[MINOR] Update docs for softmax
Project: http://git-wip-us.apache.org/repos/asf/systemml/repo Commit: http://git-wip-us.apache.org/repos/asf/systemml/commit/585b85fe Tree: http://git-wip-us.apache.org/repos/asf/systemml/tree/585b85fe Diff: http://git-wip-us.apache.org/repos/asf/systemml/diff/585b85fe Branch: refs/heads/master Commit: 585b85fe0ff7753b69e8027bf6f8df0a8b594d30 Parents: cbfb21c Author: Mike Dusenberry <[email protected]> Authored: Mon Jun 19 13:52:54 2017 -0700 Committer: Mike Dusenberry <[email protected]> Committed: Mon Jun 19 13:52:54 2017 -0700 ---------------------------------------------------------------------- scripts/nn/layers/softmax.dml | 9 +++++---- 1 file changed, 5 insertions(+), 4 deletions(-) ---------------------------------------------------------------------- http://git-wip-us.apache.org/repos/asf/systemml/blob/585b85fe/scripts/nn/layers/softmax.dml ---------------------------------------------------------------------- diff --git a/scripts/nn/layers/softmax.dml b/scripts/nn/layers/softmax.dml index 68a7bc7..84627a9 100644 --- a/scripts/nn/layers/softmax.dml +++ b/scripts/nn/layers/softmax.dml @@ -26,10 +26,11 @@ forward = function(matrix[double] scores) return (matrix[double] probs) { /* - * Computes the forward pass for a softmax classifier. The inputs - * are interpreted as unnormalized, log-probabilities for each of - * N examples, and the softmax function transforms them to normalized - * probabilities. + * Computes the forward pass for a softmax classifier. The input + * has N examples, each with D values that are interpreted as + * unnormalized, log-probabilities for each of D classes. The softmax + * function transforms these values to normalized probabilities across + * the D classes, for every example. * * This can be interpreted as a generalization of the sigmoid * function to multiple classes.
