Eisen Documentation

Eisen is a software package to enable simple and quick development and experimentation with deep learning models. It is designed for applications in healthcare.

Eisen provides a complete set of building blocks and everything in between:

Installing Eisen is extremely simple.

pip3 install --upgrade eisen

Eisen implements an opinionated API that builds directly on PyTorch. The goal of Eisen is to be:

• Extremely simple to use

• Extremely easy to understand

• Similar to other packages in the PyTorch echosystem

• Easy to contribute to

We have subdivided Eisen in several sub-packages:

• Eisen meta-package (installs everything) pip3 install eisen

• Eisen core pip3 install eisen-core

• Eisen CLI pip3 install eisen-cli

• Eisen Deploy pip3 install eisen-deploy

• Eisen Extras pip3 install eisen-extras

You can also obtain Eisen as a Docker image which can be downloaded from dockerhub check out DockerHub at https://hub.docker.com/repository/docker/eisenai/eisen

Eisen core modules are summarized in the figure below:

Building with Eisen is simple. A typical project follows the architecture summarized in the figure below:

Links

Home

• Eisen Documentation
• Links
  • Get Started
    • Obtain Eisen via PyPi
    • Obtain Eisen via DockerHub
  • Train your first model
    • Dataset and Transforms
    • Workflow components
    • Workflow
    • Running Training
  • Contributing to Eisen
    • Implementing Transforms
    • Implementing Datasets
    • Implementing I/O Transforms
    • Implementing Modules
    • Implementing Workflows
  • Datasets
  • I/O
  • Transforms
    • Imaging Transforms
  • Models
    • Segmentation Models
  • Ops
    • Losses
    • Metrics
  • Workflows
  • Hooks
    • Logging hooks
    • Artifacts generation hooks
  • Artifacts
  • Wrappers
  • Other utilities
  • General Information
  • Packaging
    • Handlers
  • Server
  • Clients
  • General Information
  • CLI
  • Intrinsics
  • Tutorials
  • Features
    • Datasets
    • Transforms
    • Workflows
    • Logging
    • Artifacts
    • CLI
    • Compatibility
    • Simplicity
• Indices and tables

Get Started

• Get Started
  • Obtain Eisen via PyPi
  • Obtain Eisen via DockerHub
• Train your first model
  • Dataset and Transforms
  • Workflow components
  • Workflow
  • Running Training
• Contributing to Eisen
  • Implementing Transforms
  • Implementing Datasets
  • Implementing I/O Transforms
  • Implementing Modules
  • Implementing Workflows

Eisen-Core API

• Datasets
• I/O
• Transforms
  • Imaging Transforms
• Models
  • Segmentation Models
• Ops
  • Losses
  • Metrics
• Workflows
• Hooks
  • Logging hooks
  • Artifacts generation hooks
• Artifacts
• Wrappers
• Other utilities

Eisen-Deploy API

• General Information
• Packaging
  • Handlers
• Server
• Clients

Eisen-CLI

• General Information
• CLI
• Intrinsics

Tutorials

• Tutorials

Features

• Features
  • Datasets
  • Transforms
  • Workflows
  • Logging
  • Artifacts
  • CLI
  • Compatibility
  • Simplicity

Indices and tables

• Index

• Module Index

• Search Page