def get_output_cost_deriv(self, outputs, training_values):
"""
Finds the cost of the network
Args:
output(List): A list of output neurons
training_values(List): A list of actual values
Returns:
cost(List): List of costs
Raises:
ValueError: If output length doesn't match training data
"""
if len(outputs) != len(training_values):
raise ValueError(
"Network output length should match training data")
costs = [outputs[i] - training_values[i] for i in range(len(outputs))]
for i in range(len(costs)):
costs[i] *= Layer.deriv_activation(outputs)[i]
return costs
