Damo-yolo retraining patience/early stopping

[Don_Apollo_Cosio]

I’m retraining DAMO-YOLO using the damoyolo_tinynasL20_T.py configuration file and the damoyolo_tinynasL20_T_420.pth pretrained weights. The total number of training epochs is set to 300. I’d like to know if there’s a parameter available to enable early stopping, so that training stops automatically if the model’s performance plateaus, or if DAMO-YOLO strictly follows a fixed training schedule, requiring me to wait until all 300 epochs are completed?