Cat-Classifier-Using-Deep-Learning

Cat classifier is a deep learning application built to differentiate between a cat and non-cat images. It is built and tested in three phases with the hardcode implementation of forward propagation and backward propagation . The optimization algorithm used is gradient descent reducing the cost of the function.
1.Implementing with the logistic regression test accuracy : 70.0%
2.Implementing with the two layer neural network the test accuracy improved to 72%.
3.Implementing with the four layer neural network the test accuracy improved to 80%.