歸納矩匹配

Inductive Moment Matching

March 10, 2025

作者: Linqi Zhou, Stefano Ermon, Jiaming Song

cs.AI

摘要

擴散模型和流匹配方法雖能生成高品質樣本，但在推理時速度緩慢，且將其蒸餾為少步模型常導致不穩定性和大量調參需求。為解決這些權衡問題，我們提出了歸納矩匹配（Inductive Moment Matching, IMM），這是一種新型生成模型，適用於單步或少量步數的採樣，並採用單階段訓練程序。與蒸餾不同，IMM無需預訓練初始化及優化兩個網絡；與一致性模型相比，IMM保證了分佈層面的收斂性，並在多種超參數和標準模型架構下保持穩定。IMM在ImageNet-256x256上僅用8步推理便以1.99的FID超越了擴散模型，並在CIFAR-10上從零開始訓練的模型中，達到了2步FID為1.98的頂尖水平。

English

Diffusion models and Flow Matching generate high-quality samples but are slow at inference, and distilling them into few-step models often leads to instability and extensive tuning. To resolve these trade-offs, we propose Inductive Moment Matching (IMM), a new class of generative models for one- or few-step sampling with a single-stage training procedure. Unlike distillation, IMM does not require pre-training initialization and optimization of two networks; and unlike Consistency Models, IMM guarantees distribution-level convergence and remains stable under various hyperparameters and standard model architectures. IMM surpasses diffusion models on ImageNet-256x256 with 1.99 FID using only 8 inference steps and achieves state-of-the-art 2-step FID of 1.98 on CIFAR-10 for a model trained from scratch.