Skip to content

A simple neural network using TensorFlow/Keras to recognize handwritten digits from the MNIST dataset. Achieves ~97.7% test accuracy. Includes training, evaluation, and visualization. Great intro to deep learning and image classification.

Notifications You must be signed in to change notification settings

dev618/NumNet-Deep-Neural-Network-for-Handwritten-Numbers

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

5 Commits
 
 
 
 
 
 

Repository files navigation

🧠 NumNet – Deep Neural Network for Handwritten Numbers

This project demonstrates how to build and train a simple neural network to recognize handwritten digits from the MNIST dataset using TensorFlow and Keras.


Python TensorFlow Keras NumPy Matplotlib MNIST Deep Learning


Tags:
deep-learning neural-networks tensorflow keras mnist digit-recognition image-classification handwritten-digits machine-learning python numpy matplotlib classification-model ai-project pattern-recognition computer-vision feedforward-network supervised-learning data-preprocessing end-to-end-project portfolio-project


🗂️ Overview

The MNIST dataset is a classic benchmark in the field of machine learning. It contains:

  • 60,000 training images
  • 10,000 test images
    Each image represents a grayscale handwritten digit (0–9) in a 28x28 pixel format.

This project builds a basic feedforward neural network to classify these digits with high accuracy using deep learning techniques.


⚙️ Key Features

✅ Loads and preprocesses the MNIST dataset
✅ Builds a sequential neural network model with:

  • Input layer (flattened 28x28 image)
  • Hidden dense layer with 128 neurons and ReLU activation
  • Output layer with 10 neurons and softmax activation

✅ Trains the model using a validation split
✅ Evaluates the model on test data
✅ Visualizes predictions on sample test images


🏁 Results

After training the model for 10 epochs, it achieves the following:

  • Training Accuracy: ~99.8%
  • Validation Accuracy: ~97.6%
  • Test Accuracy: ~97.7%

🚀 How to Use

1️⃣ Install Required Packages

pip install tensorflow matplotlib numpy

Run the Script
python num_net.py

About

A simple neural network using TensorFlow/Keras to recognize handwritten digits from the MNIST dataset. Achieves ~97.7% test accuracy. Includes training, evaluation, and visualization. Great intro to deep learning and image classification.

Topics

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published