Update README.md

README.md

@@ -224,32 +224,12 @@ The model was trained in two stages:
 1. Classifier training with the EfficientNet-B0 backbone frozen.
 2. Fine-tuning of the final feature blocks.
 
-### Phase 1
-
-![Phase 1 Training Loss](training/phase1_loss.png)
-
-### Phase 2
-
-![Phase 2 Fine-Tuning Loss](training/phase2_loss.png)
+| Phase 1 Loss | Phase 2 Loss |
+| :---: | :---: |
+| ![Phase 1](training/phase1_loss.png) | ![Phase 2](training/phase2_loss.png) |
 
 The epoch-level training history is available in
 `training/training_history.json`.
----
-
-## Explainability
-
-Grad-CAM visualizations are provided to understand which image regions contribute most strongly to predictions.
-
-Examples include:
-
-- Correct HSRP predictions
-- Correct Non-HSRP predictions
-- False Positives
-- False Negatives
-
-Location:
-
-explainability/
 
 ---
 
@@ -302,10 +282,8 @@ evaluation/
 
 training/
 ├── training_history.json
-└── training_config.yaml
-
-explainability/
-└── gradcam_examples/
+└── phase1_loss.png
+└── phase2_loss.png
 
 ```
 ---
@@ -327,10 +305,3 @@ Its sole responsibility is determining whether a cropped license plate image bel
 - Binary HSRP classification
 
 ---
-
-## Author
-
-Aniket Behera
-
-Computer Engineering
-Smart HSRP Detection & Violation Monitoring System