Table 3 Different training hyperparameters for the proposed CNN models.
From: Automated multi-model framework for malaria detection using deep learning and feature fusion
Hyperparameter | Value | Description |
|---|---|---|
Optimizer | SGDM | Stochastic Gradient Descent with Momentum |
Execution Environment | Parallel | Utilizes multiple CPU/GPU cores for faster computation |
Mini-Batch Size | 64 | Number of samples processed per training iteration |
Maximum Epochs | 30 | Number of complete passes through the training dataset |
Activation Function | ReLU | It introduces non-linearity and helps accelerate convergence by allowing models to learn complex patterns efficiently. |
Initial Learning Rate | 1.00E-05 | Starting value for learning rate |
Learning Rate Schedule | piecewise | Learning rate is adjusted at pre-defined intervals |
Validation Frequency | 50 | Validation is performed every 50 iterations |
