Installation
RustyML 0.17.0 requires Rust 1.89.0 or later. The examples use Rust edition 2024 and ndarray 0.17.2.
Create a project
-
Create a Cargo project.
cargo new rustyml-example -
Open the project directory.
cd rustyml-example -
Add these dependencies to
Cargo.toml.[dependencies] rustyml = "=0.17.0" ndarray = "=0.17.2" -
Replace
src/main.rswith the program in Your first model. -
Run the program.
cargo run
RustyML enables all public modules by default. Declare ndarray directly because the examples import its arrays and macros.
Select features
To select modules, disable default features and list the features you need. For example, this dependency supports classical models, data utilities, and evaluation metrics.
[dependencies]
rustyml = { version = "=0.17.0", default-features = false, features = ["machine_learning", "utils", "metrics"] }
ndarray = "=0.17.2"
| Feature | Public APIs |
|---|---|
machine_learning | Regression, classification, clustering, and dimensionality reduction |
neural_network | Models, layers, losses, and optimizers |
utils | Data splits, scalers, and label conversion |
metrics | Regression, classification, and clustering scores |
math | Distance functions and reduction functions |
full | All five modules above |
show_progress | Training progress bars in supported operations |
The default feature enables full.
Each of machine_learning, neural_network, utils, and metrics also enables math.
Cargo features add APIs to the build.
The show_progress feature does not enable a model module.
Import APIs
The examples use module paths to identify each API. For shorter imports, use a category prelude.
use ndarray::array;
use rustyml::prelude::machine_learning::*;
fn main() -> Result<(), rustyml::error::Error> {
let x = array![[1.0], [2.0]];
let y = array![3.0, 5.0];
let mut model = LinearRegression::new(true);
model.fit(&x, &y)?;
println!("prediction: {:.1}", model.predict(&array![[3.0]])?[0]);
Ok(())
}Output:
prediction: 7.0
Other category preludes are utils, metrics, and neural_network under rustyml::prelude.
The root prelude exports APIs from every enabled category.
With metrics enabled, its Average name refers to the metrics enum.
Use rustyml::prelude::neural_network::Average for the merge layer.