def test_waldo(): """ This method will run the test on waldo dataset. """ #dg = waldo(dimensions = (256,256)) # Initialize a dataset creator dg = waldo() # Initialize a dataset creator training_data, training_labels = dg.query_data(samples=1000) testing_data, testing_labels = dg.query_data(samples=1000) #dg._demo() num = [5, 6, 7, 8, 9, 10, 20, 40, 50, 70, 100, 120, 140] for i in range(len(num)): print 'Hidden layers: ', num[i] t0 = time.time() n = mlnn( training_data, training_labels, num[i]) # This call should return a net object that is trained. params = n.get_params( ) # This call should reaturn parameters of the model that are # fully trained. predictions = n.get_predictions( testing_data) # This call should return predictions. acc = accuracy(testing_labels, predictions) print "Accuracy of predictions on waldo data = " + str(acc) + "%" t1 = time.time() print "Time taken: ", t1 - t0 return acc
def test_waldo(): """ This method will run the test on waldo dataset. """ dg = waldo() # Initialize a dataset creator training_data, training_labels = dg.query_data(samples=100) n = mlnn(training_data, training_labels ) # This call should return a net object that is trained. params = n.get_params( ) # This call should reaturn parameters of the model that are # fully trained. testing_data, testing_labels = dg.query_data(samples=100) predictions = n.get_predictions( testing_data) # This call should return predictions. print "Accuracy of predictions on waldo data = " + str( accuracy(testing_labels, predictions)) + "%"