Closed Feather06 closed 1 year ago
Hi, it is a binary classification task with the label "1" being "fake" and "0" being "real." As a result, the larger the predicted logits, the higher the probability that the model assigns to that sample being fake, and the more negative the logits, the higher the probability that the sample is real.
Please re-open if it is still not clear.
So to predict as real or fake, can I sigmoid(logits) and put some threshold say 0.3 to classify as real or fake ?
How does the output logits value judge whether it is real or fake?