This contains implementation of Handwritten Digit Recognition Models trained on MNIST dataset
Objective : To train various classifiers/models using the following methods to classify the image sample given by MNIST dataset into one of the 10 possible classes (0 to 9) : 1. Logistic Regression 2. Multi Layer Perceptron 3. Deep Neural Network 4. Deep Convolutional Neural Network
Further, Exercise on various parameters (if applicable) such as number of hidden layers, type of activation functions, number of convolution kernels, size of convolution kernels. Also exercise on over-fitting problems with possible solutions.