Beginning Anomaly Detection Using

7UXH



WUDQVIRUP WUDQVIRUPV7R7HQVRU 

GRZQORDG 

7UXH

WHVWBGDWDVHW WRUFKYLVLRQGDWDVHWV01,67 URRW GDWD 

WUDLQ 

)DOVH



WUDQVIRUP WUDQVIRUPV7R7HQVRU

'DWDORDGHU

WUDLQBORDGHU WRUFKXWLOVGDWD'DWD/RDGHU GDWDVHW WUDLQBGDWDVHW

EDWFKBVL]H EDWFKBVL]H

VKXIIOH 

7UXH

WHVWBORDGHU WRUFKXWLOVGDWD'DWD/RDGHU GDWDVHW WHVWBGDWDVHW

EDWFKBVL]H EDWFKBVL]H

VKXIIOH 

)DOVH

Figure 3-56.  Using DataLoaders, a feature of PyTorch, to get the training and 

testing data

Chapter 3   IntroduCtIon to deep LearnIng

---

114

The procedure for loading the MNIST data might be a bit different in PyTorch, using 

data loaders instead of data frames, but you can still use data frames, arrays, and so on 

in PyTorch after converting them to tensors. The procedure is usually to convert the data 

frame to a numpy array and then to a PyTorch tensor.

Let’s move on to creating your model (Figure 

3-57

).