Conversation
There was a problem hiding this comment.
Pull request overview
This PR updates PHYSBO’s permutation-importance implementation to support models (notably BLM) whose get_post_fmean API takes only query points, and extends the Gaussian process tutorial notebooks to demonstrate building/using blm.Model.
Changes:
- Add
query_onlyoption tomisc.get_permutation_importance()to support BLM-styleget_post_fmean(X_query)calls. - Update
blm.Model.get_permutation_importance()to usequery_only=True. - Extend JP/EN GP tutorial notebooks with a new section demonstrating
gp.export_blm()and basic BLM usage (including permutation importance).
Reviewed changes
Copilot reviewed 4 out of 4 changed files in this pull request and generated 3 comments.
| File | Description |
|---|---|
src/physbo/misc/permutation_importance.py |
Adds query_only branch for BLM-style prediction API (but still has a critical MPI feature-splitting bug). |
src/physbo/blm/core/_model.py |
Routes BLM permutation importance through the updated helper with query_only=True. |
docs/sphinx/manual/ja/source/notebook/tutorial_Gaussian_process.ipynb |
Adds BLM tutorial section (includes very large stored outputs). |
docs/sphinx/manual/en/source/notebook/tutorial_Gaussian_process.ipynb |
Adds BLM tutorial section (includes very large stored outputs). |
💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.
|
After seeing copilot's review, I find that |
There was a problem hiding this comment.
Pull request overview
Copilot reviewed 7 out of 7 changed files in this pull request and generated 2 comments.
💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.
Fix
misc.permutation_importanceandblm.Model.permutation_importanceAdded a tutorial how to use
blm.ModelUpdate of
blm.Model.predict_sampling(X, N=n)N=1returnsN=X.shape[0]result(n, X.shape[0])array instead of(X.shape[0], n)to achieve the consistency withgp.Model.predict_sampling(), which returns(n, X.shape[0])