arXivDaily arXiv每日学术速递 周一至周五更新
arXiv周末暂无论文更新,休息一下吧,周末愉快~~
arXiv 2603.19360cs.LG
收藏

预热启动流匹配用于保证快速文本/图像生成

Warm-Start Flow Matching for Guaranteed Fast Text/Image Generation

Samsung AI Center Cambridge(三星AI研究中心剑桥)

Minyoung Kim

AI总结:

本文提出预热启动流匹配(WS-FM)方法,通过利用轻量级生成模型快速生成初始样本,减少流匹配算法的生成时间,同时保证生成质量。

摘要:

当前的自回归(AR)语言模型、基于扩散的文本/图像生成模型以及最近的流匹配(FM)算法能够生成高质量的文本/图像样本。然而,这些模型的推理或样本生成过程通常非常耗时且计算需求大,主要是因为需要大量函数评估对应于标记长度或扩散步数。这还需要大量GPU资源、时间和电力。本文提出了一种新的解决方案,通过保证的速度提升因子减少流匹配算法的样本生成时间,而不牺牲生成样本的质量。我们的关键思想是利用计算轻量级的生成模型,其生成时间相对于目标AR/FM模型可以忽略不计。轻量级模型生成的草稿样本质量不够理想但生成速度快,被用作FM算法的初始分布。与传统使用FM方法所采用的纯噪声(例如高斯或均匀)初始分布不同,草稿样本已经具有相当的质量,因此可以在纯噪声FM情况下将起始时间更接近结束时间而不是0。这将显著减少达到目标数据分布的时间步数,速度提升因子得到保证。我们的想法被称为预热启动FM或WS-FM,可以本质上被视为从低质量草稿样本学习到高质量样本的生成模型。作为概念验证,我们在一些合成玩具数据以及现实世界文本和图像生成任务中展示了该想法,证明了我们的想法在样本生成中提供保证的速度提升,同时不牺牲生成样本的质量。

英文摘要:

Current auto-regressive (AR) LLMs, diffusion-based text/image generative models, and recent flow matching (FM) algorithms are capable of generating premium quality text/image samples. However, the inference or sample generation in these models is often very time-consuming and computationally demanding, mainly due to large numbers of function evaluations corresponding to the lengths of tokens or the numbers of diffusion steps. This also necessitates heavy GPU resources, time, and electricity. In this work we propose a novel solution to reduce the sample generation time of flow matching algorithms by a guaranteed speed-up factor, without sacrificing the quality of the generated samples. Our key idea is to utilize computationally lightweight generative models whose generation time is negligible compared to that of the target AR/FM models. The draft samples from a lightweight model, whose quality is not satisfactory but fast to generate, are regarded as an initial distribution for a FM algorithm. Unlike conventional usage of FM that takes a pure noise (e.g., Gaussian or uniform) initial distribution, the draft samples are already of decent quality, so we can set the starting time to be closer to the end time rather than 0 in the pure noise FM case. This will significantly reduce the number of time steps to reach the target data distribution, and the speed-up factor is guaranteed. Our idea, dubbed {\em Warm-Start FM} or WS-FM, can essentially be seen as a {\em learning-to-refine} generative model from low-quality draft samples to high-quality samples. As a proof of concept, we demonstrate the idea on some synthetic toy data as well as real-world text and image generation tasks, illustrating that our idea offers guaranteed speed-up in sample generation without sacrificing the quality of the generated samples.

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。

↑