No services or servers are required. Metrics are stored in a Git repository as text files, or pointers to files in DVC storage.
• GitOps API 
  Plots are generated through DVC using Git commit SHAs or branch names, e.g.: dvc plots diff --target logs master.
• Automation
  DVCLive metrics are easy to use by any automation, DevOps, or MLOps tool such as CI/CD (including CML), custom scripts, or ML platforms.

Key Summer of Code task

Help us with the ongoing active issues of our library. We have ML-specific DVCLive with the set of good-first-issue as well as DVC with a set of good-first-issue. We decide down the road what is the best for the ODS Summer of Code participants.

We have already started to mark some issues as good first issues, and in case of dvclive all marked as such should be fine to take by newcomers. 

We will be on the lookout for this kind of issues, and will also be going through dvc's open issues and marking them too.

Call for integration collaboration 

Today only Python is supported (while DVC is language agnostic), with a minimum number of integrations with ML frameworks:

• Tensorflow/Keras
• XGBoost
• MMCV

The DVCLive team is happy to extend the functionality as needed. Please create an issue to start a discussion!

Contacts and other useful materials 

• ODS slack channel #tool_dvc
• Library GitHub 
• Good first issues