通过引入动量项加速生成模型采样，十倍减少计算量仍保质量。
Implicit Dynamical Flow Fusion (IDFF) for Generative Modeling
- 引入动量项构建新向量场，支持更大采样步长。
- 在CIFAR-10和CelebA上仅需十分之一网络评估次数，性能相当。
- 适用于图像与时间序列生成，尤其在分子模拟与海温数据上表现优异。
条件流匹配（CFM）模型可从非信息性先验生成高质量样本，但通常需要数百次网络评估（NFE）。为解决此问题，我们提出隐式动力学流融合（IDFF）：IDFF学习一个带额外动量项的新向量场，可在生成过程中采用更长步长，同时保持生成分布的保真度。结果表明，与CFM相比，IDFF将NFE降低一个数量级，且不牺牲样本质量，实现快速采样，并高效处理图像与时间序列生成任务。我们在标准基准如CIFAR-10和CelebA上评估，其似然与生成质量与CFM及扩散模型相当，且所需NFE更少。在分子模拟与海表温度（SST）等时间序列数据建模任务中，IDFF亦表现更优，凸显其跨领域通用性与有效性。
原文摘要 · Abstract (English)
Conditional Flow Matching (CFM) models can generate high-quality samples from a non-informative prior, but they can be slow, often needing hundreds of network evaluations (NFE). To address this, we propose Implicit Dynamical Flow Fusion (IDFF); IDFF learns a new vector field with an additional momentum term that enables taking longer steps during sample generation while maintaining the fidelity of the generated distribution. Consequently, IDFFs reduce the NFEs by a factor of ten (relative to CFMs) without sacrificing sample quality, enabling rapid sampling and efficient handling of image and time-series data generation tasks. We evaluate IDFF on standard benchmarks such as CIFAR-10 and CelebA for image generation, where we achieve likelihood and quality performance comparable to CFMs and diffusion-based models with fewer NFEs. IDFF also shows superior performance on time-series datasets modeling, including molecular simulation and sea surface temperature (SST) datasets, highlighting its versatility and effectiveness across different domains.\href{https://github.com/MrRezaeiUofT/IDFF}{Github Repository}
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。