autograd
A library to run the computation graphs, whose backend is rust-ndarray.
Overview
- (Higher order) automatic differentiation
- Neural net first APIs
- Pure Rust
- Dynamic/static graph construction with shared variables
Example
extern crate autograd as ag;
let ref x = placeholder;
let ref y = variable;
// `z` is a target of partial differentiation.
let ref z = 2*x*x + 3*y + 1;
// dz/dy
let ref g1 = gradients;
// dz/dx (necessary to fill the placeholder `x`)
let ref g2 = gradients;
// ddz/dx (second order derivative)
let ref gg = gradients;
// evaluation of symbolic gradients
assert_eq!;
let feed_dict = new.add;
assert_eq!;
assert_eq!;
For more, see examples or tests.
Documentation
WIP
License
MIT