Crate anofox_ml_neural_networks 

Multi-layer perceptron (MLP) neural networks for classification and regression.

This crate provides MlpClassifier and MlpRegressor, feedforward neural networks trained with backpropagation and gradient descent.

Examples

use anofox_ml_neural_networks::MlpClassifier;
use anofox_ml_core::{Fit, Predict};
use ndarray::array;

let x = array![[0.0, 0.0], [0.0, 1.0], [1.0, 0.0], [1.0, 1.0]];
let y = array![0.0, 1.0, 1.0, 0.0]; // XOR

let mlp = MlpClassifier {
    hidden_layer_sizes: vec![10, 10],
    learning_rate: 0.01,
    max_iter: 500,
    seed: 42,
    batch_size: None,
    alpha: 0.0,
    ..Default::default()
};

let model = mlp.fit(&x, &y).unwrap();
let preds = model.predict(&x).unwrap();
assert_eq!(preds.len(), 4);

Re-exports

pub use activation::Activation;

pub use mlp_classifier::FittedMlpClassifier;

pub use mlp_classifier::MlpClassifier;

pub use mlp_regressor::FittedMlpRegressor;

pub use mlp_regressor::MlpRegressor;

pub use solver::Solver;

Modules

activation

mlp_classifier

mlp_regressor

network

solver

utils