Video Compression Meets Video Generation: Latent Inter-Frame Pruning with Attention Recovery

TL;DR

This paper introduces LIPAR, a framework that reduces video generation latency by pruning redundant latent patches and recovering attention, significantly increasing throughput without quality loss.

Contribution

The paper presents a novel latent inter-frame pruning method with an attention recovery mechanism that enhances real-time video generation efficiency.

Findings

Increases video editing throughput by 1.53 times

Achieves an average of 19.3 FPS on NVIDIA RTX 4090

Maintains generation quality without additional training

Abstract

Current video generation models suffer from high computational latency, making real-time applications prohibitively costly. In this paper, we address this limitation by exploiting the temporal redundancy inherent in video latent patches. To this end, we propose the Latent Inter-frame Pruning with Attention Recovery (LIPAR) framework, which detects and skips recomputing duplicated latent patches. Additionally, we introduce a novel Attention Recovery mechanism that approximates the attention values of pruned tokens, thereby removing visual artifacts arising from naively applying the pruning method. Empirically, our method increases video editing throughput by $1.53\times$, achieving an average of 19.3 FPS on an NVIDIA RTX 4090 with the 1.3B Self-Forcing model (4-step denoising, FP16). The proposed method does not compromise generation quality and can be seamlessly integrated with the model without additional training. Our approach effectively bridges the gap between traditional compression algorithms and modern generative pipelines.