The Metropolitan Museum of Art Open Access: collection metadata and images
- Provider
- The Metropolitan Museum of Art
- License
- CC0-1.0 (source)
- Size
- 159 files
- Format
- Parquet (in sampled files)
- Stored at
- Hugging Face (https://huggingface.co/datasets/metmuseum/openaccess)
- Access
- Free and public. No account with the data provider is needed.
- Category
- Computer Vision
- Source
- https://github.com/metmuseum/openaccess
About The Metropolitan Museum of Art Open Access: collection metadata and images
The Metropolitan Museum of Art's own Hugging Face release of its Open Access collection data, covering more than 470,000 artworks spanning 5,000 years of art from around the world. Each row is one collection object with the Met's catalogue fields: objectID, accession number and year, isHighlight and isPublicDomain flags, department, object name, title, culture, period, dynasty, reign, portfolio, artist name, role, nationality, biography and dates, and links to the primary and additional image URLs. For public-domain objects the repository also embeds the image itself, fetched from the Met Collection API primaryImageSmall field with img2dataset, so these are web-resolution, compressed images rather than full-resolution masters; rows without an image are kept and can be filtered on isPublicDomain or a blank url. Uses: art and cultural heritage image classification, fine-grained attribute recognition (culture, medium, period), image retrieval, captioning and digital humanities. Measured 2026-11-14: 101 Parquet shards in default/train (about 418 GB), plus gzipped CSV metadata in source/ and openaccess.csv.gz at the top level; open with the datasets library, pandas or pyarrow. The Met has waived all rights under Creative Commons Zero (CC0 1.0) for these select datasets and public-domain images; it asks users to consider citing the Met and not to imply endorsement.
How to load this dataset in Python
from datasets import load_dataset
ds = load_dataset("metmuseum/openaccess")
print(ds)Data structure
File types in sampled files: Parquet
Files
Example file: default/train/00000.parquet
Analyze this dataset without downloading it
On Scigantic, The Metropolitan Museum of Art Open Access: collection metadata and images opens in a hosted JupyterLab notebook with the data already mounted, so you can read files where they are stored instead of copying them first. Each notebook includes an AI assistant that can see the mounted files and write and run analysis code. Open the dataset page to launch a notebook, or browse the full catalog.
This page is also available as Markdown and as JSON.
Related Computer Vision datasets
- Quick, Draw! Dataset: 50 Million Hand-Drawn Sketches in 345 Categories as Stroke Vectors and 28x28 Bitmaps (Google Creative Lab)
- uCO3D: UnCommon Objects in 3D
- Textvqa
- YouTube UGC Dataset: User-Generated Content Video Clips with Mean Opinion Scores for Video Quality Assessment and Compression (Google)
- Nemotron VLM Dataset V 2
- PUG Animals
- Minimal Video Pairs
- ABU: Airport-Beach-Urban hyperspectral anomaly/target detection benchmark