台风天赋的博客

LVLMs-info: Deepseek-VLdeepseek-vlpaper: https://arxiv.org/abs/2403.05525, code: https://github.com/deepseek-ai/DeepSeek-VL模型基本结构：基本和Qwen-VL一样，三部分，a hybrid vision encoder, a vision adaptor, and a language model（LLM）hybrid vision encoder使用SigLIP处理低