Skip to main content
9 events
when toggle format what by license comment
Aug 11, 2020 at 16:05 history edited Sycorax CC BY-SA 4.0
added 55 characters in body
Jul 6, 2020 at 19:09 history edited Sycorax CC BY-SA 4.0
added 113 characters in body
Jul 5, 2020 at 15:32 history edited Sycorax CC BY-SA 4.0
added 12 characters in body
Jul 5, 2020 at 15:27 vote accept Josh
Jul 5, 2020 at 15:22 history edited Sycorax CC BY-SA 4.0
added 202 characters in body
Jul 5, 2020 at 15:20 comment added Sycorax Yes, $p$ is the label and $q$ is the predicted probability of $y=1$. I've added clarification for the second part of your comment.
Jul 5, 2020 at 15:03 comment added Josh Thanks Syrocorax! Great suggestions. And just to complete everything here, in addition to what you wrote above, in a classification setting, when using $H(p,q)$ as your loss, you would define $q$ as the probability output of the network, and $p$ as the ground truth labels, and not the other way around, right? (moreover, this should also fully avoid the $\log\; 0 = -\inf$ problem I mentioned, right?)
Jul 5, 2020 at 13:53 history edited Sycorax CC BY-SA 4.0
added 738 characters in body
Jul 5, 2020 at 13:32 history answered Sycorax CC BY-SA 4.0