Yume-1.5: A Text-Controlled Interactive World Generation Model

TL;DR

Yume-1.5 introduces a real-time, text-controlled world generation framework that overcomes previous diffusion model limitations by integrating context compression, streaming acceleration, and interactive exploration.

Contribution

It presents a novel framework combining unified context compression, real-time streaming, and text control for interactive world generation from images or text.

Findings

Supports real-time, interactive world exploration

Achieves realistic world generation from minimal prompts

Enables text-controlled world event creation

Abstract

Recent approaches have demonstrated the promise of using diffusion models to generate interactive and explorable worlds. However, most of these methods face critical challenges such as excessively large parameter sizes, reliance on lengthy inference steps, and rapidly growing historical context, which severely limit real-time performance and lack text-controlled generation capabilities. To address these challenges, we propose \method, a novel framework designed to generate realistic, interactive, and continuous worlds from a single image or text prompt. \method achieves this through a carefully designed framework that supports keyboard-based exploration of the generated worlds. The framework comprises three core components: (1) a long-video generation framework integrating unified context compression with linear attention; (2) a real-time streaming acceleration strategy powered by bidirectional attention distillation and an enhanced text embedding scheme; (3) a text-controlled method for generating world events. We have provided the codebase in the supplementary material.