Global land embeddings on a 0.01° grid (approximately 1 km) from Matryoshka Implicit Neural Distillation (MIND), a geographic implicit neural representation.
MIND maps latitude and longitude to a 3,072-dimensional embedding organized into chunks. Early chunks capture broad geographic patterns, while later chunks add finer spatial variation. Retaining only the leading chunks lets you choose an embedding width suited to your task without retraining the model.
See the MIND project page for the paper, code, and interactive visualizations.