🧠 Handwritten Digit Recognizer using CNN & MNIST

This project is a Convolutional Neural Network (CNN) based model built to recognize handwritten digits (0–9) using the MNIST dataset. The model was trained and tested using Google Colab, and saved in .h5 format.

📁 Files Included

• Handwritten_Digit_Recognizer.ipynb — Jupyter notebook containing all code (data loading, preprocessing, model training, evaluation).
• mnist_digit_recognizer.h5 — Trained CNN model saved in HDF5 format.
• README.md — Description and overview of the project.

---

🔍 Project Overview

📌 Objective

To build a machine learning model that accurately classifies handwritten digits from images using a CNN and the MNIST dataset.

🧰 Tools & Libraries Used

• Python
• TensorFlow / Keras
• NumPy, Matplotlib
• Google Colab

---

📊 Dataset

• MNIST Dataset from TensorFlow datasets
• 60,000 training images and 10,000 test images of digits (0–9)
• Each image: 28x28 grayscale

---

🧠 Model Architecture

• Input Layer: 28x28 grayscale image
• Conv2D Layer: 32 filters (3x3) + ReLU
• MaxPooling2D
• Conv2D Layer: 64 filters (3x3) + ReLU
• MaxPooling2D
• Flatten
• Dense Layer: 128 neurons + ReLU
• Dropout: 0.2
• Output Layer: 10 neurons (softmax)

---

🏁 Training Summary

• Epochs: 5
• Batch Size: 64
• Final Accuracy:
  • Training Accuracy: ~99.4%
  • Validation Accuracy: ~99.0%

---

✅ How to Use

1. Clone the repo or download the notebook.
2. Open the .ipynb file in Google Colab or Jupyter.
3. Run all cells to train or test the model.
4. Or load the .h5 model for prediction.

from keras.models import load_model
model = load_model('mnist_digit_recognizer.h5')

---

📷 Screenshots

🧪 Training Output

🎯 Sample Predictions

📓 Notebook View

---

📌 Author

Vidit Kumar
Project completed as part of internship with Codec Technologies India.

---

🔗 Links

• Project on Google Colab
• Trained Model (.h5)