TensorState’s Documentation

Introduction

TensorState is a toolbox designed to analyze the way neural networks process information to calculate metrics about the network architecture.

Both Tensorflow and PyTorch are supported, but complex networks may prove problematic for some of the network functions (such as automatically building an efficiency model). Let us know if you run into problems.

Please send comments, suggestions, and feedback to nicholas.j.schaub@gmail.com.

Table of Contents

• Installation
  • Introduction
  • Windows and Python >= 3.6
  • Linux
  • Mac OSX
  • Troubleshooting
• List of Tutorials
  • A Simple Tensorflow Tutorial
  • A Simple PyTorch Tutorial
• Reference
  • tensorstate.TensorState
  • tensorstate.Layers
  • tensorstate.States
• What is state space?
  • An Abstract Explanation