some confusion with the several lines of the code

[FrancisLeon]

Dear author,

I am confused about some part of your code, and I will be very grateful if you provide some support.

In train.py line 90-91, I am in fact confused by the dimension operations involved in these three lines.

softmax = np.reshape(softmax, [FLAGS.M, -1, 10])
softmax = np.mean(softmax, 0)
prediction = np.argmax(softmax, 1).flatten()