zalandoresearch/fashion-mnist resource

Health v2 · maintenance only

Flags: no_releases

Adoption not part of the score

12810 stars · 3071 forks observed · 2026-08-28

What it is AI-extracted, prompt v1, taxonomy v1, 2026-08-29, confidence not recorded

Fashion-MNIST is a dataset of 70,000 28x28 grayscale Zalando article images across 10 classes, designed as a direct drop-in replacement for the original MNIST dataset. It includes a Python loader and a public benchmark dashboard for comparing machine learning algorithms.

Use cases

• benchmark image classification models on a harder MNIST alternative
• load a drop-in MNIST replacement for training CNNs
• test machine learning algorithms on fashion product images
• compare model performance on a standard vision dataset
• teach deep learning with a simple but non-trivial dataset
• evaluate GANs on small grayscale images

When to choose

• you need a standard, well-known image classification benchmark harder than MNIST
• you want a small dataset that loads quickly for prototyping vision models
• you need identical image size and split structure to MNIST for easy swapping

When to avoid

• you need high-resolution or color images
• you need real-world production-grade fashion image data
• your task requires more than 10 coarse clothing categories

Facets

dataset · maturity stable

machine-learning deep-learning computer-vision data-science benchmarking machine-learning computer-vision image-processing data-science python cross-platform fashion-mnist mnist image-classification benchmark-dataset zalando grayscale-images

Member repositories

For agents

markdown · JSON · MCP: product_card(name="zalandoresearch/fashion-mnist")