Frontier-scale digital pathology foundation models from PathAI, spanning a compact 22M-parameter variant and a 1.1B-parameter flagship trained on 551K whole-slide images.
A lightweight 22M-parameter ViT-S pathology foundation model pre-trained on 195M tiles, handling tasks from subcellular segmentation to slide-level prediction.
Self-supervised vision transformer foundation models for computational pathology, pre-trained on up to 3.1 million whole slide images from 632M to 1.9B parameters.