QuantCache: Adaptive Importance-Guided Quantization with Hierarchical Latent and Layer Caching for Video Generation
By: Junyi Wu , Zhiteng Li , Zheng Hui and more
Potential Business Impact:
Makes video creation faster without losing quality.
Recently, Diffusion Transformers (DiTs) have emerged as a dominant architecture in video generation, surpassing U-Net-based models in terms of performance. However, the enhanced capabilities of DiTs come with significant drawbacks, including increased computational and memory costs, which hinder their deployment on resource-constrained devices. Current acceleration techniques, such as quantization and cache mechanism, offer limited speedup and are often applied in isolation, failing to fully address the complexities of DiT architectures. In this paper, we propose QuantCache, a novel training-free inference acceleration framework that jointly optimizes hierarchical latent caching, adaptive importance-guided quantization, and structural redundancy-aware pruning. QuantCache achieves an end-to-end latency speedup of 6.72$\times$ on Open-Sora with minimal loss in generation quality. Extensive experiments across multiple video generation benchmarks demonstrate the effectiveness of our method, setting a new standard for efficient DiT inference. The code and models will be available at https://github.com/JunyiWuCode/QuantCache.
Similar Papers
Q&C: When Quantization Meets Cache in Efficient Image Generation
CV and Pattern Recognition
Makes AI image creation much faster.
MixCache: Mixture-of-Cache for Video Diffusion Transformer Acceleration
Graphics
Makes videos create faster without losing quality.
CacheQuant: Comprehensively Accelerated Diffusion Models
CV and Pattern Recognition
Makes AI image creation much faster and smaller.