InternVL
An open-source general-purpose vision-language model developed by the Shanghai AI Laboratory. It supports interleaved image-text input and demonstrates strong multimodal understanding capabilities across tasks such as document comprehension and visual question answering.
Core Facts
VLMs (Vision Language Models) typically consist of a visual encoder (such as ViT), a language model, and a projection layer connecting the two, with representative models including GPT-4V/4o, Claude 3.5, the LLaVA series, and InternVL
Open-source multimodal models LLaVA, InternVL, and Qwen-VL typically adopt a 'vision encoder + projection layer + language model' architecture.
Timeline (last 90 days)
Open-weight VLM refers to vision-language models whose parameters can be publicly downloaded and deployed to run locally, such as the Qwen2.5-VL and InternVL series.
A VLM is a multimodal model that integrates a visual encoder on top of a traditional LLM, with typical representatives including LLaVA, Qwen-VL, and InternVL.
All Facts (6)
VLMs (Vision Language Models) typically consist of a visual encoder (such as ViT), a language model, and a projection layer connecting the two, with representative models including GPT-4V/4o, Claude 3.5, the LLaVA series, and InternVL
80%VerifiedOpen-source multimodal models LLaVA, InternVL, and Qwen-VL typically adopt a 'vision encoder + projection layer + language model' architecture.
70%UnverifiedQwen-VL和InternVL分别来自阿里和上海AI Lab，代表国内多模态领域的前沿水平
90%Unverified开源多模态模型包括LLaVA、MiniCPM-V、InternVL、Qwen-VL等
60%UnverifiedOpen-weight VLM refers to vision-language models whose parameters can be publicly downloaded and deployed to run locally, such as the Qwen2.5-VL and InternVL series.
50%UnverifiedA VLM is a multimodal model that integrates a visual encoder on top of a traditional LLM, with typical representatives including LLaVA, Qwen-VL, and InternVL.
50%