Table 5 Model effect comparison
Model | Accuracy (%) | Flops (G) | Params (M) |
|---|---|---|---|
ResNet50 | 0.87 | 12.087 | 25.560 |
DenseNet121 | 0.88 | 8.121 | 6.960 |
ShuffleNetV1 | 0.86 | 0.127 | 0.348 |
MobileNetV1 | 0.85 | 0.952 | 2.232 |
EfficientB4(teacher) | 0.97 | 4.490 | 17.559 |
Focus-RCNet | 0.90 | 0.418 | 0.525 |
Focus-RCNet-KD | 0.92 | 0.418 | 0.525 |