RustNN
An easy to use neural network library written in Rust.
Description
RustNN is a feedforward neural network library. The library generates fully connected multi-layer artificial neural networks that are trained via backpropagation. Networks are trained using an incremental training mode.
XOR example
This example creates a neural network with 2 nodes in the input layer, a single hidden layer containing 3 nodes, and 1 node in the output layer. The network is then trained on examples of the XOR function. All of the methods called after train(&examples) are optional and are just used to specify various options that dictate how the network should be trained. When the go() method is called the network will begin training on the given examples. See the documentation for the NN and Trainer structs for more details.
use nn::{NN, HaltCondition};
// create examples of the XOR function
// the network is trained on tuples of vectors where the first vector
// is the inputs and the second vector is the expected outputs
let examples = [
(vec![0f64, 0f64], vec![0f64]),
(vec![0f64, 1f64], vec![1f64]),
(vec![1f64, 0f64], vec![1f64]),
(vec![1f64, 1f64], vec![0f64]),
];
// create a new neural network by passing a pointer to an array
// that specifies the number of layers and the number of nodes in each layer
// in this case we have an input layer with 2 nodes, one hidden layer
// with 3 nodes and the output layer has 1 node
let mut net = NN::new(&[2, 3, 1]);
// train the network on the examples of the XOR function
// all methods seen here are optional except go() which must be called to begin training
// see the documentation for the Trainer struct for more info on what each method does
net.train(&examples)
.halt_condition( HaltCondition::Epochs(10000) )
.log_interval( Some(100) )
.momentum( 0.1 )
.rate( 0.3 )
.go();
// evaluate the network to see if it learned the XOR function
for &(ref inputs, ref outputs) in examples.iter() {
let results = net.run(inputs);
let (result, key) = (results[0].round(), outputs[0]);
assert!(result == key);
}
41 Dec 25, 2022
12 Jul 17, 2022
20 Nov 7, 2022
6 Jul 1, 2022
3 Dec 26, 2022
5 Nov 4, 2020
0 Feb 21, 2021
3 Mar 7, 2022
2 Apr 1, 2022
6 Dec 29, 2022
17 Mar 27, 2023
44 May 6, 2023
6 Jun 12, 2023
2 Jul 11, 2023
199 Oct 23, 2022
41 Sep 16, 2022
19 Aug 11, 2022
354 Jan 1, 2023