Fix #20: long times break with Substack's Gamma #21
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
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Outdated suggestions cannot be applied.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.
For quizzes much longer than a model halflife, Substack's Gamma can return NaNs during
_updateRecallSingle😭(Note this doesn't happen with stdlib's more accurate Gamma function 💪!)
This PR fixes this by switching to the log domain if the default linear-domain calculation fails to converge. For the examples in #20, as well as much more extreme examples (per the new unit test, when the quiz is 1e5 times the halflife!), this is sufficient for the update step to converge.
New minor release version is included but needs to be rebuilt before release.