Table 2.

Metrics of the learning process.

Parameters	Precision	Recall	F1-Score	mAPval@0.5	mAPval@0.5:0.95
Yolov5 (78 epochs, 288 labels, val. labels 99)	0.98	0.99	0.985	0.993	0.689
Yolov5 (79 epochs, 296 labels)	0.98	1	0.989	0.993	0.721
Yolov5 (80 epochs, 244 labels)	0.98	0.99	0.985	0.993	0.686
Yolov5 (100 epochs, 345 labels)	0.98	0.98	0.978	0.993	0.708

Note: labels—the number of rivets on the images of plate fragments that participated in the training of the epoch; val. labels—the number of rivets on the images that participated in the validation.