A library to run the computation graphs, whose backend is rust-ndarray.
```rust
extern crate autograd as ag;
let ref x = ag::placeholder(&[1]);
let ref y = ag::variable(ag::init::zeros(&[1]));
// z is a target of partial differentiation.
let ref z = 2xx + 3*y + 1;
// dz/dy
let ref g1 = ag::gradients(z, &[y], None)[0];
// dz/dx (necessary to fill the placeholder x)
let ref g2 = ag::gradients(z, &[x], None)[0];
// ddz/dx (second order derivative)
let ref gg = ag::gradients(g2, &[x], None)[0];
// evaluation of symbolic gradients asserteq!(3., g1.eval()[0]); let feeddict = ag::Input::new().add(x, ag::init::fromscalar(2.)); asserteq!(8., g2.evalwithinput(feeddict)[0]); asserteq!(4., gg.eval()[0]); ```
For more, see examples or tests.
WIP
MIT