Using the sonar dataset which contains the frequencies that are reflected by the submarien that will be used to classify whether the object is a rock or mine.

Using Tensorflow created a deep neural network for creating and saving the model(checkpoint). The python code file "rockmine.py" contains comments that are easy to understand what every function does. Uses one hot encoding for the labeled output of the dataset

I have created a multilayer perceptron with 4 hidden layer and 60 neurons each layer that uses gradient descent optimizer to minimize the cost.

The output or the accuracy is as given below. Its not that great but thats what a 1000 epoch can do with a learning rate of 0.3 :)