Table 2 The performance between ResNet-50 and revised ResNet-50
From: Development of revised ResNet-50 for diabetic retinopathy detection
Model/performance | ResNet-50 | Revised ResNet-50 |
|---|---|---|
Train accuracy | 0.8981 | 0.8395 |
Train loss | 0.4805 | 0.5287 |
Validation accuracy | 0.7145 | 0.7416 |
Validation loss | 0.6603 | 0.6155 |
Test accuracy | 0.7567 | 0.7432 |
Test loss | 0.6474 | 0.6018 |