Skip to main content
. 2023 Aug 22;7:100574. doi: 10.1016/j.crfs.2023.100574

Table 3.

CNN models included in the comparison.

Model Number of Parameters
DenseNet121 (Huang et al., 2017) 7.0M
EfficientNetV2B0 (Tan and Le, 2021) 5.9M
EfficientNetV2B1 (Tan and Le, 2021) 6.9M
EfficientNetV2B2 (Tan and Le, 2021) 8.8M
EfficientNetV2B3 (Tan and Le, 2021) 12.9M
EfficientNetV2L (Tan and Le, 2021) 117.7M
EfficientNetV2M (Tan and Le, 2021) 53.2M
EfficientNetV2S (Tan and Le, 2021) 20.3M
InceptionResNetV2 (Szegedy et al., 2017) 51.8M
InceptionV3 (Szegedy et al., 2016) 21.8M
MobileNet (Howard et al., 2017) 3.2M
MobileNetV2 (Sandler et al., 2018) 2.3M
MobileNetV3Large (Howard et al., 2019) 3.0M
MobileNetV3Small (Howard et al., 2019) 0.9M
NASNetMobile (Zoph et al., 2018) 4.3M
RegNetX016 (Radosavovic et al., 2020) 7.9M
RegNetX032 (Radosavovic et al., 2020) 14.4M
RegNetY016 (Radosavovic et al., 2020) 9.9M
RegNetY032 (Radosavovic et al., 2020) 18.0M
ResNet50 (He et al., 2016a, He et al., 2016ba) 23.6M
ResNet50V2 (He et al., 2016a, He et al., 2016bb) 23.6M
Xception (Chollet, 2017) 20.9M