Autograd documentation

Author:Paxton Maeder-York, Adam Nitido, Dylan Randle and Simon Sebbagh Date:Jan 01, 2019 Version:1.0

Autograd is a forward and reverse mode Automatic Differentiation (AD) software library. Autograd also supports optimization.

To install the latest release, type:

pip install dragongrad

See the Installation notes for details.

Contents

• Introduction
• Background
• Installation
  • Pypi installation
  • GitHub Installation
  • Requirements
• Autograd Usage
  • General rules
  • Simple Differentiation Case
  • Differentiation of Functions
  • Multiple Inputs
  • Forward or Reverse Mode
  • Optimization
  • Gradient Descent
  • Adam
• Software Organization
• Implementation
  • Forward Mode
  • Reverse Mode
• Notebooks
  • Demo Notebook
  • Converge Notebook
• License
• Future Work

Indices and tables

• Index
• Module Index
• Search Page