SimpleAR：通过预训练、SFT和RL推动自回归视觉生成的前沿
SimpleAR: Pushing the Frontier of Autoregressive Visual Generation through Pretraining, SFT, and RL
AI总结：
本工作提出朴素自回归视觉生成框架SimpleAR，经训练与推理优化，仅0.5B参数即可生成高保真1024x1024图像，在GenEval等基准上表现具竞争力，SFT和GRPO可提升性能，vLLM加速后生成时间约14秒，揭示自回归生成潜力。
摘要：
本工作提出SimpleAR，一种无需复杂架构修改的朴素自回归视觉生成框架。通过对训练和推理优化的仔细探索，我们证明：1）仅用0.5B参数，模型就能生成1024x1024分辨率的高保真图像，并在具挑战性的文本到图像基准上取得具竞争力的结果，例如GenEval上0.59分、DPG上79.66分；2）监督微调（SFT）和组相对策略优化（GRPO）训练均可显著提升生成美学和提示对齐度；3）经vLLM等推理加速技术优化后，SimpleAR生成1024x1024图像的时间可缩短至约14秒。通过分享这些发现并开源代码，我们希望揭示自回归视觉生成的潜力，鼓励更多人参与该研究领域。代码见https://github.com/wdrink/SimpleAR。
英文摘要：
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.
机构、总结、中文摘要均由AI分析而来，可能存在错误，请以论文正文为准。
