Datasets:

swiss-ai
/

nasa

Parquet error: Scan size limit exceeded: attempted to read 14339265645 bytes, limit is 300000000 bytes Make sure that 1. the Parquet files contain a page index to enable random access without loading entire row groups2. otherwise use smaller row-group sizes when serializing the Parquet files

Error code:   TooBigContentError

Need help to make the dataset viewer work? Make sure to review how to configure the dataset viewer, and open a discussion for direct support.

NASA Image Library (Recaptioned)

A public-domain subset of the NASA Image and Video Library, filtered and captioned with a Qwen vision-language model. Prepared by the Swiss AI Initiative vision team.

The set contains 36,997 images, all public domain, covering spaceflight, astronomy, planetary science, Earth observation and aeronautics, from NASA centers including JPL, KSC, MSFC, JSC and GSFC.

How it was made

Images were sourced directly from NASA's public API (images-api.nasa.gov), filtered to public-domain and information-dense items, then recaptioned with Qwen3.6-27B. The caption field is generated and grounded in the image and surrounding text. See the full pipeline and scripts.

License

Every item is NASA public domain. Non-public-domain material (ESA joint content, third-party credits) was filtered out. See NASA's Media Usage Guidelines. This is an independent release by the Swiss AI Initiative, not an official NASA product.

Downloads last month: -