Python package for solid deep learning.


Eisen is an open source package that facilitates training, validation, testing and deployment of deep learning algorithms. Check out the documentation now!

Simple API

Eisen proposes simple interfaces to complex functionality. No over-engineering means developers can achieve more with less effort.

Modular Architecture

Eisen is built to be modular, so you can use only what you need! It is mostly compatible with other packages in the PyTorch universe.

Complete Workflow

Eisen is with you from training to deployment. Once your model is fit you can serve it through a HTTP interface.

Build Visually

Tired of coding? You can design experiments and define your DL workflow by visually mixing and matching Eisen building blocks.

Try it now!

Get started with our tutorials or ...

... to test Eisen on a powerful GPU directly in your browser!

Open source

Eisen is and will always be free and open source. We have built it to serve the needs of a community of developers who need a solid foundation to do science. We aim to make development simple, experiments reproducible and deployment of DL models straightforward. Our philosophy is to avoid over-engineering and keep a simple, clean and readable code base. Feel free to contribute and improve the code, or just create your own version of it.

Modular architecture

Eisen offers an opinionanted API to tap into medical image analysis and volumetric image understanding capabilties. Our architecture is inspired by Torchvision and is mostly compatible with other software packages in the PyTorch universe. You can mix and match Eisen modules and take the functionality you need into your projects. You can access Eisen by importing it as you would do with any other python packages or by taking advantage of its command line interface (CLI). Find out more here.

Innovative features

From training to deployment, Eisen is always with you. In addition to python coding, you can build your experiments visually via a convenient and simple user interface allowing users to configure every aspect of training, validation, testing and deployment in just a few clicks. Once your models are ready, you can make them available in a server-client fashion no matter the number and kind of their input and outputs. Find out more about this and other features here!

Step into Eisen

Eisen is free and open source. Our software is organized in several sub-packages that you can fork and contribute to on GitHub. Reach out on Slack.