MNIST c-GAN

A conditional Generative Adversarial Network to generate a specific "handwritten"(I know computers do not have hands) digit.

Dataset

The dataset is the famous MNIST - Handwritten Digit Classification dataset that comes built-in the tensorflow.keras.datasets module.

Dataset Analysis

Requirements

• Numpy 1.18.3
• Matplotlib 3.2.1
• Scikit-learn 0.22.2
• Tensorflow 2.2
• OpenCV 4.2

Model Architecture

Generator

Discriminator

Adversarial

Model metrics

Model Output

File Configuration

• main.ipynb - Jupyter notebook which has the analysis of dataset along with the declaration and training of the model
• resources/ - Github README resources
• generator.py - Generates the image with a CLI (Run it you will find out)

Author

Sagnik Sarkar -