In 'best_point' don't require GeneratorRun to have best_arm_predictions to predict from model #2767
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
Summary: * Add `BenchmarkProblem.get_oracle_experiment_from_params`, a method to compute an experiment where parameters are evaluated at oracle values. This will be useful once we enable inference regret. * Add a helper `get_oracle_experiment_from_experiment` to replicate the old behavior of `get_oracle_experiment`. * Remove `get_opt_trace` from `BenchmarkProblem` and absorbe that logic int `benchmark_replication`; once we have inference regret enabled, how we compute the trace should depend on the _method_, not the problem. The problem should only be responsible for computing oracle values given a parameterization. * Arc lint Differential Revision: D62250058
Summary: Context: There used to be more `BenchmarkProblem` subclasses, and they used to implement their own '__repr__' methods, so there were tests for the custom repr methods. Now `BenchmarkProblem` and its subclass `SurrogateBenchmarkProblem` get their `__repr__` methods from being data classes. These tests have become annoying because they break with any change to `BenchmarkProblem, even if just changing the order of arguments. This PR: * Removes two `test_repr` methods. Differential Revision: D62518032
…ns to predict from model Summary: Context: `get_best_parameters_from_model_predictions_with_trial_index` will only predict from a model if there are `best_arm_predictions` on the `GeneratorRun`. This doesn't make sense, since it's about to construct and fit a new model and use it to generate predicts. Any existing `best_arm_predictions` are not used. This PR: * Removes the `gr.best_arm_predictions is not None` check * Changes how some imported functions are referenced in `best_point_mixin.py` (doesn't change functionality) Reviewed By: mpolson64 Differential Revision: D62594017
facebook-github-bot
added
the
CLA Signed
|
This pull request was exported from Phabricator. Differential Revision: D62594017 |
|
This pull request has been merged in 2fc80f1. |
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.
Summary:
Context:
get_best_parameters_from_model_predictions_with_trial_indexwill only predict from a model if there arebest_arm_predictionson theGeneratorRun. This doesn't make sense, since it's about to construct and fit a new model and use it to generate predicts. Any existingbest_arm_predictionsare not used.This PR:
gr.best_arm_predictions is not Nonecheckbest_point_mixin.py(doesn't change functionality)Reviewed By: mpolson64
Differential Revision: D62594017