Beginning Anomaly Detection Using

LPSRUW

WHQVRUIORZ

DV

WI

LPSRUW

NHUDV

IURP

NHUDVGDWDVHWV

LPSRUW

PQLVW

IURP

NHUDVPRGHOV

LPSRUW

6HTXHQWLDO

IURP

NHUDVOD\HUV

LPSRUW

'HQVH'URSRXW

)ODWWHQ,QSXW

IURP

NHUDVOD\HUV

LPSRUW

&RQY'0D[3RROLQJ'

IURP

NHUDV

LPSRUW

EDFNHQG

DV

.

LPSRUW

QXPS\

DV

QS

Figure 3-14.  Importing the modules needed to create the model

Chapter 3   IntroduCtIon to deep LearnIng

---

86

Now define some variables that you will use later (Figure 

3-15

).

One pass of the entire data set through the model is called an 

epoch. The batch 

size is how many data entries pass through the model in one iteration. In this case, the 

training data passes through the model 128 entries at a time until all of the entries have 

passed through, marking the end of one epoch. The number of classes is 10 to represent 

each of the 10 digits from 0-9. These variables are also known as