🍀 Challenge: Training - 3 Layered Neural Network

Problem statement

Train the 3 layered neural network.

• Call the forward_propagation function.
• Call the error function calculate_error and save the loss in each epoch.
• Call the backpropagation function.
• Call the update_parameters function to update the weights and biases in each epoch.

Sample input

• The input features: X
• The target output: y
• The weights of the three layers: w1, w2, and w3 respectively
• The bias of the three layers b1, b2, and b3 respectively
• Total epochs: epochs
• The learning rate: learning_rate

Sample output

• The updated weights and biases at the three layers respectively, i.e., w1, b1, w2, b2,w3, and b3
• The cross-entropy loss in each epoch saved in the losses array

Coding exercise

Write your code below. It is recommended​ to solve the exercise before viewing the solution.

📝 Note: There is a train function given in the code for testing purposes. Do not modify the function signature.

Good luck!🤞