Salvato in:
Dettagli Bibliografici
Autori principali: Wei, Yi, Tang, Shunpu, Zhao, Liang, Yang, Qiangian
Natura: Preprint
Pubblicazione: 2025
Soggetti:
Accesso online:https://arxiv.org/abs/2510.16326
Tags: Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
_version_ 1866917024301580288
author Wei, Yi
Tang, Shunpu
Zhao, Liang
Yang, Qiangian
author_facet Wei, Yi
Tang, Shunpu
Zhao, Liang
Yang, Qiangian
contents Recent advances in diffusion models have driven remarkable progress in image generation. However, the generation process remains computationally intensive, and users often need to iteratively refine prompts to achieve the desired results, further increasing latency and placing a heavy burden on cloud resources. To address this challenge, we propose DiffusionX, a cloud-edge collaborative framework for efficient multi-round, prompt-based generation. In this system, a lightweight on-device diffusion model interacts with users by rapidly producing preview images, while a high-capacity cloud model performs final refinements after the prompt is finalized. We further introduce a noise level predictor that dynamically balances the computation load, optimizing the trade-off between latency and cloud workload. Experiments show that DiffusionX reduces average generation time by 15.8% compared with Stable Diffusion v1.5, while maintaining comparable image quality. Moreover, it is only 0.9% slower than Tiny-SD with significantly improved image quality, thereby demonstrating efficiency and scalability with minimal overhead.
format Preprint
id arxiv_https___arxiv_org_abs_2510_16326
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle DiffusionX: Efficient Edge-Cloud Collaborative Image Generation with Multi-Round Prompt Evolution
Wei, Yi
Tang, Shunpu
Zhao, Liang
Yang, Qiangian
Computer Vision and Pattern Recognition
Machine Learning
Recent advances in diffusion models have driven remarkable progress in image generation. However, the generation process remains computationally intensive, and users often need to iteratively refine prompts to achieve the desired results, further increasing latency and placing a heavy burden on cloud resources. To address this challenge, we propose DiffusionX, a cloud-edge collaborative framework for efficient multi-round, prompt-based generation. In this system, a lightweight on-device diffusion model interacts with users by rapidly producing preview images, while a high-capacity cloud model performs final refinements after the prompt is finalized. We further introduce a noise level predictor that dynamically balances the computation load, optimizing the trade-off between latency and cloud workload. Experiments show that DiffusionX reduces average generation time by 15.8% compared with Stable Diffusion v1.5, while maintaining comparable image quality. Moreover, it is only 0.9% slower than Tiny-SD with significantly improved image quality, thereby demonstrating efficiency and scalability with minimal overhead.
title DiffusionX: Efficient Edge-Cloud Collaborative Image Generation with Multi-Round Prompt Evolution
topic Computer Vision and Pattern Recognition
Machine Learning
url https://arxiv.org/abs/2510.16326