让图像生成更灵活：异步调节去噪时机，提升细节与文本一致性。
AsyncDiff: Asynchronous Timestep Conditioning for Enhanced Text-to-Image Diffusion Inference
- 去噪器条件设置与图像更新解耦，可动态选择更优噪声水平。
- 15步内完成SD3.5生成，10步内完成Flux生成，性能优于同步方法。
- 支持保守或激进调整，适合追求高效与可控生成的开发者。
文本到图像扩散模型通常采用同步调度，即数值积分器推进潜空间至与去噪器条件一致的时间步。本文提出异步推理机制，解耦两者，使去噪器可在不同、学习得到的时间步进行条件设定，同时保持图像更新时间表不变。一个轻量级时间步预测模块（TPM）通过分组相对策略优化（GRPO）训练，基于当前状态选择更合适的条件时间步，有效控制图像细节和纹理丰富度。部署时可通过缩放超参数在原始与非同步时间步间插值，实现保守或激进调整。为保证计算成本可控，将推理步数限制在15步（SD3.5）和10步（Flux）。在MS-COCO 2014和T2I-CompBench数据集上评估，方法优化了图像奖励、HPSv2、CLIP Score与Pick Score的复合奖励，表现持续提升。
原文摘要 · Abstract (English)
Text-to-image diffusion inference typically follows synchronized schedules, where the numerical integrator advances the latent state to the same timestep at which the denoiser is conditioned. We propose an asynchronous inference mechanism that decouples these two, allowing the denoiser to be conditioned at a different, learned timestep while keeping image update schedule unchanged. A lightweight timestep prediction module (TPM), trained with Group Relative Policy Optimization (GRPO), selects a more feasible conditioning timestep based on the current state, effectively choosing a desired noise level to control image detail and textural richness. At deployment, a scaling hyper-parameter can be used to interpolate between the original and de-synchronized timesteps, enabling conservative or aggressive adjustments. To keep the study computationally affordable, we cap the inference at 15 steps for SD3.5 and 10 steps for Flux. Evaluated on Stable Diffusion 3.5 Medium and Flux.1-dev across MS-COCO 2014 and T2I-CompBench datasets, our method optimizes a composite reward that averages Image Reward, HPSv2, CLIP Score and Pick Score, and shows consistent improvement.
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。