Xichen Pan(@xichen_pan):Modern text-to-image models are increasingly powered by large pretrained LLMs. But there is a curious mismatch: the LLM typically encodes the prompt only once, while the evolving noisy latent states are handled entirely by a newly trained generative backbone. Can pretrained multimodal prior participate in the denoising process? Introducing RepFusion. (1/12) 📄 https://t.co/WbkTtg5M79 🌐

Xichen Pan

@xichen_pan

PhD Student @NYU_Courant, Researcher @AIatMeta; Multimodal Generation | Prev: @MSFTResearch, @AlibabaGroup, @sjtu1896; More at

Joined August 2022

584 Following 746 Followers

Xichen Pan@xichen_pan

2026.06.16 01:23

Modern text-to-image models are increasingly powered by large pretrained LLMs. But there is a curious mismatch: the LLM typically encodes the prompt only once, while the evolving noisy latent states are handled entirely by a newly trained generative backbone. Can pretrained multimodal prior participate in the denoising process? Introducing RepFusion. (1/12) 📄 🌐