LLaVA: Large Language and Vision Assistant

Authors: Liu et al. Year: 2023 ArXiv/Link: https://arxiv.org/abs/2304.08485

Summary

Connects vision and language modalities using a simple projection layer, creating effective multimodal LLMs.

Key Concepts

• Multimodal fusion
• Vision-language integration
• Instruction-following
• Visual understanding
• LLM-based reasoning

Impact

Practical and efficient multimodal language models

Category

Vision-Language Models

---