Beginning Anomaly Detection Using

Figure 7-48.  The upsampling factor n

4

2   6   7   1   6   9   5

4   4

Figure 7-49.  The first entry in the input is repeated twice to form the first two 

entries in the upsampled output vector

4   

2

6   7   1   6   9   5

4   4   

2   2

Figure 7-50.  The next entry is repeated twice to form the next two entries in the 

output vector of the upsampling operation

Chapter 7   temporal Convolutional networks

---

286

And so on until you finally get Figure 

7-52

.

Going back to the model, each upsampling layer is then connected to a one- 

dimensional convolutional layer, and the pair of upsampling layer and one-dimensional 

convolutional layer repeats again until the final output is passed through a softmax 

function to result in the output/prediction.