🧠 What it Does: Your program learns to recognize handwritten digits (0–9) using deep learning.
🔧 What You Did (Step-by-Step): Loaded the MNIST dataset → A collection of 70,000 handwritten digit images.
Preprocessed the data → Scaled pixel values to 0–1 and reshaped images for the model.
Built a CNN model → A type of deep learning model that works great with images.
Used data augmentation → Slightly altered training images to help the model generalize better.
Trained the model → It learned patterns from training images over 10 rounds (epochs).
Tested the model → Achieved 99.38% accuracy on new, unseen images — excellent!
Visualized performance → Accuracy graph showed how well it learned over time.
Saved the model → So you can use it later without retraining.
✅ Final Result: Your model can now predict handwritten digits with 99.38% accuracy!