We introduce LLaVA-Critic, the first open-source large multimodal model (LMM) designed as
a generalist evaluator to assess performance across a wide range of multimodal tasks …