SimpleAR: Pushing the Frontier of Autoregressive Visual Generation through Pretraining, SFT, and RL

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Wang, Junke, Tian, Zhi, Wang, Xun, Zhang, Xinyu, Huang, Weilin, Wu, Zuxuan, Jiang, Yu-Gang
Format: Preprint
Published: 2025
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866915244162416640
author Wang, Junke
Tian, Zhi
Wang, Xun
Zhang, Xinyu
Huang, Weilin
Wu, Zuxuan
Jiang, Yu-Gang
author_facet Wang, Junke
Tian, Zhi
Wang, Xun
Zhang, Xinyu
Huang, Weilin
Wu, Zuxuan
Jiang, Yu-Gang
contents This work presents SimpleAR, a vanilla autoregressive visual generation framework without complex architecure modifications. Through careful exploration of training and inference optimization, we demonstrate that: 1) with only 0.5B parameters, our model can generate 1024x1024 resolution images with high fidelity, and achieve competitive results on challenging text-to-image benchmarks, e.g., 0.59 on GenEval and 79.66 on DPG; 2) both supervised fine-tuning (SFT) and Group Relative Policy Optimization (GRPO) training could lead to significant improvements on generation aesthectics and prompt alignment; and 3) when optimized with inference acceleraton techniques like vLLM, the time for SimpleAR to generate an 1024x1024 image could be reduced to around 14 seconds. By sharing these findings and open-sourcing the code, we hope to reveal the potential of autoregressive visual generation and encourage more participation in this research field. Code is available at https://github.com/wdrink/SimpleAR.
format Preprint
id arxiv_https___arxiv_org_abs_2504_11455
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle SimpleAR: Pushing the Frontier of Autoregressive Visual Generation through Pretraining, SFT, and RL
Wang, Junke
Tian, Zhi
Wang, Xun
Zhang, Xinyu
Huang, Weilin
Wu, Zuxuan
Jiang, Yu-Gang
Computer Vision and Pattern Recognition
This work presents SimpleAR, a vanilla autoregressive visual generation framework without complex architecure modifications. Through careful exploration of training and inference optimization, we demonstrate that: 1) with only 0.5B parameters, our model can generate 1024x1024 resolution images with high fidelity, and achieve competitive results on challenging text-to-image benchmarks, e.g., 0.59 on GenEval and 79.66 on DPG; 2) both supervised fine-tuning (SFT) and Group Relative Policy Optimization (GRPO) training could lead to significant improvements on generation aesthectics and prompt alignment; and 3) when optimized with inference acceleraton techniques like vLLM, the time for SimpleAR to generate an 1024x1024 image could be reduced to around 14 seconds. By sharing these findings and open-sourcing the code, we hope to reveal the potential of autoregressive visual generation and encourage more participation in this research field. Code is available at https://github.com/wdrink/SimpleAR.
title SimpleAR: Pushing the Frontier of Autoregressive Visual Generation through Pretraining, SFT, and RL
topic Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2504.11455