The 4,000 labeled synthetic datasets, 3,000 unlabeled field datasets for "Massive-scale unlabeled field and labeled synthetic seismic datasets of global shelf-edge scale clinothems"

The 4,000 labeled synthetic datasets, 3,000 unlabeled field datasets curated by the Computational Interpretation Group (CIG) for for automatic relative geologic time (RGT) estimation, Hui Gao, Xinming Wu, Jintao Li, Xiaoming Sun, and Jiarun Yang are the main contributors to the dataset. 1. The 4,000 labeled synthetic datasets: Synthetic seismic clinothem datasets generated by geological and geophysical forward modeling (stratigraphic forward modeling, geological structural deformation modeling, and geophysical forward modeling), each paired with dense relative geological time (RGT) labels, faults labels, and corresponding geological priors (seismic normal vectors, and automatically extracted local horizion segments). 2. The 3,000 unlabeled field datasets: Curated field seismic clinothem datasets collected from multiple sedimentary basins (Taranaki, Northern Carnarvon, North Sea, Browse, Beagle, Offshore Canterbury, and Colville Basins), representing diverse stratigraphic architectures and geological settings.