Adjust Hellinger pairwise distance to vaoid NaNs #189
Merged
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
divyegala
approved these changes
Mar 30, 2021
| [=] __device__(value_t input) { return sqrt(1 - input); }, | ||
| [=] __device__(value_t input) { | ||
| // Adjust to replace NaN in sqrt with 0 if input to sqrt is negative | ||
| value_t rectifier = (1 - input) > 0; |
There was a problem hiding this comment.
This should be bool. Just combine this with next line?
cjnolet
requested changes
Mar 30, 2021
cjnolet
added
improvement
|
Even when the inputs are L1 normalized this can arise. Running A solution could be to run csr_row_normalize_l1 and keep this rectifier for negative epsilon, or to add an |
|
@gpucibot merge |
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Outdated suggestions cannot be applied.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.
This change will fix NaNs that arise in Hellinger pairwise distance when the input to sqrt is negative.
These kind of inputs can happen even when the inputs are normalized row-wise to 1 due to accumulation of rounding approximation.