Fix PowerLayer gradient check failures by reducing step size#1840
Merged
jeffdonahue merged 1 commit intoBVLC:devfrom Feb 6, 2015
Merged
Fix PowerLayer gradient check failures by reducing step size#1840jeffdonahue merged 1 commit intoBVLC:devfrom
jeffdonahue merged 1 commit intoBVLC:devfrom
Conversation
The gradient checker fails on certain elements of the PowerLayer checks, but only 1-3 sometimes fail out of the 120 elements tested. This is not due to any numerical issue in the PowerLayer, but the distribution of the random inputs for the checks. boost 1.56 switched the normal distribution RNG engine from Box-Muller to Ziggurat.
jeffdonahue
added a commit
that referenced
this pull request
Feb 6, 2015
Fix PowerLayer gradient check failures by reducing step size
Contributor
|
Cool, thanks Evan |
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
See #1252. This reduces the step size of the gradient checker so that the finite-difference approximation holds better. Tests now pass with boost 1.56+, which rewrote the normal distribution RNG -- although the switch to Ziggurat from Box-Muller was supposed to be transparent, it seems it wasn't.
Thanks to @jeffdonahue for the inference to boost RNG since it is always by the fillers regardless of CPU / GPU mode!