Update MetadataToParameter.transform_experiment_data to retain the original #4116
Conversation
This file contains hidden or 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
facebook#4115) Summary: If any progression value is missing in `ObservationFeatures.metadata`, `MapKeyToFloat._transform_observation_feature` fills it with the upper bound of the corresponding parameter. This encodes the assumption that any data without a progression value corresponds to a completed trial (as a trial can be completed with `client.complete_trial` without providing progression). The initial implementation of `transform_experiment_data` did not account for the possibility of NaN values in the corresponding index. This diff updates the method to check for NaN progression values and fill them with the target progression. Reviewed By: ltiao Differential Revision: D79750521
…iginal Summary: This transform was popping the corresponding element from the metadata before adding it as a feature. This is fine for the most part, but it doesn't really have a significant benefit over just leaving the original in the metadata. The problem with popping the element from the metadata is when doing this with `ExperimentData`. Even if we `deepcopy` a pandas dataframe, it does not deepcopy the underlying dictionary values. If we then transform the copy, we end up removing the metadata from the original as well, which leads to errors when doing cross validation with an adapter that uses these transforms (since we end up transforming multiple copies, and only the first one has the metadata). This diff updates the transfrom to not remove the element from metadata. Differential Revision: D79805146
meta-cla
bot
added
the
CLA Signed
Contributor
|
This pull request was exported from Phabricator. Differential Revision: D79805146 |
Codecov Report❌ Patch coverage is
Additional details and impacted files@@ Coverage Diff @@
## main #4116 +/- ##
==========================================
- Coverage 96.07% 96.07% -0.01%
==========================================
Files 566 566
Lines 57027 57044 +17
==========================================
+ Hits 54790 54806 +16
- Misses 2237 2238 +1 ☔ View full report in Codecov by Sentry. 🚀 New features to boost your workflow:
|
Contributor
|
This pull request has been merged in e71e598. |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
3 participants
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.
Summary:
This transform was popping the corresponding element from the metadata before adding it as a feature. This is fine for the most part, but it doesn't really have a significant benefit over just leaving the original in the metadata.
The problem with popping the element from the metadata is when doing this with
ExperimentData. Even if wedeepcopya pandas dataframe, it does not deepcopy the underlying dictionary values. If we then transform the copy, we end up removing the metadata from the original as well, which leads to errors when doing cross validation with an adapter that uses these transforms (since we end up transforming multiple copies, and only the first one has the metadata).This diff updates the transfrom to not remove the element from metadata.
Differential Revision: D79805146