Table 2 Comparison of average Accuracy, Precision, Recall, F1 score, Times, Flop and Params of different network models.
From: Lightweight grape leaf disease recognition method based on transformer framework
Methods | Accuracy | Precision | Recall | F1 score | Top Acc | Flop(G) | Params (M) | Times (ms) |
|---|---|---|---|---|---|---|---|---|
ConvNext V2 | 95.89 | 96.02 | 95.98 | 95.97 | 96.41 | 4.45 | 27.79` | 28 |
InceptionNext | 93.81 | 92.75 | 92.73 | 92.72 | 95.74 | 4.20 | 28.04 | 32 |
DenseNet121 | 98.30 | 98.06 | 98.06 | 98.04 | 99.71 | 2.83 | 7.89 | 37 |
ResNet50 | 97.71 | 97.70 | 97.69 | 97.65 | 99.75 | 4.13 | 25.55 | 14 |
EfficientNet V2 | 95.73 | 95.59 | 95.52 | 95.39 | 98.16 | 2.85 | 21.305 | 26 |
Mobilenet V4 | 88.16 | 88.14 | 88.05 | 87.50 | 93.78 | 0.18 | 2.46 | 6.9 |
GhostNetV2 | 91.55 | 94.47 | 91.61 | 91.46 | 94.07 | 0.18 | 4.87 | 23 |
Deit3 | 95.95 | 95.92 | 95.89 | 95.82 | 98.62 | 4.24 | 21.97 | 7 |
EfficientFormer V2 | 96.82 | 96.11 | 96.09 | 96.06 | 99.04 | 1.23 | 12.63 | 23 |
MobileVit V2 | 93.45 | 92.64 | 92.58 | 92.55 | 96.20 | 1.41 | 4.87 | 14 |
SwinTransformer V2 | 96.29 | 96.29 | 96.23 | 96.18 | 98.25 | 4.51 | 28.33 | 17 |
TinyVit | 96.20 | 95.54 | 95.43 | 95.46 | 97.70 | 1.19 | 12.07 | 10 |
CaitNet | 93.43 | 92.89 | 93.44 | 92.87 | 98.54 | 8.63 | 46.82 | 15 |
MvitV2 | 80.86 | 80.56 | 80.58 | 80.31 | 89.97 | 3.97 | 24.07 | 17 |
DLVTNet | 98.48 | 98.48 | 98.47 | 98.46 | 99.79 | 0.49 | 1.05 | 8 |