Home·Curated Catalog·Computer Vision
👁 Curated Catalog · Computer Vision

MNIST

70,000 handwritten digits — the canonical intro-ML benchmark.

LQS 83 · gold ✓ Commercial OK 70K images 50 MB Binary · JPG Released 1998
Browse commercial Computer Vision → Visit original source ↗
Source: yann.lecun.com · maintained by Yann LeCun et al.
70K
images
50 MB
Size on disk
83
LQS · gold
1998
First released

About this dataset

MNIST is the most widely-used digit classification benchmark in machine learning. Curated by Yann LeCun from a subset of NIST Special Database 3 and 1, it contains 60,000 training + 10,000 test images of 28×28 handwritten digits (0–9), perfectly balanced across classes. It's the de facto 'hello world' for image classification and every ML library ships with loaders for it.

Maintainer
License
Formats
Binary · JPG

LabelSets Quality Score

LQS is our 7-dimension quality score, computed from the dataset's published statistics. See methodology →

83
out of 100
gold tier

Solid dataset with some trade-offs

Composite score computed from the 7 dimensions below: completeness, uniqueness, validation health, size adequacy, format compliance, label density, and class balance.

Completeness 92
No public completeness metric; using prior for 'research_release' datasets.
Uniqueness 68
Minimal deduplication disclosed.
Validation 92
Labels produced by domain experts or trained annotators.
Size adequacy 81
70,000 images — below 100,000 target for Computer Vision, but usable.
Format compliance 95
Industry-standard format — drop-in compatible with mainstream tooling.
Label density 52
Average 1.0 labels per item (sparse).
Class balance 90
Near-uniform class distribution.

What it's used for

Common tasks and benchmarks where MNIST is the default or competitive choice.

Sample statistics

What's actually in the dataset — from the maintainer's published stats.

60K train + 10K test. 28×28 grayscale. Perfectly balanced across 10 classes (7K per class). Expert-curated from NIST handwritten samples.

License

MNIST is distributed under Public Domain. This is a third-party public dataset; LabelSets indexes and scores it but does not host or redistribute the data. Always verify current license terms with the maintainer before commercial use.

Need commercial-licensed Computer Vision data?

LabelSets sellers offer paid computer vision datasets with what public datasets often can't give you:

Browse paid Computer Vision → Sell your dataset

Similar public datasets

Other entries in the Computer Vision catalog.

Frequently Asked Questions

MNIST is distributed under Public Domain, which generally permits commercial use. Always verify the current license terms with the maintainer (Yann LeCun et al.) before using in a commercial product.
MNIST contains 70,000 images. 60K train + 10K test. 28×28 grayscale. Perfectly balanced across 10 classes (7K per class). Expert-curated from NIST handwritten samples.
MNIST is maintained by Yann LeCun et al. and is available at http://yann.lecun.com/exdb/mnist/. LabelSets indexes and scores this dataset for discoverability but does not redistribute it.
LQS is a 7-dimension quality score (completeness, uniqueness, validation, size adequacy, format compliance, label density, class balance) computed from the dataset's published statistics. Composite scores map to tiers: platinum (≥90), gold (≥75), silver (≥60), bronze (<60). Read the full methodology.