Python Deep Learning Projects

Evaluating the model

Now, evaluate the convolution model with max pooling on the test data:

# evaluate modelloss, acc = model.evaluate(X_test, y_test)print('Test loss:', loss)print('Accuracy:', acc)

The following is the output of the preceding code:

Figure 8.21: Printout of the evaluation of the convolution classifier with max pooling

We can see that the model is 97.88% accurate on the test data, 97.72% on the validation data, and 99.74% on the train data. The convolution model with pooling gives the same level of performance as the convolution model without pooling, but with four times less parameters.

In this case, we can clearly see from ...