DiCoDe: Diffusion-Compressed Deep Tokens for Autoregressive Video Generation with Language Models

TL;DR

DiCoDe introduces a scalable autoregressive video generation method using diffusion-compressed deep tokens, enabling efficient training and high-quality video synthesis with language models, and demonstrating promising results across various model sizes.

Contribution

The paper presents DiCoDe, a novel approach that leverages diffusion-trained deep tokens for scalable autoregressive video generation with language models, achieving high compression and quality.

Findings

Achieves 1000x token compression enabling efficient training.

Performs comparably to existing methods in video quality.

Scaling up model size improves performance consistently.

Abstract

Videos are inherently temporal sequences by their very nature. In this work, we explore the potential of modeling videos in a chronological and scalable manner with autoregressive (AR) language models, inspired by their success in natural language processing. We introduce DiCoDe, a novel approach that leverages Diffusion-Compressed Deep Tokens to generate videos with a language model in an autoregressive manner. Unlike existing methods that employ low-level representations with limited compression rates, DiCoDe utilizes deep tokens with a considerable compression rate (a 1000x reduction in token count). This significant compression is made possible by a tokenizer trained through leveraging the prior knowledge of video diffusion models. Deep tokens enable DiCoDe to employ vanilla AR language models for video generation, akin to translating one visual "language" into another. By treating videos as temporal sequences, DiCoDe fully harnesses the capabilities of language models for autoregressive generation. DiCoDe is scalable using readily available AR architectures, and is capable of generating videos ranging from a few seconds to one minute using only 4 A100 GPUs for training. We evaluate DiCoDe both quantitatively and qualitatively, demonstrating that it performs comparably to existing methods in terms of quality while ensuring efficient training. To showcase its scalability, we release a series of DiCoDe configurations with varying parameter sizes and observe a consistent improvement in performance as the model size increases from 100M to 3B. We believe that DiCoDe's exploration in academia represents a promising initial step toward scalable video modeling with AR language models, paving the way for the development of larger and more powerful video generation models.