DeltaQuant: 4-bit Video Diffusion Models with Spatiotemporal Delta Smoothing
Xingyang Li, Samuel Tesfai, Zhekai Zhang, Haocheng Xi, Shuo Yang, Lvmin Zhang, Yufei Sun, Kelly Peng, Maneesh Agrawala, Ion Stoica, Kurt Keutzer, Jun-Yan Zhu
Abstract
tokens) to 4 bits while keeping core tokens in FP8. This decomposition substantially reduces quantization error with minimal overhead. For weight quantization, DeltaQuant incorporates SVDQuant's low-rank decomposition to further reduce quantization error. We also implement an efficient kernel that translates DeltaQuant's computational benefits into real-world speedups. Extensive experiments on Wan2.2 I2V, Wan2.2 T2V, and LTX-Video T2V demonstrate that DeltaQuant maintains high generation fidelity. On Wan2.2, it compresses model size by 2.9× and reduces memory footprint by 2.3×. DeltaQuant is compatible with efficient attention mechanisms and few-step distillation. When integrated with these techniques, it achieves an additional 3.0× acceleration, for a total 111.8× end-to-end speedup.
Ask Lune
Know this paper in minutes, not an afternoon.
Lune can read every page of it and the top-tier papers that cite it, and gives you its claims, its evidence and its limits, each one cited to the paper it comes from. No skimming, no AI slop.
Lineage
- SDXL: Improving Latent Diffusion Models for High-Resolution Image Synthesis cites SDEdit: Guided Image Synthesis and Editing with Stochastic Differential Equations
- SDXL: Improving Latent Diffusion Models for High-Resolution Image Synthesis cites Denoising Diffusion Probabilistic Models
- SDEdit: Guided Image Synthesis and Editing with Stochastic Differential Equations cites Denoising Diffusion Probabilistic Models
Builds on30
- Denoising Diffusion Probabilistic ModelsJonathan Ho, Ajay Jain, Pieter AbbeelNeurIPS 2020 · 35,902 citations
- SDXL: Improving Latent Diffusion Models for High-Resolution Image SynthesisDustin Podell, Zion English, Kyle Lacey, Andreas Blattmann et al.ICLR 2024 · 4,569 citations
- SDEdit: Guided Image Synthesis and Editing with Stochastic Differential EquationsChenlin Meng, Yutong He, Yang Song, Jiaming Song et al.ICLR 2022 · 2,128 citations
- SmoothQuant: Accurate and Efficient Post-Training Quantization for Large Language ModelsGuangxuan Xiao, Ji Lin, Mickaël Seznec, Hao Wu et al.ICML 2023 · 1,493 citations
- FlashAttention-3: Fast and Accurate Attention with Asynchrony and Low-precisionJay Shah, Ganesh Bikshandi, Ying Zhang, Vijay Thakkar et al.NeurIPS 2024 · 727 citations
Related papers
- ViDiT-Q: Efficient and Accurate Quantization of Diffusion Transformers for Image and Video GenerationTianchen Zhao, Tongcheng Fang, Haofeng Huang, Rui Wan et al.ICLR 2025
- TwinQuant: Learnable Subspace Decomposition for 4-Bit LLM QuantizationHaodong WANG, Junjie Liu, Zicong Hong, Qianli Liu et al.ICML 2026
- Faster Video Diffusion with Trainable Sparse AttentionPeiyuan Zhang, Yongqi Chen, Haofeng Huang, Will Lin et al.NeurIPS 2025 · 6 citations
- Unlocking Data-free Low-bit Quantization with Matrix Decomposition for KV Cache CompressionPeiyu Liu, Ze-Feng Gao, Xin Zhao, Yipeng Ma et al.ACL 2024 · 2 citations
- ZeroQuant: Efficient and Affordable Post-Training Quantization for Large-Scale TransformersZhewei Yao, Reza Yazdani Aminabadi, Minjia Zhang, Xiaoxia Wu et al.NeurIPS 2022 · 816 citations
