Target property value #170
Closed
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
This was referenced Feb 15, 2024
|
Closing since this code has been added in PR #172 |
kspieks
added a commit
that referenced
this pull request
Feb 15, 2024
This PR supersedes #165 and #170 and resolves #168. #165 implemented a sampler which would divide molecules based on their molecular weight (either ascending or descending), and #170 added basically the same thing but for arbitrary target values. This PR just refactors #170 as if #165 where already implemented, reducing code duplication quite a bit. There is also some small internal cleanup, namely 00eeca6, 98b3efd, and 3b8bca1 (as well as some long-overdue CI updates).
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Outdated suggestions cannot be applied.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.
As suggested in issue #168, this PR introduces support for an extrapolative sampler that splits the data by target property. A flag controls whether the testing set contains data points with the largest or smallest target y values.
To document a few general notes,
• This sampler sorts the indices by property value. Since no clusters are created, the
TargetPropertyclass actually behaves as an interpolation sampler behind the scenes when it's called in main.py.• The other consequence of sorting is that the results are always deterministic. This sampler does not use the
random_stateargument. If users want to obtain multiple splits, they can vary the amount of data that is assigned to the training vs. testing setLet me know if I missed anything. Looking forward to your feedback!