why do not apply softmax before predicting class?

Question

MaybeShewill-CV opened this issue 5 years ago · 1 comments

HI I wonder why don't you apply softmax on the inference logits before you decode the predicted class in file https://github.com/rishizek/tensorflow-deeplab-v3/blob/master/deeplab_model.py line 159~161. Thanks for your reply

Answer 1 · 2019-05-31T07:03:03.000Z

@rishizek Thanks I have solved this problem:)