T2I-R1: Reinforcing Image Generation with Collaborative Semantic-level and Token-level CoT

Dongzhi Jiang, Ziyu Guo, Renrui Zhang, Zhuofan Zong, Hao Li, Le Zhuo, Shilin Yan, Pheng-Ann Heng, Hongsheng Li

CUHK· Shanghai AI Laboratory· CPII under InnoHK

Autoregressive Image Generation; Chain-of-Thought (CoT) Reasoning; Multimodal Large Language Model

Abstract

Recent advancements in large language models have demonstrated how chain-of-thought (CoT) and reinforcement learning (RL) can improve performance. However, applying such reasoning strategies to the visual generation domain remains largely unexplored. In this paper, we present T2I-R1, a novel reasoning-enhanced text-to-image generation model, powered by RL with a bi-level CoT reasoning process.

Specifically, we identify two levels of CoT that can be utilized to enhance different stages of generation:

the semantic-level CoT for high-level planning of the prompt
the token-level CoT for low-level pixel processing during patch-by-patch generation.

To better coordinate these two levels of CoT, we introduce BiCoT-GRPO with an ensemble of generation rewards, which seamlessly optimizes both generated CoTs within the same training step.

By applying our reasoning strategies to the baseline model, Janus-Pro, we achieve superior performance with 13% improvement on T2I-CompBench and 19% improvement on the WISE benchmark, even surpassing the state-of-the-art model FLUX.1. All the training code is in the supplementary material and will be made public.