LLM-grounded Video Diffusion Models
This work addresses the challenge of intricate spatiotemporal video generation for AI and creative applications, representing an incremental improvement by integrating existing LLM and diffusion components in a novel way.
The paper tackles the problem of generating videos from complex text prompts by introducing LLM-grounded Video Diffusion (LVD), which uses a large language model to create dynamic scene layouts that guide a diffusion model, resulting in significantly improved performance over baseline methods in generating videos with desired attributes and motion patterns.
Text-conditioned diffusion models have emerged as a promising tool for neural video generation. However, current models still struggle with intricate spatiotemporal prompts and often generate restricted or incorrect motion. To address these limitations, we introduce LLM-grounded Video Diffusion (LVD). Instead of directly generating videos from the text inputs, LVD first leverages a large language model (LLM) to generate dynamic scene layouts based on the text inputs and subsequently uses the generated layouts to guide a diffusion model for video generation. We show that LLMs are able to understand complex spatiotemporal dynamics from text alone and generate layouts that align closely with both the prompts and the object motion patterns typically observed in the real world. We then propose to guide video diffusion models with these layouts by adjusting the attention maps. Our approach is training-free and can be integrated into any video diffusion model that admits classifier guidance. Our results demonstrate that LVD significantly outperforms its base video diffusion model and several strong baseline methods in faithfully generating videos with the desired attributes and motion patterns.