Module datasets 

Canonical ML dataset loaders

Provides convenient one-liner access to well-known ML datasets for tutorials, examples, and benchmarking.

Example

use alimentar::datasets::{iris, mnist, cifar10};

// Load Iris dataset (embedded, no download)
let iris = iris()?;
println!("Iris: {} samples", iris.len());

// Load MNIST (downloads from HuggingFace Hub on first use)
let mnist = mnist()?;
let (train, test) = mnist.split()?;

Structs

Cifar10Dataset

CIFAR-10 image classification dataset

Cifar100Dataset

CIFAR-100 image classification dataset

DatasetSplit

Split information for train/test datasets

FashionMnistDataset

Fashion-MNIST clothing classification dataset

IrisDataset

The Iris flower dataset

MnistDataset

MNIST handwritten digits dataset

Constants

CIFAR10_CLASSES

CIFAR-10 class names

CIFAR100_COARSE_CLASSES

CIFAR-100 coarse class names (20 superclasses)

CIFAR100_FINE_CLASSES

CIFAR-100 fine class names (100 classes)

FASHION_MNIST_CLASSES

Fashion-MNIST class names

Traits

CanonicalDataset

A canonical ML dataset with train/test split support

Functions

cifar10

Load CIFAR-10 dataset (embedded 100-sample subset)

cifar100

Load CIFAR-100 dataset (embedded 100-sample subset)

fashion_mnist

Load Fashion-MNIST dataset (embedded 100-sample subset)

iris

Load the Iris dataset

mnist

Load MNIST dataset (embedded 1000-sample subset)