arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2510.21167cs.CV
收藏

分块流匹配:改进流匹配模型以实现高效高质量生成

Blockwise Flow Matching: Improving Flow Matching Models For Efficient High-Quality Generation

Korea University(高丽大学);KAIST(韩国科学技术院)

Dogyun Park, Taehoon Lee, Minseok Joo, Hyunwoo J. Kim

AI总结:

针对流匹配模型难以同时捕捉不同时间步特征且推理成本高的问题,提出分块流匹配框架,将生成轨迹分段建模,并引入语义特征引导与轻量特征残差近似,在ImageNet上实现2.1至4.9倍加速。

摘要:

近年来,流匹配模型在广泛领域内推动了高保真数据生成的边界。它通常采用单个大型网络来学习从噪声到数据的整个生成轨迹。尽管这些模型有效,但这种设计难以同时捕捉不同时间步上的不同信号特征,并且由于需要迭代评估整个模型而产生了大量推理成本。为了解决这些局限性,我们提出了分块流匹配(BFM),一种新颖的框架,它将生成轨迹划分为多个时间片段,每个片段由较小但专门的速速块建模。这种分块设计使每个块能够在其指定区间内有效专门化,从而提高推理效率和样本质量。为了进一步提升生成保真度,我们引入了一个语义特征引导模块,该模块明确地根据与预训练表示对齐的语义丰富特征来条件化速度块。此外,我们提出了一种轻量级特征残差近似策略,在显著降低推理成本的同时保持语义质量。在ImageNet 256x256上的大量实验表明,BFM在现有流匹配方法上建立了显著改进的帕累托前沿,在相当的生成性能下实现了推理复杂度2.1倍至4.9倍的加速。代码可在https://github.com/mlvlab/BFM获取。

英文摘要:

Recently, Flow Matching models have pushed the boundaries of high-fidelity data generation across a wide range of domains. It typically employs a single large network to learn the entire generative trajectory from noise to data. Despite their effectiveness, this design struggles to capture distinct signal characteristics across timesteps simultaneously and incurs substantial inference costs due to the iterative evaluation of the entire model. To address these limitations, we propose Blockwise Flow Matching (BFM), a novel framework that partitions the generative trajectory into multiple temporal segments, each modeled by smaller but specialized velocity blocks. This blockwise design enables each block to specialize effectively in its designated interval, improving inference efficiency and sample quality. To further enhance generation fidelity, we introduce a Semantic Feature Guidance module that explicitly conditions velocity blocks on semantically rich features aligned with pretrained representations. Additionally, we propose a lightweight Feature Residual Approximation strategy that preserves semantic quality while significantly reducing inference cost. Extensive experiments on ImageNet 256x256 demonstrate that BFM establishes a substantially improved Pareto frontier over existing Flow Matching methods, achieving 2.1x to 4.9x accelerations in inference complexity at comparable generation performance. Code is available at https://github.com/mlvlab/BFM.

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。

↑