ESA/Hubble Full-Resolution Space Images with Captions (2,706 Images)

Provider
Maxwell Weinzierl (UT Dallas); images and captions by ESA/Hubble
License
CC-BY-4.0 (source)
Size
202 files
Format
Parquet (in sampled files)
Stored at
Hugging Face (https://huggingface.co/datasets/Supermaxman/esa-hubble)
Access
Free and public. No account with the data provider is needed.
Category
Astronomy & Space
Source
https://github.com/Supermaxman/hubble-diffusion

About ESA/Hubble Full-Resolution Space Images with Captions (2,706 Images)

Full-resolution processed images from the NASA/ESA Hubble Space Telescope published by ESA/Hubble at esahubble.org, each paired with the title and caption text written by ESA/Hubble science writers and with archive metadata. 2,706 images (galaxies, nebulae, star clusters, gravitational lenses, planetary nebulae and solar system targets, from instruments including WFPC2, ACS and WFC3) average about 2,881 by 3,267 pixels and 44 MB each, up to 432 MB. Fields include the image, text (title plus description), ESA/Hubble image id such as heic0604a or opo9941a, image type, credit line, source URL and release date. Data are 200 Parquet shards under data/ (about 111 GB, measured 2026-11-16); stream with Hugging Face datasets rather than downloading whole. Built for text-to-image generation (Hubble Diffusion), and also usable for astronomical image captioning, retrieval and vision-language evaluation. These are colour-composited outreach images, not calibrated science frames; use MAST for photometry. ESA/Hubble releases its images, videos and web texts under CC-BY-4.0 provided the credit given with each image is shown; cite Weinzierl 2023 and keep each image's credit.

How to load ESA/Hubble Full-Resolution Space Images with Captions (2,706 Images) in Python

from datasets import load_dataset
ds = load_dataset("Supermaxman/esa-hubble")
print(ds)

Data structure

File types in sampled files: Parquet

Files

Example file: data/train-00000-of-00200-ca85803776ee0942.parquet

Analyze this dataset without downloading it

On Scigantic, ESA/Hubble Full-Resolution Space Images with Captions (2,706 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 Astronomy & Space datasets