MGit: A Model Versioning and Management System

TL;DR

MGit is a system designed to efficiently manage, version, and track the lineage of machine learning models, reducing storage costs and improving collaboration and update processes.

Contribution

It introduces a lineage graph and storage optimizations for model management, enabling efficient versioning, provenance tracking, and automatic updates.

Findings

Reduces lineage graph storage by up to 7x

Enables automatic downstream model updates

Facilitates better collaboration and testing

Abstract

Models derived from other models are extremely common in machine learning (ML) today. For example, transfer learning is used to create task-specific models from "pre-trained" models through finetuning. This has led to an ecosystem where models are related to each other, sharing structure and often even parameter values. However, it is hard to manage these model derivatives: the storage overhead of storing all derived models quickly becomes onerous, prompting users to get rid of intermediate models that might be useful for further analysis. Additionally, undesired behaviors in models are hard to track down (e.g., is a bug inherited from an upstream model?). In this paper, we propose a model versioning and management system called MGit that makes it easier to store, test, update, and collaborate on model derivatives. MGit introduces a lineage graph that records provenance and versioning information between models, optimizations to efficiently store model parameters, as well as abstractions over this lineage graph that facilitate relevant testing, updating and collaboration functionality. MGit is able to reduce the lineage graph's storage footprint by up to 7x and automatically update downstream models in response to updates to upstream models.