Get rid of SurrogateRunner and SurrogateBenchmarkProblem #2954
+13
−303
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
facebook-github-bot
added
the
CLA Signed
|
This pull request was exported from Phabricator. Differential Revision: D64855010 |
|
This pull request was exported from Phabricator. Differential Revision: D64855010 |
esantorella
added a commit
to esantorella/Ax
that referenced
this pull request
Oct 24, 2024
Summary: Pull Request resolved: facebook#2954 **Context**: This diff cuts over over uses of `SurrogateRunner` to use `ParamBasedTestProblemRunner` with a `test_problem` that is the newly introduced `SurrogateTestFunction`, and the following diff after that will bring us down to only one runner class for benchmarking by merging `ParamBasedTestProblemRunner` into `BenchmarkRunner`. Having only one runner will make it easier to enable asynchronous benchmarks. Currently, SurrogateRunner had its own logic for tracking when trials are completed, which would make it difficult to work in with asynchronicity. **Note on naming**: Some names have become non-intuitive in the process of benchmarking. To accord with some future changes I hope to make, I called a new class SurrogateTestFunction, whereas SurrogateParamBasedTestProblem would be more in line with the old naming. The name changes I hope to make: * ParamBasedTestProblemRunner -> nothing, absorbed into BenchmarkRunner * ParamBasedTestProblem -> TestFunction, to emphasize that all it does is generate data (rather than more generally specify the problem we are solving) and that it is deterministic, and to differentiate it from BenchmarkProblem. BenchmarkTestFunction would also be a candidate. * BoTorchTestProblem -> BoTorchTestFunction **Changes in this diff**: * Introduces SurrogateTestFunction, a ParamBasedTestProblem for * Removes SurrogateRunner. There is some loss of (unused) functionality here; dict-valued 'noise_std' is no longer supported. I expect we will eventually want that, but it doesn't need to be done now. * Changed `predict_for_tensor` utility function to use a `SurrogateTestFunction` rather than a `SurrogateRunner` * Removed `SurrogateRunner` from the json_store registry. There is no point in keeping it around for backward compatibility since it was just deserializing to a dummy `SyntheticRunner` anyway. Differential Revision: D64855010
esantorella
force-pushed
the
export-D64855010
branch
2 times, most recently
from
f8bb4ae to
a7434fd
Compare
esantorella
added a commit
to esantorella/Ax
that referenced
this pull request
Oct 25, 2024
Summary: **Context**: This diff cuts over over uses of `SurrogateRunner` to use `ParamBasedTestProblemRunner` with a `test_problem` that is the newly introduced `SurrogateTestFunction`, and the following diff after that will bring us down to only one runner class for benchmarking by merging `ParamBasedTestProblemRunner` into `BenchmarkRunner`. Having only one runner will make it easier to enable asynchronous benchmarks. Currently, SurrogateRunner had its own logic for tracking when trials are completed, which would make it difficult to work in with asynchronicity. **Note on naming**: Some names have become non-intuitive in the process of benchmarking. To accord with some future changes I hope to make, I called a new class SurrogateTestFunction, whereas SurrogateParamBasedTestProblem would be more in line with the old naming. The name changes I hope to make: * ParamBasedTestProblemRunner -> nothing, absorbed into BenchmarkRunner * ParamBasedTestProblem -> TestFunction, to emphasize that all it does is generate data (rather than more generally specify the problem we are solving) and that it is deterministic, and to differentiate it from BenchmarkProblem. BenchmarkTestFunction would also be a candidate. * BoTorchTestProblem -> BoTorchTestFunction **Changes in this diff**: * Introduces SurrogateTestFunction, a ParamBasedTestProblem for * Removes SurrogateRunner. There is some loss of (unused) functionality here; dict-valued 'noise_std' is no longer supported. I expect we will eventually want that, but it doesn't need to be done now. * Changed `predict_for_tensor` utility function to use a `SurrogateTestFunction` rather than a `SurrogateRunner` * Removed `SurrogateRunner` from the json_store registry. There is no point in keeping it around for backward compatibility since it was just deserializing to a dummy `SyntheticRunner` anyway. Differential Revision: D64855010
|
This pull request was exported from Phabricator. Differential Revision: D64855010 |
Codecov ReportAll modified and coverable lines are covered by tests ✅
Additional details and impacted files@@ Coverage Diff @@
## main #2954 +/- ##
========================================
Coverage 95.66% 95.66%
========================================
Files 486 485 -1
Lines 48944 48832 -112
========================================
- Hits 46821 46716 -105
+ Misses 2123 2116 -7 ☔ View full report in Codecov by Sentry. |
esantorella
added a commit
to esantorella/Ax
that referenced
this pull request
Oct 25, 2024
Summary: Pull Request resolved: facebook#2954 **Context**: This diff cuts over over uses of `SurrogateRunner` to use `ParamBasedTestProblemRunner` with a `test_problem` that is the newly introduced `SurrogateTestFunction`, and the following diff after that will bring us down to only one runner class for benchmarking by merging `ParamBasedTestProblemRunner` into `BenchmarkRunner`. Having only one runner will make it easier to enable asynchronous benchmarks. Currently, SurrogateRunner had its own logic for tracking when trials are completed, which would make it difficult to work in with asynchronicity. **Note on naming**: Some names have become non-intuitive in the process of benchmarking. To accord with some future changes I hope to make, I called a new class SurrogateTestFunction, whereas SurrogateParamBasedTestProblem would be more in line with the old naming. The name changes I hope to make: * ParamBasedTestProblemRunner -> nothing, absorbed into BenchmarkRunner * ParamBasedTestProblem -> TestFunction, to emphasize that all it does is generate data (rather than more generally specify the problem we are solving) and that it is deterministic, and to differentiate it from BenchmarkProblem. BenchmarkTestFunction would also be a candidate. * BoTorchTestProblem -> BoTorchTestFunction **Changes in this diff**: * Introduces SurrogateTestFunction, a ParamBasedTestProblem for * Removes SurrogateRunner. There is some loss of (unused) functionality here; dict-valued 'noise_std' is no longer supported. I expect we will eventually want that, but it doesn't need to be done now. * Changed `predict_for_tensor` utility function to use a `SurrogateTestFunction` rather than a `SurrogateRunner` * Removed `SurrogateRunner` from the json_store registry. There is no point in keeping it around for backward compatibility since it was just deserializing to a dummy `SyntheticRunner` anyway. Differential Revision: D64855010
esantorella
force-pushed
the
export-D64855010
branch
from
a7434fd to
027cfba
Compare
esantorella
added a commit
to esantorella/Ax
that referenced
this pull request
Oct 25, 2024
Summary: **Context**: This diff cuts over over uses of `SurrogateRunner` to use `ParamBasedTestProblemRunner` with a `test_problem` that is the newly introduced `SurrogateTestFunction`, and the following diff after that will bring us down to only one runner class for benchmarking by merging `ParamBasedTestProblemRunner` into `BenchmarkRunner`. Having only one runner will make it easier to enable asynchronous benchmarks. Currently, SurrogateRunner had its own logic for tracking when trials are completed, which would make it difficult to work in with asynchronicity. **Note on naming**: Some names have become non-intuitive in the process of benchmarking. To accord with some future changes I hope to make, I called a new class SurrogateTestFunction, whereas SurrogateParamBasedTestProblem would be more in line with the old naming. The name changes I hope to make: * ParamBasedTestProblemRunner -> nothing, absorbed into BenchmarkRunner * ParamBasedTestProblem -> TestFunction, to emphasize that all it does is generate data (rather than more generally specify the problem we are solving) and that it is deterministic, and to differentiate it from BenchmarkProblem. BenchmarkTestFunction would also be a candidate. * BoTorchTestProblem -> BoTorchTestFunction **Changes in this diff**: * Introduces SurrogateTestFunction, a ParamBasedTestProblem for surrogates * Removes SurrogateRunner. There is some loss of (unused) functionality here; dict-valued 'noise_std' is no longer supported. I expect we will eventually want that, but it doesn't need to be done now. * Changed `predict_for_tensor` utility function to use a `SurrogateTestFunction` rather than a `SurrogateRunner` * Removed `SurrogateRunner` from the json_store registry. There is no point in keeping it around for backward compatibility since it was just deserializing to a dummy `SyntheticRunner` anyway. Differential Revision: D64855010
|
This pull request was exported from Phabricator. Differential Revision: D64855010 |
esantorella
added a commit
to esantorella/Ax
that referenced
this pull request
Oct 25, 2024
Summary: **Context**: This diff cuts over over uses of `SurrogateRunner` to use `ParamBasedTestProblemRunner` with a `test_problem` that is the newly introduced `SurrogateTestFunction`, and the following diff after that will bring us down to only one runner class for benchmarking by merging `ParamBasedTestProblemRunner` into `BenchmarkRunner`. Having only one runner will make it easier to enable asynchronous benchmarks. Currently, SurrogateRunner had its own logic for tracking when trials are completed, which would make it difficult to work in with asynchronicity. **Note on naming**: Some names have become non-intuitive in the process of benchmarking. To accord with some future changes I hope to make, I called a new class SurrogateTestFunction, whereas SurrogateParamBasedTestProblem would be more in line with the old naming. The name changes I hope to make: * ParamBasedTestProblemRunner -> nothing, absorbed into BenchmarkRunner * ParamBasedTestProblem -> TestFunction, to emphasize that all it does is generate data (rather than more generally specify the problem we are solving) and that it is deterministic, and to differentiate it from BenchmarkProblem. BenchmarkTestFunction would also be a candidate. * BoTorchTestProblem -> BoTorchTestFunction **Changes in this diff**: * Introduces SurrogateTestFunction, a ParamBasedTestProblem for surrogates * Removes SurrogateRunner. There is some loss of (unused) functionality here; dict-valued 'noise_std' is no longer supported. I expect we will eventually want that, but it doesn't need to be done now. * Changed `predict_for_tensor` utility function to use a `SurrogateTestFunction` rather than a `SurrogateRunner` * Removed `SurrogateRunner` from the json_store registry. There is no point in keeping it around for backward compatibility since it was just deserializing to a dummy `SyntheticRunner` anyway. Differential Revision: D64855010
Summary: **Context**: This diff cuts over over uses of `SurrogateRunner` to use `ParamBasedTestProblemRunner` with a `test_problem` that is the newly introduced `SurrogateTestFunction`, and the following diff after that will bring us down to only one runner class for benchmarking by merging `ParamBasedTestProblemRunner` into `BenchmarkRunner`. Having only one runner will make it easier to enable asynchronous benchmarks. Currently, SurrogateRunner had its own logic for tracking when trials are completed, which would make it difficult to work in with asynchronicity. **Note on naming**: Some names have become non-intuitive in the process of benchmarking. To accord with some future changes I hope to make, I called a new class SurrogateTestFunction, whereas SurrogateParamBasedTestProblem would be more in line with the old naming. The name changes I hope to make: * ParamBasedTestProblemRunner -> nothing, absorbed into BenchmarkRunner * ParamBasedTestProblem -> TestFunction, to emphasize that all it does is generate data (rather than more generally specify the problem we are solving) and that it is deterministic, and to differentiate it from BenchmarkProblem. BenchmarkTestFunction would also be a candidate. * BoTorchTestProblem -> BoTorchTestFunction **Changes in this diff**: * Introduces SurrogateTestFunction, a ParamBasedTestProblem for surrogates * Removes SurrogateRunner. There is some loss of (unused) functionality here; dict-valued 'noise_std' is no longer supported. I expect we will eventually want that, but it doesn't need to be done now. * Changed `predict_for_tensor` utility function to use a `SurrogateTestFunction` rather than a `SurrogateRunner` * Removed `SurrogateRunner` from the json_store registry. There is no point in keeping it around for backward compatibility since it was just deserializing to a dummy `SyntheticRunner` anyway. Differential Revision: D64855010
esantorella
force-pushed
the
export-D64855010
branch
from
027cfba to
712a0e1
Compare
|
This pull request was exported from Phabricator. Differential Revision: D64855010 |
esantorella
added a commit
to esantorella/Ax
that referenced
this pull request
Oct 28, 2024
Summary: **Context**: This diff cuts over over uses of `SurrogateRunner` to use `ParamBasedTestProblemRunner` with a `test_problem` that is the newly introduced `SurrogateTestFunction`, and the following diff after that will bring us down to only one runner class for benchmarking by merging `ParamBasedTestProblemRunner` into `BenchmarkRunner`. Having only one runner will make it easier to enable asynchronous benchmarks. Currently, SurrogateRunner had its own logic for tracking when trials are completed, which would make it difficult to work in with asynchronicity. **Note on naming**: Some names have become non-intuitive in the process of benchmarking. To accord with some future changes I hope to make, I called a new class SurrogateTestFunction, whereas SurrogateParamBasedTestProblem would be more in line with the old naming. The name changes I hope to make: * ParamBasedTestProblemRunner -> nothing, absorbed into BenchmarkRunner * ParamBasedTestProblem -> TestFunction, to emphasize that all it does is generate data (rather than more generally specify the problem we are solving) and that it is deterministic, and to differentiate it from BenchmarkProblem. BenchmarkTestFunction would also be a candidate. * BoTorchTestProblem -> BoTorchTestFunction **Changes in this diff**: * Introduces SurrogateTestFunction, a ParamBasedTestProblem for surrogates * Removes SurrogateRunner. There is some loss of (unused) functionality here; dict-valued 'noise_std' is no longer supported. I expect we will eventually want that, but it doesn't need to be done now. * Changed `predict_for_tensor` utility function to use a `SurrogateTestFunction` rather than a `SurrogateRunner` * Removed `SurrogateRunner` from the json_store registry. There is no point in keeping it around for backward compatibility since it was just deserializing to a dummy `SyntheticRunner` anyway. Differential Revision: D64855010
esantorella
added a commit
to esantorella/Ax
that referenced
this pull request
Oct 28, 2024
Summary: Pull Request resolved: facebook#2954 **Context**: This diff cuts over over uses of `SurrogateRunner` to use `ParamBasedTestProblemRunner` with a `test_problem` that is the newly introduced `SurrogateTestFunction`, and the following diff after that will bring us down to only one runner class for benchmarking by merging `ParamBasedTestProblemRunner` into `BenchmarkRunner`. Having only one runner will make it easier to enable asynchronous benchmarks. Currently, SurrogateRunner had its own logic for tracking when trials are completed, which would make it difficult to work in with asynchronicity. **Note on naming**: Some names have become non-intuitive in the process of benchmarking. To accord with some future changes I hope to make, I called a new class SurrogateTestFunction, whereas SurrogateParamBasedTestProblem would be more in line with the old naming. The name changes I hope to make: * ParamBasedTestProblemRunner -> nothing, absorbed into BenchmarkRunner * ParamBasedTestProblem -> TestFunction, to emphasize that all it does is generate data (rather than more generally specify the problem we are solving) and that it is deterministic, and to differentiate it from BenchmarkProblem. BenchmarkTestFunction would also be a candidate. * BoTorchTestProblem -> BoTorchTestFunction **Changes in this diff**: * Introduces SurrogateTestFunction, a ParamBasedTestProblem for surrogates * Removes SurrogateRunner. There is some loss of (unused) functionality here; dict-valued 'noise_std' is no longer supported. I expect we will eventually want that, but it doesn't need to be done now. * Changed `predict_for_tensor` utility function to use a `SurrogateTestFunction` rather than a `SurrogateRunner` * Removed `SurrogateRunner` from the json_store registry. There is no point in keeping it around for backward compatibility since it was just deserializing to a dummy `SyntheticRunner` anyway. Differential Revision: D64855010
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: This diff cuts over over uses of
SurrogateRunnerto useParamBasedTestProblemRunnerwith atest_problemthat is the newly introducedSurrogateTestFunction, and the following diff after that will bring us down to only one runner class for benchmarking by mergingParamBasedTestProblemRunnerintoBenchmarkRunner. Having only one runner will make it easier to enable asynchronous benchmarks. Currently, SurrogateRunner had its own logic for tracking when trials are completed, which would make it difficult to work in with asynchronicity.Note on naming: Some names have become non-intuitive in the process of benchmarking. To accord with some future changes I hope to make, I called a new class SurrogateTestFunction, whereas SurrogateParamBasedTestProblem would be more in line with the old naming.
The name changes I hope to make:
Changes in this diff:
predict_for_tensorutility function to use aSurrogateTestFunctionrather than aSurrogateRunnerSurrogateRunnerfrom the json_store registry. There is no point in keeping it around for backward compatibility since it was just deserializing to a dummySyntheticRunneranyway.Differential Revision: D64855010