Table 1 Common fine-tuning parameters and their corresponding values for all three candidate models.
From: A fuzzy rank-based deep ensemble methodology for multi-class skin cancer classification
Category | Parameter | Value |
|---|---|---|
Model Architecture | pre-trained weights | Imagenet |
| Â | input shape | (96, 96, 3) |
| Â | batch size | 32 |
| Â | layer.trainable | True |
Model Compilation | optimizer | Adam |
| Â | loss function | Sparse_categorical_crossentropy |
ReduceLROnPlateau | monitor | Val_loss |
| Â | factor | 0.1 |
| Â | patience | 5 |
| Â | verbose | 1 |
EarlyStopping | monitor | Val_loss |
| Â | patience | 20 |
| Â | verbose | 1 |
| Â | restore_best_weights | True |
