让扩散模型自动学习随提示和步骤变化的采样参数。
Learning Sampling Parameters for Diffusion Models
- 用强化学习训练大模型生成随提示和时间步变化的参数调度
- 在Flux.1和Stable Diffusion 3.5上胜过基线，人类评分胜率达68.12%
- 适合想提升生成质量但不想手动调参的研究者和开发者
文本到图像扩散模型在推理时涉及多个采样参数，包括提示词、反向提示词、无分类器引导系数和噪声调度。这些参数通常一次性手动设定并固定使用，尽管不同提示和生成阶段可能需要不同的参数值。本文提出LeSAMP框架，实现基于提示条件、随时间变化的采样参数学习。将参数选择建模为强化学习问题：给定用户提示，训练大型语言模型输出所选采样参数的调度方案。通过人类偏好模型和视觉语言模型作为裁判提供的奖励进行优化。在Flux.1 [dev]和Stable Diffusion 3.5上评估，相比基线，LeSAMP在人类偏好评分下最高获得68.12%胜率，在VLM-as-a-judge下达73.37%胜率。用户研究进一步验证，最高胜率达59.46%。结果表明，学习到的采样参数策略为提升扩散模型输出提供了一种与现有后训练方法互补的新途径。
原文摘要 · Abstract (English)
Text-to-image diffusion models expose many inference-time sampling parameters, including prompts, negative prompts, classifier-free guidance scales, and noise schedules. These parameters are typically manually chosen once and then held fixed across prompts and denoising timesteps, even though different prompts and stages of generation can benefit from different parameter values. We introduce LeSAMP, a framework for learning prompt-conditioned, timestep-varying sampling parameters. We formulate parameter selection as a reinforcement learning problem: Given a user prompt, a large language model is trained to emit schedules for the chosen sampling parameters. We optimize our model using rewards from human preference models and VLM-as-a-judge. We evaluate our model on Flux.1 [dev] and Stable Diffusion 3.5, and find that compared to baselines, LeSAMP has a win rate of up to 68.12% using human preference scores and 73.37% using VLM-as-a-judge. These gains are validated in a user study where we achieve win rates of up to 59.46% over previous baselines. Our results suggest that learned sampling-parameter policies provide a complementary approach to existing post-training methods for improving diffusion model outputs.
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。