MNIST handwritten digits
70,000 labelled images of handwritten digits - the standard first test that a neural network implementation works at all.
A set of 60,000 training images and 10,000 test images of handwritten digits, assembled by the National Institute of Standards and Technology (the NIST in MNIST) in the 1980s. Chollet describes solving it as the "Hello World" of deep learning: it is small enough to train on a laptop, yet real enough that a broken implementation fails visibly. Use it to check that a network learns at all before pointing it at anything harder.
Key columns
Representative fields, refer to the source for the full, authoritative schema.
| Column | Type | Description |
|---|---|---|
| pixel0 … pixel783 | integer | Greyscale intensity 0–255 for each cell of the 28×28 image, flattened row-major. |
| label | integer | The digit shown, 0–9. |
License: Free for research and educational use; see the source page for terms.