 |
The model successfully learns the training data when the accuracy steadily increases and eventually approaches 99.97%.
After some early swings, the validation accuracy stabilizes at 97.98%, indicating a significant improvement as well as shown in Figure 4.
The model appears to generalize well with unseen data, as seen by the close convergence of training and validation accuracy and the absence of notable overfitting.
The validation loss eventually levels off between 0.1 and 0.2, suggesting that there is little overfitting and that the model performs well in generalization, even on the validation set.
|
IEEE/ICACT20250100 Slide.11
[Big Slide]