ENHANCED MULTIMODAL LANGUAGE MODELS WITH MOME

Summary from the TLDR AI newsletter

Researchers propose the Mixture of Multimodal Experts (MoME) to tackle the issue of task interference in generalist Multimodal Large Language Models (MLLMs).