| Training Report for Pneumonia CNN Model | |
| ============================================================ | |
| Dataset: chest-xray-pneumonia | |
| Model: Custom CNN (5 Conv Blocks) | |
| Date: 2025-05-25 18:49:57 | |
| Hyperparameters: | |
| - Epochs: 12 | |
| - Batch Size: 32 | |
| - Learning Rate: 0.001 (RMSprop) | |
| - Loss Function: Binary Crossentropy | |
| Number of Parameters: 1,246,401 | |
| Dataset Size: | |
| - Training: 4694 | |
| - Validation: 522 | |
| - Test: 624 | |
| Class Distribution (Train): | |
| - PNEUMONIA: 3487 | |
| - NORMAL: 1207 | |
| Performance Metrics: | |
| - Test Loss: 1.2874 | |
| - Test Accuracy: 63.62% | |
| Classification Report: | |
| precision recall f1-score support | |
| PNEUMONIA 0.99 0.42 0.59 390 | |
| NORMAL 0.51 0.99 0.67 234 | |
| accuracy 0.64 624 | |
| macro avg 0.75 0.71 0.63 624 | |
| weighted avg 0.81 0.64 0.62 624 | |
| ============================================================ |