Autumn LogoAutumn

> git clone

The Hacker's Machine Intelligence Platform.
The Fastest Machine Learning Framework [1]. And Working On Making Deep Learning 100x More Efficient


The Autumn Architecture

A Machine Intelligence Platform Engineered By Software Developers Not Scientists. Modular, Performant, Portable, Minimal Technical Debt.

Autumn Leaf Collenchyma


Autumn is a modular, portable, end-to-end Platform engineered for developing performant Machine Intelligence applications. It consits of two Frameworks, Leaf and Collenchyma.


Leaf is an imperative, expressive, cross-platform Machine Intelligence Framework, that has a clear architecture and no external dependecies besides Collenchyma. Leaf is easy to maintain, easy to use and highly performant.


Collenchyma is a portable, platform-agnostic, high-performance computation Framework. Applications build with Collenchyma can be deployed and executed in parallel on one or many hetereogenous devices on any host machine.

Leaf was influenced by the brilliant people behind Torch, Caffe, Tensorflow, Rust and numerous research papers. With Autumn we try to transcend the opposing concepts in Machine Intelligence, bringing research and applications closer together.

- Autumn Founders, MJ and Max

Autumn Features

Build maintainable Machine Intelligence applications, that can be deployed into the cloud for production, executed for testing on your desktop and shipped to smartphones, IoT and embedded devices. No code changes required.


Autumn enables developers to build maintainable applications, train and run the models efficiently, reducing development and maintainance costs and increasing productivity.

> More Benchmarks


Autumn enables developers to understand the whole stack, write maintainable plugins and safely implement the latest research outcomes into their applications.

> More Code


Autumn gives you the choice to run, deploy and test your Machine Intelligence applications whereever you want. Any machine, any OS, any hardware. If it doesn't run on iOS, Windows, Android or on a chip without an operating system, it's a bug, let us know about it by creating an issue on Github.

>More Collenchyma

Autumn Examples

Here are some of the ways developers and companies can use the Autumn Platform for building Machine Intelligence driven applications.

Autumn in Production


Engineer reliable, production-ready applications, minimize technical debt and keep your applications easily maintainable.

Autumn and Big Data


Hook Autumn into your big data environment, for scaled learning and fast performance on stream computation thanks to Rusts zero-copy.

Autumn and IoT


Deploy trained models to IoT and embedded devices for on-device, no-internet-required Machine Learning.


We love the ML community very much. Leaf und Collenchyma are free for everyone. You can even choose between the MIT or Apache 2.0 license - whatever strikes your fancy.


With a bit of luck you find us online on the #rust-machine-learing IRC at, but we are always approachable on Gitter/Leaf and Gitter/Collenchyma. If you found a bug or have a feature request, you can create an issue on Github/Leaf and Github/Collenchyma. And for more private matters, send us a mail straight to our inbox


Learn from examples. Refer to Leafs' and Collenchymas' documentation - btw you can help us create a higher-level documentation for the community. Or browse through our Mozilla talk slides.

Let's Hack AI.

Machine Learning researchers came a long way and now we hackers build the tools to make Machine Intelligence accessable for everyone. Autumn provides the platform to make your contributions and extensions open and instantly available to other developers, researchers and companies. Click on the button to start hacking on AI.

> GET STARTED! or apply for a job