Table 5 Training performance of all models on different numbers of Epochs
Model | Epochs | Train loss | Valid loss | Training accuracy | Validation accuracy |
|---|---|---|---|---|---|
The proposed FireXnet model | 1 | 1.0284 | 1.0284 | 58.73% | 65.64% |
99 | 0.0252 | 0.0474 | 98.54% | 98.31% | |
100 | 0.0267 | 0.0349 | 98.52% | 98.42% | |
DenseNet-201 | 1 | 1.0973 | 0.5369 | 58.59% | 83.33% |
99 | 0.0416 | 0.0364 | 97.24% | 97.12% | |
100 | 0.0395 | 0.0495 | 97.67% | 97.63% | |
VGG16 | 1 | 1.4186 | 1.1477 | 45.94% | 44.30% |
99 | 0.1531 | 0.1233 | 95.39% | 95.91% | |
100 | 0.1480 | 0.1293 | 95.65% | 96.05% | |
MobileNetV2 | 1 | 1.2676 | 0.5303 | 49.99% | 80.99% |
99 | 0.0412 | 0.0839 | 99.38% | 97.51% | |
100 | 0.0421 | 0.0714 | 99.27% | 97.95% | |
InceptionResNetV2 | 1 | 1.2576 | 0.7130 | 45.72% | 83.92% |
99 | 0.0801 | 0.0664 | 97.37% | 97.66% | |
100 | 0.0992 | 0.0716 | 96.89% | 97.51% | |
InceptionV3 | 1 | 1.1651 | 0.5102 | 52.16% | 90.64% |
99 | 0.0790 | 0.0706 | 97.59% | 97.66% | |
100 | 0.0762 | 0.0719 | 97.51% | 97.51% |