SARLO-80: Worldwide Slant SAR Language Optic Dataset 80cm

The SARLO-80 dataset has been released, comprising 119,566 triplets of very-high-resolution (VHR) synthetic aperture radar (SAR) slant-range patches, aligned optical imagery, and natural-language descriptions. This dataset, built from approximately 2,500 global scenes with native resolutions ranging from 20cm to 2m, utilizes complex-valued SAR measurements and includes standardized 80cm slant-range grids. It is significant for practitioners as it facilitates multimodal learning and alignment tasks in SAR and optical domains, offering fixed train/validation/test splits and full preprocessing code for reproducibility in vision-language training.