dormouse_model_10-10_01-38-13.keras

Hazel Dormouse Detection Model: dormouse_model_10-10_01-38-13.keras

Training Parameters:
{'input_shape': (224, 224, 3), 'num_classes': 2, 'batch_size': 16, 'epochs': 55, 'augmentation': True, 'early_stopping': True, 'early_stop_patience': 15}

Dataset Size (after augmentation of the train set):

• Train: 1990 images
• Validation: 38 images

Validation Set Performance:

• val_loss: 1.4759
• MaP: 0.6592, MaP@[IoU=50]: 0.8810, MaP@[IoU=75]: 0.7944
• Recall@[max_detections=1]: 0.7121, Recall@[max_detections=10]: 0.7121

Training Metrics:
box_loss: 0.3600 - class_loss: 0.1254 - loss: 0.4854 - val_box_loss: 1.0628 - val_class_loss: 0.4506 - val_loss: 1.4759 - MaP: 0.6592 - MaP@[IoU=50]: 0.8810 - MaP@[IoU=75]: 0.7944 - MaP@[area=small]: 0.0000e+00 - MaP@[area=medium]: 0.5628 - MaP@[area=large]: 0.7118 - Recall@[max_detections=1]: 0.7121 - Recall@[max_detections=10]: 0.7121 - Recall@[max_detections=100]: 0.7121 - Recall@[area=small]: 0.0000e+00 - Recall@[area=medium]: 0.6038 - Recall@[area=large]: 0.7400