This project uses a neural network to classify images of handwritten numbers into their correct values accurately. The accuracy is about 95%.
The dataset used = MNIST (Modified National Institute of Standards and Technology database) has 60000 training sets and 10000 testing sets.
Keras API is used to define a neural network which is then built by TensorFlow.
Within Keras, the sequential and Dense model is used to specify a simple single input-output sequential layer of the network in the hidden layer The Dense model in Keras ensures that the outputs of all the neurons in a layer are being inputted to the next layer
Activation Function - A function added to obtain the results of non-linear inputs
Ones used here:
ReLU- The rectified linear activation function or ReLU for short is a piecewise linear function that will output the input directly if it is positive, otherwise, it will output zero
Softmax - Calculates the probabilities of occurrences of the output layer and helps in classifying to the max probability class
3 epochs or iterations are run in this project