Uni-MoE: Scaling Unified Multimodal LLMs with Mixture of Experts

Open paper

Benchmark	Model	Rank	Results
visual-question-answering-on-mm-vet	Uni-MoE	#125	GPT-4 score: 32.8