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

TL;DR

T2I-R1 introduces a novel reasoning-enhanced text-to-image generation model that employs bi-level chain-of-thought reasoning and reinforcement learning to improve image quality and coherence, surpassing existing models on multiple benchmarks.

Contribution

The paper proposes a new bi-level CoT reasoning framework with reinforcement learning for text-to-image generation, integrating semantic and token-level reasoning to enhance performance.

Findings

Achieved 13% improvement on T2I-CompBench

Achieved 19% improvement on WISE benchmark

Surpassed state-of-the-art model FLUX

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: (1) the semantic-level CoT for high-level planning of the prompt and (2) 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 generation 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. Code is available at: https://github.com/CaraJ7/T2I-R1