Issue 636 one output node for single-class classification #641
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
we used to use 2 outputs for "binary" classificaiton tasks, but we choose the convention of using 1 output node for 1 class problems going forward.
Sometimes this line seems to be needed, sometimes it doesnt(?) Commenting the line out avoids an error when I run the notebook, though the test in test_cnn.py runs without error with or without the line.
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Outdated suggestions cannot be applied.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.
We choose to use the convention that one-class problems should have a CNN architecture with one output node. In the past, we often used two nodes for such "binary" classification problems ("present" and "absent") with a softmax layer, but this is confusing and doesn't make much sense. Resolves #636 and #495 - in which having two classes for a one-class model caused misleading metrics (eg, averaged precision across the "present" and "absent" classes)