David Dinkevich

Abstract

Fewer latent tokens promise cheaper diffusion. Yet aggressive latent compression creates a reconstruction-generation dilemma: adding channels improves reconstruction but makes generation harder. We introduce APEX, short for Asynchronous Prefix denoising for EXtreme compression, which pairs a hierarchical latent space with asynchronous denoising over channel groups. Our autoencoder orders channel prefixes from structure to detail, while one shared denoiser generates them at staggered local times without adding tokens. On ImageNet-512, APEX outperforms existing methods at every evaluated channel count under 64× and 128× compression. At matched training and guided sampling compute, it also outperforms a 256-token model at every evaluated budget, with larger gains at lower compute.

Abstract

Story2Board is a training-free framework for expressive storyboard generation from natural language. It addresses spatial composition, background evolution, and narrative pacing in addition to subject identity. We introduce a lightweight consistency framework composed of Latent Panel Anchoring, which preserves a shared character reference across panels, and Reciprocal Attention Value Mixing, which softly blends visual features between token pairs with strong reciprocal attention. Together, these mechanisms enhance coherence without architectural changes or fine-tuning, enabling state-of-the-art diffusion models to generate visually diverse yet consistent storyboards. We use an off-the-shelf language model to convert free-form stories into grounded panel-level prompts. We also introduce the Rich Storyboard Benchmark and a Scene Diversity metric for evaluating layout variation and background-grounded storytelling alongside consistency.