Flexible Diffusion Modeling of Long Videos

William Harvey, Saeid Naderiparizi, Vaden Masrani, Christian Dietrich Weilbach, Frank Wood

Keywords: generative modeling, denoising diffusion probabilistic model, video modeling

TL;DR: We apply diffusion generative models to video. To model long videos with bounded computational resources, we present an architecture which can flexibly generate, condition on, or ignore any subsets of video frames.

Abstract: We present a framework for video modeling based on denoising diffusion probabilistic models that produces long-duration video completions in a variety of realistic environments. We introduce a generative model that can at test-time sample any arbitrary subset of video frames conditioned on any other subset and present an architecture adapted for this purpose. Doing so allows us to efficiently compare and optimize a variety of schedules for the order in which frames in a long video are sampled and use selective sparse and long-range conditioning on previously sampled frames. We demonstrate improved video modeling over prior work on a number of datasets and sample temporally coherent videos over 25 minutes in length. We additionally release a new video modeling dataset and semantically meaningful metrics based on videos generated in the CARLA autonomous driving simulator.

Community Implementations: [ 3 code implementations](https://www.catalyzex.com/paper/arxiv:2205.11495/code)