Accelerating Multi-Condition T2I Generation via Adaptive Condition Offloading and Pruning

2026-05-09

AI Summary

This work addresses the significant latency in multi-conditional text-to-image (T2I) generation caused by high computational and communication overhead from preprocessing and control optimization. To mitigate this, the authors propose an edge-device collaborative architecture featuring a subtask manager that jointly optimizes conditional inference offloading and bandwidth allocation. A lightweight condition contribution assessment mechanism is introduced to enable dynamic condition scaling and redundancy pruning. By integrating heuristic task scheduling, feature activation analysis, and condition overlap evaluation, the system efficiently executes multi-conditional generation under edge-device collaboration. Experimental results demonstrate that the proposed approach reduces generation latency by nearly 25% and improves average generation quality by 6%, substantially outperforming existing baselines.

Abstract

Text-to-image (T2I) generation using multiple conditions enables fine-grained user control on the generated image. Yet, incorporating multi-condition inputs incurs substantial computation and communication overhead, due to additional preprocessing subtasks and control optimizations. It hence leads to unacceptable generation latency. In this paper, we propose an end-edge collaborative system design to accelerate multi-condition T2I generation through adaptive condition offloading and pruning. Extensive offline profiling reveal that, different conditions exhibit significant diversity in computation and communication costs. To this end, we propose a \textit{Subtask Manager} that jointly optimizes condition inference offloading and bandwidth allocation using a heuristic algorithm, balancing local and edge execution delays to minimize overall preprocessing latency. Then, we design a lightweight feature-driven \textit{Conditioning Scale Estimator} that evaluates the contribution of each condition by analyzing its feature activation strength and overlap with other conditions. This allows adaptive conditioning scale selection and pruning of insignificant conditions, thereby accelerating the denoising process. Extensive experimental results show that our system reduces latency by nearly 25\% and improves 6\% average generation quality, outperforming other benchmarks.

Problem

• multi-condition T2I generation
• computation overhead
• communication overhead
• generation latency

Innovation

• adaptive condition offloading
• condition pruning
• multi-condition T2I generation
• edge-cloud collaboration
• feature-driven conditioning