Elara Math

Elara Math is a Rust-native math library, with (current or planned support for):

• Tensors (n-dimensional arrays)* (supported)
• Linear algebra with tensors (supported)
• Numerical solvers for integration & differential equations (WIP)
• Automatic differentiation (supported)

*: GPU tensors are not available yet, but GPU acceleration is planned to be added in the future

Shoutouts: Acknowledgements

As an example, here is a working tiny neural network using elara-math.

use elara_math::prelude::*;

const EPOCHS: usize = 10000;
const LR: f64 = 1e-5;

fn main() {
    let train_data = tensor![
        [0.0, 0.0, 1.0],
        [1.0, 1.0, 1.0],
        [1.0, 0.0, 1.0],
        [0.0, 1.0, 1.0]];
    let train_labels = tensor![
        [0.0],
        [1.0],
        [1.0],
        [0.0]
    ].reshape([4, 1]);
    let mut weights = Tensor::rand([3, 1]);
    for epoch in 0..EPOCHS {
        let output = train_data.matmul(&weights).relu();
        let loss = elara_math::mse(&output, &train_labels);
        println!("Epoch {}, loss: {:?}", epoch, loss);
        loss.backward();
        let adjustment = weights.grad() * LR;
        weights = weights - Tensor::new(adjustment);
        weights.zero_grad();
    }
    let pred_data = tensor![[1.0, 0.0, 0.0]];
    let pred = &pred_data.matmul(&weights).relu();
    println!("Weights after training: {:?}", weights);
    println!("Prediction [1, 0, 0] -> {:?}", pred.borrow().data);
}

Developing

To develop elara-math, first clone the repository:

git clone https://github.com/elaraproject/elara-math

Then, copy over the pre-commit githook:

cp .githooks/pre-commit .git/hooks/pre-commit && chmod a+x .git/hooks/pre-commit

You should then be all set to start making changes!