SparseXceptionMIL (SparseXMIL)
github.com/loic-lb/sparsexmilSparseXMIL is a framework designed for whole slide image (WSI) analysis in digital pathology. It uses sparse convolutions within an Xception architecture to efficiently model spatial context for Multiple Instance Learning (MIL), addressing GPU memory limitations.
Sourced from
- GitHub — github.com/loic-lb/sparsexmil