Large generative models (for example, language and diffusion models) enable high-quality text and image synthesis but are hard to train or adapt in cross-device federated settings due to heavy computation and communication and statistical/system heterogeneity. We propose FedGen-Edge, a framework that decouples a frozen, pre-trained global backbone from lightweight client-side adapters and federates only the adapters. Using Low-Rank Adaptation (LoRA) constrains client updates to a compact subspace, which reduces uplink traffic by more than 99 percent versus full-model FedAvg, stabilizes aggregation under non-IID data, and naturally supports personalization because each client can keep a locally tuned adapter. On language modeling (PTB) and image generation (CIFAR-10), FedGen-Edge achieves lower perplexity/FID and faster convergence than strong baselines while retaining a simple FedAvg-style server. A brief ablation shows diminishing returns beyond moderate LoRA rank and a trade-off between local epochs and client drift. FedGen-Edge offers a practical path toward privacy-preserving, resource-aware, and personalized generative AI on heterogeneous edge devices.

翻译：大型生成模型（例如语言模型和扩散模型）能够实现高质量的文本与图像合成，但由于计算与通信开销巨大以及统计/系统异构性，在跨设备联邦设置中难以训练或适配。我们提出FedGen-Edge框架，该框架将冻结的预训练全局主干网络与轻量级客户端适配器解耦，并仅对适配器进行联邦训练。采用低秩适配（LoRA）技术将客户端更新约束在紧凑子空间内，相比完整模型的FedAvg减少99%以上的上行流量，在非独立同分布数据下稳定聚合过程，并天然支持个性化——每个客户端可保留本地调优的适配器。在语言建模（PTB）与图像生成（CIFAR-10）任务中，FedGen-Edge在保持简单FedAvg式服务器架构的同时，实现了比强基线更低的困惑度/FID值与更快的收敛速度。简要消融实验表明：适度LoRA秩以上收益递减，且本地训练轮数与客户端漂移存在权衡。FedGen-Edge为异构边缘设备上实现隐私保护、资源感知且个性化的生成式AI提供了可行路径。