Table 5 Model complexities for the TrashBox dataset.
From: Enhancing trash classification in smart cities using federated deep learning
Technique | Computational complexity (GMac) | Number of parameters (M) | Model size (MB) |
|---|---|---|---|
ResNeXt50 | 33.53 | 25.04 | 95.94 |
ResNet101 | 61.56 | 44.56 | 170.68 |
ResNet50 | 32.33 | 25.57 | 97.89 |
ResNet34 | 28.82 | 21.8 | 83.33 |
MobileNetV3-Small | 232.6 | 2.54 | 9.83 |
MobileNetV3-Large | 894.64 | 5.48 | 21.11 |
MobileNetV2 | 1.25 | 2.23 | 8.75 |
GoogLeNet | 11.83 | 6.63 | 25.52 |
ResNext | 31.14 | 83.47 | 319.46 |
