Table 3 Comparison experiments of different models.
From: A new method for Tomicus classification of forest pests based on improved ResNet50 algorithm
Model | Accuracy | Loss | Params | Flops |
|---|---|---|---|---|
ResNet18 | 80.5% | 0.517 | 11.7 M | 1.8G |
ResNet34 | 77.2% | 0.562 | 21.8 M | 3.7G |
ResNet50 | 83.3% | 0.442 | 25.6 M | 4.1G |
ResNet101 | 83.9% | 0.426 | 44.6 M | 7.9G |
MobileNetV2 | 83.2% | 0.441 | 3.5 M | 0.3G |
VGG16 | 85.4% | 0.393 | 138.4 M | 15.5G |
VGG19 | 81.8% | 0.513 | 143.7 M | 19.6G |
VIT-B/16 | 71.9% | 0.891 | 86.6 M | 16.9G |
EffientNetV2 | 76.8% | 0.669 | 21.5 M | 2.9G |
AlexNet | 85.9% | 0.357 | 57.0 M | 0.7G |
DEMNet | 92.8% | 0.220 | 1.6M | 0.7G |
