Beginning Anomaly Detection Using

Figure 3-31.  Print the shapes of the transformed data

Figure 3-32.  The resulting output

Chapter 3   IntroduCtIon to deep LearnIng

---

94

Note  the \ character tells python that you want to continue to the next line. 

Without it, the code would not run because python doesn’t see the end of the string 

denoted by the second “, but what \ tells python is to continue on the next line.

Now you can move on to defining and compiling your model.

Run the code in Figure 

3-33

.

PRGHO 6HTXHQWLDO

PRGHODGG &RQY' NHUQHOBVL]H  

DFWLYDWLRQ UHOX 

LQSXWBVKDSH LQSXWBVKDSH

PRGHODGG &RQY'   

DFWLYDWLRQ UHOX

PRGHODGG 0D[3RROLQJ' SRROBVL]H 

PRGHODGG 'URSRXW 

PRGHODGG )ODWWHQ

PRGHODGG 'HQVH DFWLYDWLRQ UHOX

PRGHODGG 'URSRXW 

PRGHODGG 'HQVH QBFODVVHV

DFWLYDWLRQ VRIWPD[

PRGHOFRPSLOH ORVV NHUDVORVVHVFDWHJRULFDOB

FURVVHQWURS\

RSWLPL]HU NHUDVRSWLPL]HUV$GDP 

PHWULFV > DFFXUDF\ @

PRGHOVXPPDU\

Figure 3-33.  Code to define a deep learning model and add layers to it

Chapter 3   IntroduCtIon to deep LearnIng

---

95

In Keras, the 

sequential model is a stack of layers. The Conv2D is a two-dimensional 

convolutional layer.

In convolutional neural networks, a