MiniGPT-4: Enhancing Vision-Language Understanding with Advanced Large Language Models

2024-01-16 · ICLR 2024 poster · anchor

Findings

IC-632BLIP-2 succeeds on only 5 out of 100 advanced compositional vision-language tasks
IC-633BLIP-2, LLaVA, and mPLUG-Owl show a trade-off between caption length and hallucination rate on COCO
IC-634Video-ChatGPT achieves the highest correctness, detail, and contextual scores among four video understanding baselines
IC-635InstructBLIP achieves the highest overall MMBench score (44.0) among five evaluated vision-language models