Table 1.

YOLO models training time on the same machine with an identical training dataset. The YOLO model descriptions can be found in Ref. [27] for v5 and in Ref. [24] for v7

Model	Parameters (millions)	Image size (pixels)	Epoch	Total time ($\times {10}^3$s)	Time per epoch
YOLOv5s	7.2	640	1000	15.5	15.5
YOLOv5m	21.2	640	1000	27.1	27.1
YOLOv5l	46.5	640	1000	39.4	39.4
YOLOv5x	86.7	640	1000	50.7	50.7
YOLOv5s6	12.6	1280	1000	17.5	17.5
YOLOv5m6	35.7	1280	1000	30.4	30.4
YOLOv5l6	76.8	1280	1000	44.5	44.5
YOLOv5x6	140.7	1280	1000	54.6	54.6
YOLOv7-tiny	6.2	640	1000	14.4	14.4
YOLOv7	36.9	640	1000	26.7	26.7
YOLOv7-x	71.3	640	1000	45.2	45.2
YOLOv7-w6	70.04	1280	1000	66.7	66.7
YOLOv7-e6	97.2	1280	1000	100.4	100.4
YOLOv7-d6	154.7	1280	1000	140.4	140.4
YOLOv7-e6e	151.7	1280	1000	135.0	135.0