Drew's Comic NewsroomAbout

Comics on the day’s news

Claude spent the weekend fixing other models’ bad habitsCoyote vs. Acme opens at 95% on the TomatometerAn ETF is a basket that trades like a stockSome quantum-gravity clues may be two maps of the same hillSara Bareilles turns a Lincoln Center lobby into a grief archiveBáez saved it, McGonigle ended it, Tigers 2-1A US Open guard’s wife weaponizes Tsitsipas quotesThe book nobody can read is still unreadClaude spent the weekend fixing other models’ bad habitsCoyote vs. Acme opens at 95% on the TomatometerAn ETF is a basket that trades like a stockSome quantum-gravity clues may be two maps of the same hillSara Bareilles turns a Lincoln Center lobby into a grief archiveBáez saved it, McGonigle ended it, Tigers 2-1A US Open guard’s wife weaponizes Tsitsipas quotesThe book nobody can read is still unread
Four-panel full-color science comic. Panel 1: Lunar Reconnaissance Orbiter mosaic tiles covering the Moon. Panel 2: open-source badges for Hugging Face and GitHub. Panel 3: cartoon researchers mapping craters and irregular mare patches. Panel 4: polar ice prospectivity heatmap near lunar poles.
Four panels: LRO mosaic Moon tiles, Hugging Face and GitHub open release, crater and irregular mare mapping, polar ice prospectivity heatmap. Science comic. · Comic: Topics / Drew’s Comic Newsroom. Source: NASA Science.

Science and Space

NASA-IBM Lunar Foundation Model opens on Hugging Face and GitHub

What happened

NASA Science, dated September 10, 2026, reports the launch of the NASA-IBM Lunar Foundation Model — among the first open-source AI foundation models built specifically for lunar science. The model is hosted publicly on Hugging Face with the complete codebase on GitHub. It was trained primarily on data from NASA’s Lunar Reconnaissance Orbiter (LRO), using roughly 2 million image tiles (more than 1 million high-resolution camera images at 1-meter resolution and nearly 964,000 multispectral images at 100-meter resolution), plus terrain and imagery from missions such as GRAIL, Lunar Prospector, and JAXA’s SELENE.

Researchers can fine-tune the pre-trained model for tasks including crater mapping, spotting young volcanic features known as irregular mare patches, and estimating where ice may be stable near the lunar poles — including permanently shadowed regions. NASA Science frames the release as open science: machine learning-ready pre-training datasets and benchmarks ship alongside the model, which is integrated into the open-source TerraTorch toolkit. Ignore wrong training counts or closed-source claims that may appear inside comic art; the verified filing is ~2M LRO tiles, Hugging Face + GitHub, open release.

A dated open-model drop with named lunar tasks is a clean Science slot. Readers get the Sept. 10 stamp, NASA-IBM collaboration, ~2M LRO tiles, Hugging Face and GitHub hosting, crater mapping, irregular mare patches, polar ice prospectivity, and the open-science framing — not a rewrite of every baseline comparison in the companion paper.

Why it matters

Readers get the instrument heritage (LRO-led training), the ~2 million tile count, the public Hugging Face and GitHub release, and the three headline adaptation tasks: craters, irregular mare patches, and polar ice prospectivity. Keep politics out. Keep the NASA Science URL on the page. Bright lunar gray and ice-cyan. White gutters. An open foundation model with a receipt is the strip.

Conclusion

NASA and IBM’s Lunar Foundation Model, announced Sept. 10 via NASA Science, is open on Hugging Face and GitHub after training on roughly 2 million LRO tiles — supporting crater mapping, irregular mare patch detection, and polar ice prospectivity as open science. Source: https://science.nasa.gov/science-research/artificial-intelligence-lunar-foundation-model/

Source: NASA Science