Compressing Language Models for Specialized Domains

Miles Williams, George Chrysostomou, Vitor Jeronymo +1

Language models (LMs) excel at tasks across diverse domains, yet require substantial computational resources during inference. Compression techniques such as pruning and quantizati…