Crate auto_diff

An auto-difference library

Introduction

This is yet another auto-difference library for deep neural network. The focus is easy on use and dynamic computation graph building.

Install

Add auto-diff = “0.5” to the [dependencies] section of your project Cargo.toml file.

Features

The forward operators support a commonly used set, including:

1. getter/setter,
2. index and slicing,
3. +, -, *, / and matmul,
4. speciall functions,
5. statistics,
6. linear algebra,
7. random number generator.

The corresponding gradient is work-in-progress.

One feature of auto-diff is the auto-difference is in background and don’t get in your way if only forward calculation is needed. Thus it can be used without syntax like variable place holder.

Example

Re-exports

pub use var::Var;

pub use err::AutoDiffError;

Modules

collection

compute_graph

err

op

optim

Gradient based optimization.

serde

var

var_inner

Macros

var_f64