feature/autosigma #6
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
…y sigma_sq (assuming it is trained). Overridable with the apply_sigma_sq kwarg.
bwpriest
added a commit
that referenced
this pull request
Dec 9, 2021
* Changed SigmaSq to its own class type. APIs no longer allow directly setting sigma_sq values, and instead only indicate whether to train them. * Update README.md missed dropping a reference to old sigma_sq api pattern * Hotfix/learn-sigma (#5) * Fixed a bug causing sigma_sqs to always go unlearned. * Changed default behavior throughout to automatically scale variance by sigma_sq (assuming it is trained). Overridable with the apply_sigma_sq kwarg. (#6) * Added option to all high-level apis to return crosswise_dists and pairwise_dists. (#7) * Added make_regress_tensors and make_train_tensors helper functions and dispersed them throughout the code. Simplified optimization API. Cleaned up some documentation. (#8) * Added CONTRIBUTING.md document. (#9) * Refactored BenchmarkGP to share new gp API. Added _sq_rel_err test function. Explored bug in rbf computation of sigma_sq, and turned off related test cases. See issue #1. (#11) * Feature/notebooks (#12) * First pass on a univariate regression tutorial, describing the basics of MuyGPyS. * Cleaned up some comments and documentation. * Linked tutorial notebook into docs and inserted links. * bugfix: changed return distances semantics and documentation to work correctly. * some miscellaneous cleanup. * Added regression API tutorial. * Edited readme to remove code and added links to notebooks. * Added matplotlib to docs/requirements.txt. (#13) * hopefully fixed rtd.yml (#14) * Docs (#15) * formatting fix to get readthedocs to build the project for notebooks * Hotfix/sigma sq approx (#16) * Addresses sigma_sq inference bug in issue #10 * Fixed test chassis to avoid overly smooth curves for to guarantee well-behaved estimators. * Fixed a bug in the sigma_sq testing chassis. (#18) * incremented minor version number.
bwpriest
added a commit
that referenced
this pull request
Jan 19, 2022
* Changed SigmaSq to its own class type. APIs no longer allow directly setting sigma_sq values, and instead only indicate whether to train them. * Update README.md missed dropping a reference to old sigma_sq api pattern * Hotfix/learn-sigma (#5) * Fixed a bug causing sigma_sqs to always go unlearned. * Changed default behavior throughout to automatically scale variance by sigma_sq (assuming it is trained). Overridable with the apply_sigma_sq kwarg. (#6) * Added option to all high-level apis to return crosswise_dists and pairwise_dists. (#7) * Added make_regress_tensors and make_train_tensors helper functions and dispersed them throughout the code. Simplified optimization API. Cleaned up some documentation. (#8) * Added CONTRIBUTING.md document. (#9) * Refactored BenchmarkGP to share new gp API. Added _sq_rel_err test function. Explored bug in rbf computation of sigma_sq, and turned off related test cases. See issue #1. (#11) * Feature/notebooks (#12) * First pass on a univariate regression tutorial, describing the basics of MuyGPyS. * Cleaned up some comments and documentation. * Linked tutorial notebook into docs and inserted links. * bugfix: changed return distances semantics and documentation to work correctly. * some miscellaneous cleanup. * Added regression API tutorial. * Edited readme to remove code and added links to notebooks. * Added matplotlib to docs/requirements.txt. (#13) * hopefully fixed rtd.yml (#14) * Docs (#15) * formatting fix to get readthedocs to build the project for notebooks * Hotfix/sigma sq approx (#16) * Addresses sigma_sq inference bug in issue #10 * Fixed test chassis to avoid overly smooth curves for to guarantee well-behaved estimators. * Fixed a bug in the sigma_sq testing chassis. (#18) * incremented minor version number. * implemented a faster crosswise_distances function. (#19) * Modified training procedure to use pure functions. scipy_optimize_from_* functions now return an optimized model, leaving the model parameter object unaffected. (#21) This is a breaking update. loo_crossval signature changed significantly, and scipy_optimize_from_* functions now return an optimized MuyGPS object rather than modifying one in place. * Fixed an import bug introduced in last PR (#22) * feature/hyp_fix (#23) * Moved opt function preparation into KernelFn and MuyGPS classes to avoid yucky dict issues. * hyperparameter bounds no longer take the value "fixed". They instead default to (0.0, 0.0). The fixed status of hyperparameters is now accessed via Hyperparameter.fixed(). * SigmaSq no longer defaults to unlearned. Now defaults to [1.0], and added the SigmaSq.trained() boolean function to query if it has been set. * Added correct initialization checking to test chassis. * Small clean up. * Updated relevant documentation. * Relax scipy requirement to build on Python > 3.7 (#25) * Relax scipy requirement to build on Python > 3.7 Scipy 1.4.1 appears to be [incompatible with Python > 3.7][1]. Relaxing the exact version requirement makes it possible to install MuyGPyS on newer versions of Python. [1]: https://github.com/scipy/scipy/blob/adc4f4f7bab120ccfab9383aba272954a0a12fb0/setup.py#L44-L46 * incremented hnswlib version to 0.6.0. Co-authored-by: Ben Priest <bwpriester@gmail.com> * Adding more functional indirection to streamline optimization (#26) * Broke Matern static functions out into different versions for each nu setting to avoid boolean checks at each evaluation during optimization. * streamlined the loo_crossval objective function by wrapping it in a caller that only accepts the arguments that change. * Updated README to point to stable version of documentation. * Removed all outputs and timings from univariate regression tutorial. Co-authored-by: Imène Goumiri <imene.goumiri@gmail.com>
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.
No description provided.