Fudan Proposes Four-Level Framework for Agentic Visual Generation

Fudan-University · hf · 2026-09-09

A Fudan University team released Agentic Visual Generation, a framework paper classifying agentic visual generation into four levels by how directly the controller decides over generation operations:

The framework maps the evolution from traditional generative models to agentic control paradigms in visual generation.

Original post →

More from Multimodal

Multimodal channel →