Benchmarking infrastructure underpinning the pathology-FM field from the Mahmood Lab. HEST-1k (NeurIPS 2024, ~422★) pairs histology with spatial transcriptomics for histology↔expression benchmarking, and TRIDENT (~602★, actively maintained) is a large-scale whole-slide-image processing toolkit; the companion Patho-Bench (~152★) standardizes pathology-FM evaluation. Widely used data and tooling for models like TITAN and UNI/CONCH.

Dataset

datasetsciencebiologyinfrastructure