Let the model consist of three fully connected layers (for example, keras): 
You need to train this model in dataset, and then use two models with the same weights as follows: 
Where Xa and Xb - data from two different classes How to implement it in tensorflow?
Depth Depth
26 one eleven
|