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

TL;DR

This paper introduces SnapGen-V, a highly efficient diffusion-based video generation model that can produce a 5-second video on a mobile device in just 5 seconds, making high-quality video synthesis accessible on edge devices.

Contribution

We develop a compact, efficient diffusion model with a specialized architecture and adversarial fine-tuning, enabling fast, high-quality video generation on mobile hardware.

Findings

Generates 5-second videos on iPhone 16 PM within 5 seconds.

Reduces denoising steps to 4 for efficiency.

Achieves comparable quality to server-side models with much faster speed.

Abstract

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.