Semantically Aligned KV Cache Compression for Efficient Agentic Autoregressive Video Generation
cs.CV, cs.LG
Submitted: 2026-11-08
Updated: 2026-11-08
Terminology
Sources
- MAGI-1: Autoregressive Video Generation at Scale
- SkyReels-V2: Infinite-length Film Generative Model
- Past- and Future-Informed KV Cache Policy with Salience Estimation in Autoregressive Video Diffusion
- LongLive-2.0: An NVFP4 Parallel Infrastructure for Long Video Generation
- CriticalKV: Optimizing KV Cache Eviction from an Output Perturbation Perspective
- NEWTON: Agentic Planning for Physically Grounded Video Generation
- Towards Memory-Efficient Autoregressive Video Generation via Instance-Specific Parametric Absorption
- ViD-GPT: Introducing GPT-style Autoregressive Generation in Video Diffusion Models
- OBCache: Optimal Brain KV Cache Pruning for Efficient Long-Context LLM Inference
- Efficient Autoregressive Video Diffusion with Dummy Head
- End-to-End Training for Autoregressive Video Diffusion via Self-Resampling
- ResidualKV: Residual-Based KV Cache Compression for Efficient Long-Context Inference
- StoryAgent: Customized Storytelling Video Generation via Multi-Agent Collaboration
- ViMax: Agentic Video Generation
- Dynamic Low-Rank Sparse Adaptation for Large Language Models
- CommitKV: Commit-Aligned KV Cache Compression for Multimodal Agents
- Planning with Sketch-Guided Verification for Physics-Aware Video Generation
- Forcing-KV: Hybrid KV Cache Compression for Efficient Autoregressive Video Diffusion Models
- PackCache: A Training-Free Acceleration Method for Unified Autoregressive Video Generation via Compact KV-Cache
- MomentKV: Closing the Directional Gap in KV Cache Eviction for Long-Context Inference
Related papers
- TEA: Text Encoder Alignment for Robust Concept Erasure in Text-to-Image Models
- Do Egocentric Video-Language Models Capture Both Hand- and Object-Centric Cues?
- Aura: Consistent Multi-Subject Video Generation via VLM-Grounded Semantic Alignment
- PVRF: All-in-one Adverse Weather Removal via Prior-modulated and Velocity-constrained Rectified Flow
- DLWM: Diverse Latent World Models for Efficient Multimodal Reasoning
- Out of Sight, Still in Mind: Token Compression for Omni-LLMs