Pretraining Billion-scale Geospatial Foundational Models on Frontier

no code implementations • 17 Apr 2024 • Aristeidis Tsaris, Philipe Ambrozio Dias, Abhishek Potnis, Junqi Yin, Feiyi Wang, Dalton Lunga

Although large FMs have demonstrated significant impact in natural language processing and computer vision, efforts toward FMs for geospatial applications have been restricted to smaller size models, as pretraining larger models requires very large computing resources equipped with state-of-the-art hardware accelerators.

Scene Classification Self-Supervised Learning