This repo contains two Jupyter Notebook tutorials and a Google Earth Engine script presented at the 2026 IEEE M2GARSS Tutorial session in April 2026 in Marrakech, Morocco.
-
(open in Colab) Fine-tuning Prithvi EO 2.0 foundation model for flood mapping. This tutorial uses the Sen1Floods11 dataset.
-
(open in Colab) Embedding generation from Prithvi EO 2.0, TerraMind v1 and SSL4EO DINOv2 for burn scar mapping. This tutorial uses the HLS Burn Scars dataset.
-
(open in GEE code editor) Detecting land cover change using Google AlphaEarth annual embeddings.
The Jupyter Notebooks presented in this tutorial were adopted from example notebooks prepared by the TerraTorch development team (source)