Leaf Disease detection

Hi,

I need help finalizing my plant disease detection project.

Current setup:

- Model: CNN + MobileNetV2 (already trained)
- Model file: .keras format available
- API is already working and giving predictions correctly
- Frontend is also working

What I need you to do:

1. Create an evaluate.py script that:

- Loads the trained model (.keras file)
- Loads dataset (PlantVillage / similar structured dataset)
- Generates Confusion Matrix
- Generates Classification Report (precision, recall, f1-score)

2. Ensure:

- Correct class label mapping (no mismatch between dataset and model)
- No changes in existing API behavior/output
- Dataset should be properly loaded using ImageDataGenerator or tf.data

3. Important:

- Current prediction output from API should remain unchanged
- Evaluation should be done separately (do not break API)

4. Optional (if possible):

- Plot confusion matrix using matplotlib/seaborn
- Print class names clearly

Issue currently faced:

- Some confusion between similar diseases (e.g., Blight vs Septoria)
- Need evaluation metrics to demonstrate model performance

Goal:

- Make the project ready for final submission with proper evaluation (confusion matrix + report)
- Keep system stable and working

Let me know if you need dataset structure or model details.