By clicking “Accept”, you agree to the storing of cookies on your device to enhance site navigation, analyze site usage, and assist in our marketing efforts. View our Privacy Policy for more information.
DenyAccept
XetHub for

Model iteration and development

Train and deploy ML models faster with flexible, scalable storage that adapts to your workflow.

Store your code and data in one place for reliable reproducibility, guaranteed provenance, and unlimited time travel.

Try XetHub now
XetHub benchmark

Scale to tomorrow’s ML storage needs today

XetHub is built for ML projects at massive scale, with an experience that starts on your laptop.
‍

Seamlessly scale projects to 10 TB and above
No limits on individual file size or number of files per project
Automatic block-level deduplication

“We rely on computer vision and ML to deliver on Gather AI’s mission. XetHub has enabled our ML team to be more productive, collaborate efficiently, and iterate quickly.”

Daniel Maturana

Co-founder and Chief ML Scientist
Open case study

41%

reduction in repository size and transfer time

4

data silos eliminated by switching to XetHub

51%

cost savings over using EBS, Git LFS, and DVC

Automatic time travel - batteries included

Access any version of a file or repository from your choice of tool -browse the XetHub UI, list from your CLI, stream through Python, or clone from Git.

Dead simple import from existing object stores
Track parallel development through branches
Easily compare experiments with one command
Debug across time to identify data/model drift and anomalies
Full reproducibility across data, model, and code changes
Get started for free

Mount models in seconds

XetHub is designed for fast streaming of data and models from anywhere. Speed up your iteration cadence with a tool that makes waiting obsolete.

At the edge: Work on projects larger than local disk without juggling I/O considerations. Write your training code as if you were on production from day one, no translation needed.
In the data center: Keep expensive GPUs busy with training, instead of idling during downloads. XetHub mount provides a 25% speedup on your first epoch of training and consistent performance for subsequent epochs.
Work with whatever frameworks and tools you like: XetHub Mount creates a read-only virtual filesystem in seconds.
Accelerate even further with powerful and tunable prefetching and caching options for optimal efficiency.
Get started for free

More features

Time travel

Mistakes happen. But with infinite time travel, you can recover with ease.

Difference tracking

Easily see how your work has evolved across branches and commits.

Big data friendly

Never worry about outgrowing your platform again. XetHub can handle petabytes of data.

Instant repository mount

Access large data like its local without worrying about download time or local storage space.

Managed storage

We handle the storage so you can do the important things.

Explore and visualize

Easily gain insight from your data with custom visualizations and interactive dashboards.

CSV previews

Automatically generated profiles to help you understand the large tables in your data.

All major OSes supported

Use the tools and devices you already have.

Simplify your ML workflow today

Stop juggling multiple tools and streamline your workflow with XetHub.
Get started for free