Questions about Logq(y|hL)

Hi!
I'm reading the code of your paper and I meet a question about the calculation of  Logq(y|h_L) in equation(9). In the code, you calculate the cross-entropy of the outputs and the labels and use this cross-entropy to represent the average of  Logq(y|h_L) . But I can't understand why they are equal. Could you explain this for me?
Looking forward to your reply! Thank you!