pure-cifar-10 0.1.1

Pure Python/NumPy CIFAR-10 dataset loader

pure_cifar_10

A pure Python, minimal-dependency loader for the CIFAR-10 dataset. This package provides the CIFAR-10 images and labels as NumPy arrays, with automatic downloading and caching, requiring only numpy and tqdm.

Features

• Minimal Dependencies: Only requires numpy and tqdm. No heavy machine learning frameworks like PyTorch or TensorFlow.
• Automatic Handling: Downloads the CIFAR-10 dataset automatically on first use and caches it locally.
• Pure NumPy: Returns standard numpy.ndarray objects for easy integration into any pipeline.
• Progress Visualization: Uses tqdm to show download and loading progress bars.
• Simple API: Mirrors the clean, functional style of the mnist_datasets package.

Installation

You can install the package directly from PyPI using pip:

pip install pure_cifar_10

Usage

from pure_cifar_10 import CIFAR10
loader = CIFAR10() 
train_data, train_labels, test_data, test_labels = loader.load_all()

Optionally specify a custom folder and load train and test data as required.

loader = CIFAR10(folder="/tmp/data) 
train_data, train_labels = loader.load() # train=True by default 
test_data, test_labels = loader.load(train=False)
print(train_data.shape, train_labels.shape)
print(test_data.shape, test_labels.shape)

(50000, 3, 32, 32) (50000,)
(10000, 3, 32, 32) (10000,)

classes = loader.class_names
print(classes)
('airplane', 'automobile', 'bird', 'cat', 'deer', 'dog', 'frog', 'horse', 'ship', 'truck')

Data Format

• Images: The data arrays have shape (N, 3, 32, 32), where N is the number of samples. This is a channels-first format (3 color channels, Red-Green-Blue). Pixel values are float32 in the range [0, 255], as stored in the original dataset.
• Labels: The labels are 1D numpy arrays of dtype int64, containing integers from 0 to 9.

Class Names

The dataset contains 10 mutually exclusive classes: 0. airplane

1. automobile
2. bird
3. cat
4. deer
5. dog
6. frog
7. horse
8. ship
9. truck