Minute-Long Videos with Dual Parallelisms

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Wang, Zeqing, Zheng, Bowen, Yang, Xingyi, Tan, Zhenxiong, Xu, Yuecong, Wang, Xinchao
Format: Preprint
Published: 2025
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866918037552103424
author Wang, Zeqing
Zheng, Bowen
Yang, Xingyi
Tan, Zhenxiong
Xu, Yuecong
Wang, Xinchao
author_facet Wang, Zeqing
Zheng, Bowen
Yang, Xingyi
Tan, Zhenxiong
Xu, Yuecong
Wang, Xinchao
contents Diffusion Transformer (DiT)-based video diffusion models generate high-quality videos at scale but incur prohibitive processing latency and memory costs for long videos. To address this, we propose a novel distributed inference strategy, termed DualParal. The core idea is that, instead of generating an entire video on a single GPU, we parallelize both temporal frames and model layers across GPUs. However, a naive implementation of this division faces a key limitation: since diffusion models require synchronized noise levels across frames, this implementation leads to the serialization of original parallelisms. We leverage a block-wise denoising scheme to handle this. Namely, we process a sequence of frame blocks through the pipeline with progressively decreasing noise levels. Each GPU handles a specific block and layer subset while passing previous results to the next GPU, enabling asynchronous computation and communication. To further optimize performance, we incorporate two key enhancements. Firstly, a feature cache is implemented on each GPU to store and reuse features from the prior block as context, minimizing inter-GPU communication and redundant computation. Secondly, we employ a coordinated noise initialization strategy, ensuring globally consistent temporal dynamics by sharing initial noise patterns across GPUs without extra resource costs. Together, these enable fast, artifact-free, and infinitely long video generation. Applied to the latest diffusion transformer video generator, our method efficiently produces 1,025-frame videos with up to 6.54$\times$ lower latency and 1.48$\times$ lower memory cost on 8$\times$RTX 4090 GPUs.
format Preprint
id arxiv_https___arxiv_org_abs_2505_21070
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Minute-Long Videos with Dual Parallelisms
Wang, Zeqing
Zheng, Bowen
Yang, Xingyi
Tan, Zhenxiong
Xu, Yuecong
Wang, Xinchao
Computer Vision and Pattern Recognition
Diffusion Transformer (DiT)-based video diffusion models generate high-quality videos at scale but incur prohibitive processing latency and memory costs for long videos. To address this, we propose a novel distributed inference strategy, termed DualParal. The core idea is that, instead of generating an entire video on a single GPU, we parallelize both temporal frames and model layers across GPUs. However, a naive implementation of this division faces a key limitation: since diffusion models require synchronized noise levels across frames, this implementation leads to the serialization of original parallelisms. We leverage a block-wise denoising scheme to handle this. Namely, we process a sequence of frame blocks through the pipeline with progressively decreasing noise levels. Each GPU handles a specific block and layer subset while passing previous results to the next GPU, enabling asynchronous computation and communication. To further optimize performance, we incorporate two key enhancements. Firstly, a feature cache is implemented on each GPU to store and reuse features from the prior block as context, minimizing inter-GPU communication and redundant computation. Secondly, we employ a coordinated noise initialization strategy, ensuring globally consistent temporal dynamics by sharing initial noise patterns across GPUs without extra resource costs. Together, these enable fast, artifact-free, and infinitely long video generation. Applied to the latest diffusion transformer video generator, our method efficiently produces 1,025-frame videos with up to 6.54$\times$ lower latency and 1.48$\times$ lower memory cost on 8$\times$RTX 4090 GPUs.
title Minute-Long Videos with Dual Parallelisms
topic Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2505.21070