{"success":true,"data":{"author":"Yann LeCun, Corinna Cortes, Christopher J.C. Burges","bucketAccessType":"public","bucketName":"https://azureopendatastorage.blob.core.windows.net/mnist","bucketPrefix":"","category":"Machine Learning & AI","cloudProvider":"azure","createdAt":"2026-11-12T19:20:37.689Z","description":"70,000 grayscale 28x28 images of handwritten digits (60,000 training, 10,000 test), each labelled 0-9, in the original IDX format (gzip) and as CSV.\n\nThis copy is Microsoft's, from Azure Open Datasets. The top level holds the four original IDX files (train/t10k images and labels, gzip) and the same data as CSV (one row per image, 784 pixel columns). A `processed/` folder holds Microsoft's preprocessed copy.\n\nAbout 1.4 GB in 30 files. Read anonymously from Azure Blob Storage.","displayTitle":"MNIST Handwritten Digits","id":"73be5823-ea64-5139-a0a8-02ccc061a7f4","license":"CC BY-SA 3.0","referenceUrl":"https://learn.microsoft.com/azure/open-datasets/dataset-mnist","summary":"70,000 grayscale 28x28 images of handwritten digits (60,000 training, 10,000 test), each labelled 0-9, in the original IDX format (gzip) and as CSV.","title":"MNIST Handwritten Digits","updatedAt":"2026-11-12T19:46:03.385Z","visibility":"public","websiteUrl":"http://yann.lecun.com/exdb/mnist/"}}