InternLM-XComposer2-4KHD: A Pioneering Large Vision-Language Model Handling Resolutions from 336 Pixels to 4K HD
Research paper by Xiaoyi Dong and collaborators. Published 2024-04-09. Topics: Large Vision-Language Model, LVLM, high-resolution understanding, patch division paradigm, dynamic resolution.
Resource information
- Authors
- Xiaoyi Dong, Pan Zhang, Yuhang Zang, Yuhang Cao, Linke Ouyang, Songyang Zhang, Haodong Duan, Wenwei Zhang, Yining Li, Hang Yan, Yang Gao, Zhe Chen, Xinyue Zhang, Wei Li, Jingwen Li, Wenhai Wang, Kai Chen, Conghui He, Xingcheng Zhang, Jifeng Dai, Yu Qiao, Dahua Lin, Jiaqi Wang
- Research topics
- Large Vision-Language Model, LVLM, high-resolution understanding, patch division paradigm, dynamic resolution, automatic patch configuration, Vision Transformer, ViT, training image aspect ratios, consistent performance enhancements, GPT-4V, Gemini Pro
- Publication date
- 2024-04-09
- Publication
- arXiv
- Last record update
- Nov 3, 2026