Overview

This project aims at classifying the digits in the MNIST dataset into two categories (binary) - zero (label 1) and non-zero (label 2) and multi-class (0 - 9). For the purpose of classification, SVM and CNN (using Keras) have been implemented. Each of them are implemented in different Jupyter notebooks.

Software

• The analysis and implementation has been done using Python 3.6 version.
• The versions of the external libraries used are as mentioned below:
  • NumPy == 1.16.4
  • Scikit-Learn == 0.21.1
  • Matplotlib == 3.0.3

Example