Search papers, labs, and topics across Lattice.
This paper introduces Progressive Seed Pruning (PSP), a novel method that enhances inference-time scaling for diffusion models by allowing aggressive pruning of less promising seed trajectories early in the process. By evaluating multiple seeds upfront and selectively focusing on the most promising candidates, PSP effectively utilizes a fixed compute budget to improve both automated and human evaluation scores in image generation tasks. The results demonstrate that PSP outperforms traditional methods like best-of-$N$, importance-sampling, and tree-search baselines, achieving superior prompt alignment while maintaining a constant memory footprint.
Pruning less promising seed trajectories early can significantly boost image generation quality without increasing compute costs.
Diffusion and flow-matching models dominate conditional image generation, yet inference-time scaling for these models is far less developed than for autoregressive language models. Because final quality is highly sensitive to the initial noise seed, many approaches spend extra compute on seed search or resampling under a black-box reward, but typically maintaining a constant memory footprint throughout inference. We show that relaxing this constraint enables an underexplored inference-time scaling axis: by front-loading exploration, evaluating many seeds early, and pruning aggressively, we can use a fixed compute budget more effectively. \emph{Progressive Seed Pruning} (\PSP) scores intermediate denoised estimates and progressively narrows the candidate set so that only promising trajectories are fully denoised, while keeping the total number of model evaluations fixed. Across diffusion and flow-matching backbones, \PSP \ consistently improves reward-guided selection and achieves higher GenEval scores (automated) and better human evaluation on prompt-alignment than best-of-$N$, importance-sampling, and tree-search baselines at matched compute. Project page: https://www.vision.caltech.edu/psp. Code: https://github.com/rogerioagjr/psp.
