Version Control

TECHNICAL DEFINITION

Version control is a system that manages changes to documents, code, and other information over time, enabling multiple collaborators to work concurrently, track revisions, merge changes, and revert to previous states, essential for collaborative software and ML development.

BACKGROUND

Generative artificial intelligence (GenAI) is a subfield of artificial intelligence (AI) that uses generative models to generate text, images, videos, audio, software code or other forms of data. These models learn the underlying patterns and structures of their training data, and use them to generate new data in response to input, which often takes the form of natural language prompts.

SYNONYMS & ALIASES

• Source control
• Revision control
• VCS
• Code versioning

USAGE NOTE

Version control is fundamental for managing code, data, and model artifacts in MLOps workflows.

DEVELOPERS

Organizations developing technology related to Version Control.

• DVC.org

  DVC (Data Version Control) is an open-source version control system for machine learning projects, designed to make ML models and datasets shareable and reproducible. It works with Git to version large files, models, and data.

• MLflow

  MLflow is an open-source platform for managing the end-to-end machine learning lifecycle, including experiment tracking, model packaging, and model versioning. It allows users to track experiments, share and deploy models, and manage the complete ML lifecycle with version control for models and code.

• Weights & Biases (W&B)

  Weights & Biases is an MLOps platform that provides tools for experiment tracking, model versioning, dataset versioning, and collaboration for machine learning teams. It helps track, visualize, and version ML experiments and models.

• Comet ML

  Comet ML offers an MLOps platform for machine learning teams, providing tools for experiment tracking, model management, and data versioning. It allows users to version code, data, models, and environments for reproducibility and collaboration.

• PromptLayer

  PromptLayer is an AI engineering platform specifically designed for prompt management and versioning. It allows engineers to track, version control, and A/B test their prompts, enabling more robust prompt engineering workflows.

• GitHub

  GitHub is the world's largest platform for hosting software development projects using Git. It provides robust version control for code, configurations, and increasingly, prompt definitions and related assets in AI engineering projects.

• Hugging Face

  Hugging Face provides a platform and tools for building, training, and deploying machine learning models. Their Hugging Face Hub offers version control for models, datasets, and even 'Spaces' (hosted ML applications), which can include prompt-driven applications.

• ClearML

  ClearML is an open-source MLOps platform that provides an end-to-end solution for experiment management, model versioning, data versioning, and pipeline orchestration. It helps ML teams track, reproduce, and manage their AI development processes.

• Pachyderm

  Pachyderm is a data versioning and data pipeline platform specifically built for machine learning. It provides Git-like version control for data, enabling reproducible data transformations and ML pipelines.