FedALT: Federated Fine-Tuning through Adaptive Local Training with Rest-of-the-World LoRA