ZeroSmooth: Training-free Diffuser Adaptation for High Frame Rate Video Generation

TL;DR

ZeroSmooth introduces a training-free, plug-and-play method for high frame rate video generation by interpolating frames in diffusion models, maintaining temporal consistency without additional training.

Contribution

The paper presents a novel training-free video interpolation approach for diffusion models, enabling high frame rate video synthesis without extra training or large datasets.

Findings

Effective frame interpolation comparable to trained models

Maintains temporal consistency in generated videos

Applicable to multiple diffusion-based video models

Abstract

Video generation has made remarkable progress in recent years, especially since the advent of the video diffusion models. Many video generation models can produce plausible synthetic videos, e.g., Stable Video Diffusion (SVD). However, most video models can only generate low frame rate videos due to the limited GPU memory as well as the difficulty of modeling a large set of frames. The training videos are always uniformly sampled at a specified interval for temporal compression. Previous methods promote the frame rate by either training a video interpolation model in pixel space as a postprocessing stage or training an interpolation model in latent space for a specific base video model. In this paper, we propose a training-free video interpolation method for generative video diffusion models, which is generalizable to different models in a plug-and-play manner. We investigate the non-linearity in the feature space of video diffusion models and transform a video model into a self-cascaded video diffusion model with incorporating the designed hidden state correction modules. The self-cascaded architecture and the correction module are proposed to retain the temporal consistency between key frames and the interpolated frames. Extensive evaluations are preformed on multiple popular video models to demonstrate the effectiveness of the propose method, especially that our training-free method is even comparable to trained interpolation models supported by huge compute resources and large-scale datasets.