仅用5亿参数的自回归模型实现1024分辨率高质量图像生成
SimpleAR: Pushing the Frontier of Autoregressive Visual Generation through Pretraining, SFT, and RL
- 基于预训练、SFT与强化学习优化，不改架构实现高效生成
- 在GenEval和DPG上分别达0.59和79.66分，质量媲美复杂模型
- 结合vLLM推理加速，单图生成仅需约14秒，适合实用部署
本文提出SimpleAR，一种无需复杂结构修改的纯自回归视觉生成框架。通过精心设计的训练与推理优化，我们证明：1）仅使用0.5B参数即可生成1024x1024高保真图像，在挑战性文本到图像基准测试中表现优异，如GenEval得分为0.59，DPG得分为79.66；2）监督微调（SFT）与组相对策略优化（GRPO）均显著提升生成美感与提示对齐能力；3）结合vLLM等推理加速技术，生成一张1024x1024图像耗时可降至约14秒。通过开源代码，我们希望揭示自回归视觉生成的潜力，推动该领域发展。
原文摘要 · Abstract (English)
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.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。