LIPAR: Training-free Latent Inter-Frame Pruning with Attention Recovery
cs.CV
Submitted: 2026-03-06
Updated: 2026-10-31
Project page: https://dennismenn.github.io/lipar
Terminology
Sources
- StreamDiffusionV2: A Streaming System for Dynamic and Interactive Video Generation
- HunyuanVideo: A Systematic Framework For Large Video Generative Models
- VidToMe: Video Token Merging for Zero-Shot Video Editing
- Timestep Embedding Tells: It's Time to Cache for Video Diffusion Model
- Scalable Diffusion Models with Transformers
- MotionStream: Real-Time Video Generation with Interactive Motion Controls
- Wan: Open and Advanced Large-Scale Video Generative Models
- Sparse VideoGen: Accelerating Video Diffusion Transformers with Spatial-Temporal Sparsity
- Sparse VideoGen2: Accelerate Video Generation with Sparse Attention via Semantic-Aware Permutation
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