XetHub for

Machine Learning

Develop Models Faster and More Collaboratively With Git Workflows

How Gather AI Accelerated Their Model Development Process with XetHub

Model Iteration Just Got Faster.

Train models faster, experiment fearlessly, and collaborate across your team with XetHub. Streamline your process by keeping data, code, and models in one place.

  • Store your code, data, and models in one place.
  • Quickly iterate and deploy models using standard Git commands.
  • Simplify your workflow by replacing data storage and model management platforms with XetHub.

As we performed our technical evaluation of XetHub, we found that it scaled well as our repo sizes got larger. It was easy to adopt and required almost no training for the engineers on the ML team.

The usage-based pricing model makes it easy to align our costs with system utilization, unlike some other models based on team size.”

Daniel Maturana, Co-Founder and Chief ML Scientist


reduction in repository size and transfer time


data silos eliminated by switching to XetHub


cost savings over using EBS, Git LFS, and DVC

Work Faster and Smarter

Most ML teams use GitHub for code and store their data somewhere else. It’s inefficient and error-prone, and the development process is further bogged down by large volumes of data.

XetHub brings data and code together, enabling fast, seamless collaboration in a Git-enabled environment at scale. It allows ML teams to manage code, models, and metadata in lock-step, replacing Git LFS, DVC, and EBS/S3 in their workflow. This reduces friction between team members, leading to a better model development experience.

Data & Model Management, Simplified

ML models are trained using massive data sets, which can cost thousands of dollars annually to store and take forever to be downloaded. Using XetHub’s automatic block-level deduplication reduces a repository’s size on average 40%, improving development efficiency by reducing download times and lower costs. The smaller repository size also means faster deployment, which translates into saved time for ML engineers.

XetHub also offers a local file system mount option as well as a web file explorer to provide quick ways to access and understand a repository’s contents. Xet Mount allows ML teams to instantly gain read-only access to a repository as if it were a folder on their machine, regardless of its size. Developers no longer have to suffer through long download times just to check a dataset’s contents.

Integrates with Your Existing ML Workflows

Rather than having to move all your workflows to adopt a monolithic platform, XetHub works within your existing developer workflows and supports any filetype of any size.

Stay in the flow with free branching, data remixing, and generation of new features and aggregates - with the confidence that you won’t lose your work.

XetHub logo with white text

Delightful collaboration on code and data at any scale.

Copyright current year XetData