Daily Paper Cast
Daily Paper Cast
Jingwen Liang, Gengyu Wang
TurboDiffusion: Accelerating Video Diffusion Models by 100-200 Times
21 minutes Posted Dec 26, 2025 at 3:11 am.
0:00
21:22
Download MP3
Show notes

🤗 Upvotes: 51 | cs.CV, cs.AI, cs.LG

Authors:

Jintao Zhang, Kaiwen Zheng, Kai Jiang, Haoxu Wang, Ion Stoica, Joseph E. Gonzalez, Jianfei Chen, Jun Zhu

Title:

TurboDiffusion: Accelerating Video Diffusion Models by 100-200 Times

Arxiv:

http://arxiv.org/abs/2512.16093v1

Abstract:

We introduce TurboDiffusion, a video generation acceleration framework that can speed up end-to-end diffusion generation by 100-200x while maintaining video quality. TurboDiffusion mainly relies on several components for acceleration: (1) Attention acceleration: TurboDiffusion uses low-bit SageAttention and trainable Sparse-Linear Attention (SLA) to speed up attention computation. (2) Step distillation: TurboDiffusion adopts rCM for efficient step distillation. (3) W8A8 quantization: TurboDiffusion quantizes model parameters and activations to 8 bits to accelerate linear layers and compress the model. In addition, TurboDiffusion incorporates several other engineering optimizations. We conduct experiments on the Wan2.2-I2V-14B-720P, Wan2.1-T2V-1.3B-480P, Wan2.1-T2V-14B-720P, and Wan2.1-T2V-14B-480P models. Experimental results show that TurboDiffusion achieves 100-200x speedup for video generation even on a single RTX 5090 GPU, while maintaining comparable video quality. The GitHub repository, which includes model checkpoints and easy-to-use code, is available at https://github.com/thu-ml/TurboDiffusion.