# Autoregressive Multimodal

VILA-U is a foundational model that unifies vision-language understanding and generation tasks, achieving efficient multimodal processing through a single autoregressive framework.

Janus is a novel autoregressive framework that unifies multimodal understanding and generation. By decoupling visual encoding, it addresses the limitations of previous methods and enhances the flexibility of the framework.

Anole is an open-source autoregressive multimodal model capable of generating interleaved image-text sequences without relying on stable diffusion technology.