Back to researchers
Silvio Savarese
BLIP-2 and frozen-encoder multimodal LLMs
Co-authored BLIP-2: a key step toward efficient vision-language models built around LLM backbones.
Highlights
MultimodalVision-languageLLMs
Focus: BLIP-2 and frozen-encoder multimodal LLMs
Why it matters: Co-authored BLIP-2: a key step toward efficient vision-language models built around LLM backbones.
Research Areas
MultimodalVision-languageLLMs