Table 11.

Evaluation results of the GTSRB data set. The results are grouped by network sizes in FLOPs. The first group contains larger networks, with the second group showing comparatively smaller models.

Model	Widening Factor k	Mean Acc	kFLOPs
EffNet V2	0.3	90.4%	704.5
MobileNetV2	0.31	90.74%	710.7
MobileNet	0.23	88.15%	533.0
ShuffleNet	0.23	88.99%	540.7
IoTNet-3-2	0.22	93.17%	531.0
EffNet V1	0.15	91.79%	344.1
IoTNet-3-2	0.18	88.82%	301.6