AI Paper Atlas

Research map / Image generation / Text-to-image diffusion

Proxy-Tuning: Tailoring Multimodal Autoregressive Models for Subject-Driven Image Generation

Yi Wu, Shengju Qian, Lingting Zhu, Lei Liu, Wandi Qiao, Ziqiang Li, Lequan Yu, Bin Li

CVPR 2026 · Image generation › Text-to-image diffusion · approach: Inference-time T2I improvement · 6 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
fine-tune AR model on diffusion-supervisor proxy images (Proxy-Tuning) instead of direct LoRALlamaGen · DreamBooth (SDXL supervisor)DINO: 0.5088 -> 0.7436Improved
Proxy-Tuning instead of direct LoRALumina-mGPT · DreamBooth (SDXL supervisor)DINO: 0.4504 -> 0.7834Improved
AR student vs diffusion supervisor on subject fidelityLumina-mGPT · DreamBooth (SDXL supervisor)CLIP-I: 0.8002 -> 0.8074Improved
Proxy-Tuning on prompt adherence (CLIP-T)LlamaGen · DreamBooth (SDXL supervisor)CLIP-T: 0.3225 -> 0.2772Hurt
Proxy-Tuning on prompt adherence (CLIP-T)Lumina-mGPT · DreamBooth (SDXL supervisor)CLIP-T: 0.3101 -> 0.3118Mixed
apply Proxy-Tuning to diffusion students instead of AR (ablation)SDXL · DreamBoothDINO: 0.7272 -> 0.6890Hurt
apply Proxy-Tuning to diffusion students (ablation)SD3 · DreamBoothCLIP-I: 0.7822 -> 0.7596Hurt
learn multiple subjects jointly in one AR modelLlamaGen (SDXL) · Multi-subject DreamBoothDINO: 0.7040 -> 0.7176Improved
Proxy-Tuning vs fine-tuning AR on prompt fidelity (human)AR model · User studyPrompt fidelity: 2.98 -> 4.52Improved
Proxy-Tuning for FLUX supervisor on LlamaGen subject fidelityLlamaGen · DreamBooth (FLUX supervisor)CLIP-I: 0.8023 -> 0.7974No clear effect
vary number of proxy images and LoRA rank (ablation)LlamaGen · DreamBooth (SDXL supervisor)CLIP-I / DINONo clear effect
geometric data augmentation instead of proxy data (ablation)AR model · DreamBoothsemantic editabilityHurt

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy