Dataset Card for Fashion-MNIST

Dataset Details

Dataset Description

Fashion-MNIST is a dataset of 70,000 grayscale images, each 28×28 pixels, representing 10 different classes of clothing and accessories. It serves as a drop-in replacement for the original MNIST dataset but provides a more challenging benchmark for machine learning models. The dataset was introduced by Zalando Research to address the limitations of MNIST, which primarily contains handwritten digits.

Dataset Sources

• Homepage: https://github.com/zalandoresearch/fashion-mnist?tab=readme-ov-file#license
• Paper: Xiao, H., Rasul, K., & Vollgraf, R. (2017). Fashion-mnist: a novel image dataset for benchmarking machine learning algorithms. arXiv preprint arXiv:1708.07747.

Dataset Structure

Total images: 70,000

Classes: 10 (T-shirt/top, Trouser, Pullover, Dress, Coat, Sandal, Shirt, Sneaker, Bag, Ankle boot)

Splits:

• Train: 60,000 images

• Test: 10,000 images

Image specs: PNG format, 28×28 pixels, Grayscale

Example Usage

Below is a quick example of how to load this dataset via the Hugging Face Datasets library.

from datasets import load_dataset  

# Load the dataset  
dataset = load_dataset("randall-lab/fashion-mnist", split="train", trust_remote_code=True)  
# dataset = load_dataset("randall-lab/fashion-mnist", split="test", trust_remote_code=True)  

# Access a sample from the dataset  
example = dataset[0]  
image = example["image"]  
label = example["label"]  

image.show()  # Display the image  
print(f"Label: {label}")

Citation

BibTeX:

@online{xiao2017/online, author = {Han Xiao and Kashif Rasul and Roland Vollgraf}, title = {Fashion-MNIST: a Novel Image Dataset for Benchmarking Machine Learning Algorithms}, date = {2017-08-28}, year = {2017}, eprintclass = {cs.LG}, eprinttype = {arXiv}, eprint = {cs.LG/1708.07747}, }