NVLM: Open Frontier-Class Multimodal LLMs
Research paper by Wenliang Dai and collaborators. Published 2024-09-17. Topics: multimodal large language models, decoder-only multimodal LLMs, cross-attention-based models, 1-D tile-tagging design, tile-based dynamic high-resolution images.
Resource information
- Authors
- Wenliang Dai, Nayeon Lee, Boxin Wang, Zhuoling Yang, Zihan Liu, Jon Barker, Tuomas Rintamaki, Mohammad Shoeybi, Bryan Catanzaro, Wei Ping
- Research topics
- multimodal large language models, decoder-only multimodal LLMs, cross-attention-based models, 1-D tile-tagging design, tile-based dynamic high-resolution images, multimodal pretraining, supervised fine-tuning datasets, multimodality, vision-language tasks, text-only performance, math and reasoning data
- Publication date
- 2024-09-17
- Publication
- arXiv
- Last record update
- Nov 3, 2026