LinCa: Accelerating Diffusion Models via Learnable Decomposed Feature Caching

2026-08-18

AI Summary

为解决扩散模型迭代采样计算成本高的问题，提出LinCa框架，通过可学习的可逆网络分解和预测特征，提高加速比同时保持高质量。

Abstract

Diffusion models have achieved remarkable success in image and video generation, yet the high computational cost of iterative sampling remains a critical bottleneck for practical deployment. Feature caching has emerged as a promising acceleration paradigm by reusing or predicting intermediate features across timesteps. However, existing training-free methods apply uniform prediction strategies that cannot adapt to the heterogeneous feature dynamics, causing significant quality degradation under high acceleration ratios. We propose LinCa, a feature caching framework based on learnable invertible networks. LinCa decomposes cached features into sub-components with distinct continuity properties via a lightweight invertible network and applies differentiated prediction orders matched to each component. The strict invertibility guarantees lossless reconstruction back to the original feature space, forming a unified Decompose-Predict-Reconstruct pipeline. By training separate predictors for different models and timestep segments, LinCa adapts to heterogeneous feature dynamics. Experiments on FLUX, Qwen-Image, and HunyuanVideo demonstrate that LinCa, with less than 0.2% additional parameters, significantly outperforms existing methods and maintains near-lossless quality at 5-7x speedup. Code: https://github.com/QHR69/LinCa

Problem

• diffusion models
• computational cost
• feature caching
• heterogeneous feature dynamics
• acceleration

Innovation

• Learnable Invertible Networks
• Feature Decomposition
• Adaptive Prediction
• Heterogeneous Feature Dynamics
• Decompose-Predict-Reconstruct Pipeline