视频生成速度提升100-200倍，质量几乎不变。
TurboDiffusion: Accelerating Video Diffusion Models by 100-200 Times
- 用低比特注意力和稀疏线性注意力加速计算
- 通过步骤蒸馏与8位量化实现100-200倍提速
- 适合需要快速生成视频的开发者和研究者
我们提出TurboDiffusion，一个视频生成加速框架，可在保持视频质量的前提下，将端到端扩散模型生成速度提升100-200倍。该框架主要依赖三项技术：(1) 注意力加速：采用低比特SageAttention和可训练稀疏线性注意力（SLA）；(2) 步骤蒸馏：使用rCM实现高效蒸馏；(3) W8A8量化：将模型参数与激活值量化至8位，加速线性层并压缩模型。此外还包含多项工程优化。在Wan2.2-I2V-14B-720P、Wan2.1-T2V-1.3B-480P、Wan2.1-T2V-14B-720P和Wan2.1-T2V-14B-480P模型上实验显示，即使仅用单张RTX 5090 GPU，也能实现100-200倍加速，且视频质量相当。代码与模型权重已开源。
原文摘要 · Abstract (English)
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.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。