gp sampler: GPy -> scikit-learn #69
Merged
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
marcofavoritobi
force-pushed
the
from_GPy_to_sklearn
branch
from
c97ac3b to
531a97a
Compare
Codecov Report
❗ Your organization needs to install the Codecov GitHub app to enable full functionality. Additional details and impacted files@@ Coverage Diff @@
## main #69 +/- ##
==========================================
+ Coverage 91.36% 91.38% +0.01%
==========================================
Files 42 42
Lines 1737 1729 -8
==========================================
- Hits 1587 1580 -7
+ Misses 150 149 -1
|
muxator
reviewed
Sep 19, 2023
marcofavoritobi
force-pushed
the
from_GPy_to_sklearn
branch
from
531a97a to
3bdeb7c
Compare
| @@ -37,6 +37,7 @@ | |||
| batch_id # unused variable (black_it/schedulers/base.py:77) | |||
| batch_id # unused variable (black_it/schedulers/round_robin.py:47) | |||
| MABEpsilonGreedy # unused class (black_it/schedulers/rl/agents/epsilon_greedy.py:26) | |||
| seed # unused variable (black_it/schedulers/rl/envs/base.py:59) | |||
There was a problem hiding this comment.
for some reason this vulture error was activated after the first commit of this PR.
marcofavoritobi
force-pushed
the
from_GPy_to_sklearn
branch
2 times, most recently
from
fe677e7 to
21a93ad
Compare
BREAKING CHANGE: This commit updates the implementation of GaussianProcessSampler. Instead of relying on GPy, now we use the implementation of Gaussian Processes provided by the scikit-learn package. The change introduces a different behaviour of the GaussianProcessSampler, even if the same random seed is provided. This is witnessed by the updated expected sampled parameters and losses in tests/test_calibrator.py. Another notable difference is in the keyword arguments for the GaussianProcessSampler class constructor. In particular: - the keyword argument 'max_iters' has been removed - the keyword argument 'jitter' has been added.
marcofavoritobi
force-pushed
the
from_GPy_to_sklearn
branch
from
21a93ad to
01447d4
Compare
Comment on lines
-79
to
+80
| self.max_iters = max_iters | ||
| self.optimize_restarts = optimize_restarts | ||
| self.acquisition = acquisition | ||
| self._gpmodel: Optional[GPRegression] = None | ||
| self.jitter = jitter |
There was a problem hiding this comment.
optimize_restarts, acquisition and jitter should be protected attributes. We can fix this in another PR.
|
Thank you @marcofavoritobi @muxator! Merging now |
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.
Proposed changes
I change the backend of the GPSampler using the scikit-learn implementation instead of the GPy one.
Fixes
This also solves #36.
Types of changes
What types of changes does your code introduce?
Put an
xin the boxes that apply