Federated learning (FL) has recently been used to collaboratively fine-tune foundation models (FMs) across multiple clients. Notably, federated low-rank adaptation (LoRA)-based fine-tuning methods ...