CtrlCache: Accelerating Interactive Video World Models with Control-Aware Caching
cs.CV
Submitted: 2026-10-06
Updated: 2026-10-06
Comments: 18 pages. Project page: https://wrecklong.github.io/CtrlCache/
Project page: https://wrecklong.github.io/CtrlCache/ABSTRACT
License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
Terminology
Sources
- One Threshold Is Not Enough: Prompt-Invariant Caching Schedules for Video Diffusion
- $\Delta$-DiT: A Training-Free Acceleration Method Tailored for Diffusion Transformers
- TaoCache: Structure-Maintained Video Generation Acceleration
- Infinite Worlds with Versatile Interactions
- MineWorld: a Real-Time and Open-Source Interactive World Model on Minecraft
- Matrix-game 2.0: An open-source, real-time, and streaming interactive world model
- Light Interaction: Training-Free Inference Acceleration for Interactive Video World Models
- Yume: An Interactive World Generation Model
- Advancing Open-source World Models
- FORA: Fast-Forward Caching in Diffusion Transformer Acceleration
- WorldPlay: Towards Long-Term Geometric Consistency for Real-Time Interactive World Modeling
- Matrix-Game 3.0: Real-Time and Streaming Interactive World Model with Long-Horizon Memory
- WBench: A Comprehensive Multi-turn Benchmark for Interactive Video World Model Evaluation
- X-Cache: Cross-Chunk Block Caching for Few-Step Autoregressive World Models Inference
- Less is Enough: Training-Free Video Diffusion Acceleration via Runtime-Adaptive Caching
Related papers
- Loss Knows Best: Detecting Annotation Errors in Videos via Loss Trajectories
- AnchorWeave: World-Consistent Video Generation with Retrieved Local Spatial Memories
- Benchmarking the Robustness of Foundation Models for Mammography under Domain Shift
- MambaX-Net: Dual-Input Mamba-Enhanced Cross-Attention Network for Longitudinal MRI Segmentation
- TeleOCR: Navigating Document Parsing Across Digital and Camera-Captured Documents
- A Survey on Efficient Vision-Language-Action Models