Federated learning to train perception model (vision encoder)

Fine-tuning and knowledge distillation to personalize AD-LLM