A novel Multimodal Large Language Model (MLLM) architecture, designed to structurally align visual and textual embeddings. - AIDC-AI/Ovis

It's good at generating alt text for images!