Explore live open-source projects and AI models.

Official code for CVPR 2022 (Oral) paper "Deep Visual Geo-localization Benchmark"

GEO-Bench: Toward Foundation Models for Earth Monitoring

GEOBench-VLM: Benchmarking Vision-Language Models for Geospatial Tasks

[CVPR' 26] Benchmarking PhD-Level Coding in 3D Geometric Computer Vision

A toolbox for benchmarking SOTA discriminative and generative geometry estimation models.

GeoAnalystBench: A GeoAI benchmark for assessing large language models for spatial analysis workflow and code generation

TerraFM is a scalable foundation model for unified multisensor Earth observation, trained on 18.7M Sentinel-1/2 tiles and achieving state-of-the-art results on GEO-Bench and Copernicus-Bench.

No description available from source.

LLM-agents benchmark set of geospatial tasks requiring multi-step tool use; and LLM-as-Judge based evaluation framework.

Code for GEO-Bench V2 datasets. Made in collaboration with TUM, IBM and ServiceNow.