Fix issues in maxout layer #22031
Merged
Fix issues in maxout layer #22031
Conversation
This file contains 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
martinwicke
approved these changes
Sep 5, 2018
martinwicke
added
kokoro:force-run
tensorflow-copybara
pushed a commit
that referenced
this pull request
Sep 11, 2018
PiperOrigin-RevId: 212399243
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
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Outdated suggestions cannot be applied.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.
Originally pointed out by @ilblackdragon in this pull request:
shape[axis] should = num_units instead of = -1.
The original implementation causes some trouble. For example, if I am to maxout a tensor x = (?, 32, 32, 256) and reduce it into a tensor of shape (?, 32, 32, 128), by using
x = tf.contrib.layers.maxout(x, 128), I get a tensor of shape (?, 32, 32, ?) instead of (?, 32, 32, 128).The shape of the tensor after the maxout can be inferred with num_units, so there is no need to use -1 here. This also causes inconvenience and raises errors when using a dense layer afterwards (or any layer that needs the value of the last dimension).
In addition, the original documentation is problematic in saying that "num_units should be a multiple of 'axis'", and this PR fixes the description.