DeepSeek-VL

model
DeepSeek
Vision-language model with dynamic tiling encoder for high-resolution image understanding.
Paper (arXiv)GitHubHuggingFace

Paper

Citations 45
arXiv HTML
multimodalopen-weight

Related