Lumen: Parameter-Efficient Alignment of Pretrained Vision and Language Encoders for Zero-Shot Computational Pathology
Lumen, which aligns frozen Virchow2 and BioMedBERT backbones using rank-4 adapters and projection heads using rank-4 adapters and projection heads, yields strong and transferable performance at patch and slide level while training only a small fraction of the parameters.
Kiarash Tajbakhsh, Abdelrahman Faqieh, Michael Jopiti et al.
· 0 citations