Since we have fine-tuned the model by training our custom dense layers on our data, let us check the performance of the model.
Use evaluate method on the vgg_model to get the accuracy of its performance on the test data.
vgg_model_loss, vgg_model_acc = vgg_model.<< your code comes here >>(test_set_x_orig,test_set_y_orig)
Print the accuracy vgg_model_acc.
print('Test accuracy using VGG16 model as the base:', vgg_model_acc)
You could play around with different value of hyper-parameters, like the learning rate, number of epochs, number of dropout neurons, dropout value, etc, and get a better model which yields better performance.
Note - Having trouble with the assessment engine? Follow the steps listed here
Loading comments...