Table 2.
Layered architecture of applied deep learning models.
| Models | Convolutional layer | Pooling layer | Dense Layer | Other architectural features | Parameters |
|---|---|---|---|---|---|
| VGG19 | 16 | 5 | 3 | None | 143 M |
| Inception V3 | 48 | 8 | 3 | Batch normalization | 23 M |
| EfficientNetB3 | 9 | 7 | 2 | Squeeze-and-excitation | 6.8 M |
| ResNet152V2 | 152 | 39 | 3 | Residual connections | 60.2 M |
| ResNet50V2 | 50 | 13 | 3 | Residual connections | 25.6 M |
| MobileNetV2 | 18 | 6 | 2 | Inverted residual connections | 3.5 M |
| Xception | 36 | 11 | 3 | Depthwise separable convolutions | 22.9 M |
| Densenet169 | 169 | 5 | 3 | Dense connections | 14.3 M |
| EfficientNetB0 | 5 | 4 | 2 | Squeeze-and-excitation | 5.3 M |
| InceptionResNetV2 | 164 | 41 | 3 | Batch normalization | 55 M |