Qwen2-VL: Enhancing Vision-Language Model’s Perception of the World at Any Resolution2024-09-25VLMsVLMs, LLMs, Qwens