On 8/3/23, NASA, IBM, and Hugging Face announced the public release of a Geospatial FM for NASA Earth Observation data. In collaboration with Clark University’s Center for Geospatial Analytics, the European Space Agency (ESA), USGS, and Oak Ridge National Laboratory, IBM and NASA IMPACT developers constructed a foundation model using NASA’s Harmonized Landsat Sentinel-2 (HLS) dataset using IBM’s watsonx FM stack and Cloud Vela supercomputer. The model, data, and pipelines are collectively called Prithvi and were released to the public as a temporal Vision transformer via the Hugging Face platform. So far, the release was captured by 40+ news outlets globally, including Engadget (with additional mentions in Engadget The Morning After newsletter and Yahoo Finance) VentureBeat and CIO, and 70+ social posts for a total global reach of 10.8M, according to metrics captured by IBM’s communication team. For more details about the HLS Geospatial FM release and the partnership between NASA and IBM, please visit IBM’s press release and the feature article on NASA’s Earthdata website.
Foundation models were applied to HLS imagery, such as this true color composite image of irrigated agricultural fields near Sadat City, about 80 km northwest of Cairo, Egypt. Credit: HLS/NASA IMPACT.