SnapGen-V: Generating a Five-Second Video within Five Seconds on a Mobile Device

• Published in: arXiv.org 2024-12
• Main Authors: Wu, Yushu, Zhang, Zhixing, Li, Yanyu, Xu, Yanwu, Kag, Anil, Yang, Sui, Coskun, Huseyin, Ma, Ke, Lebedev, Aleksei, Hu, Ju, Metaxas, Dimitris, Wang, Yanzhi, Tulyakov, Sergey, Ren, Jian
• Format: Article
• Language: English
• Subjects: Algorithms; Cloud computing; Diffusion layers; Image processing; Image resolution; Smartphones
• ISSN: 2331-8422

We have witnessed the unprecedented success of diffusion-based video generation over the past year. Recently proposed models from the community have wielded the power to generate cinematic and high-resolution videos with smooth motions from arbitrary input prompts. However, as a supertask of image generation, video generation models require more computation and are thus hosted mostly on cloud servers, limiting broader adoption among content creators. In this work, we propose a comprehensive acceleration framework to bring the power of the large-scale video diffusion model to the hands of edge users. From the network architecture scope, we initialize from a compact image backbone and search out the design and arrangement of temporal layers to maximize hardware efficiency. In addition, we propose a dedicated adversarial fine-tuning algorithm for our efficient model and reduce the denoising steps to 4. Our model, with only 0.6B parameters, can generate a 5-second video on an iPhone 16 PM within 5 seconds. Compared to server-side models that take minutes on powerful GPUs to generate a single video, we accelerate the generation by magnitudes while delivering on-par quality.