Skip to content

Commit 00b4ee4

Browse files
Balandatfacebook-github-bot
authored andcommitted
Update changelog for 0.5.1 (#930)
Summary: Pull Request resolved: #930 Reviewed By: lena-kashtelyan Differential Revision: D30734039 Pulled By: Balandat fbshipit-source-id: 8619e13129b2e89af3979c2a7d71af04d5b58bea
1 parent 304f5e2 commit 00b4ee4

File tree

1 file changed

+33
-0
lines changed

1 file changed

+33
-0
lines changed

CHANGELOG.md

+33
Original file line numberDiff line numberDiff line change
@@ -2,6 +2,39 @@
22

33
The release log for BoTorch.
44

5+
## [0.5.1] - Sep 2, 2021
6+
7+
#### Compatibility
8+
* Require GPyTorch >=1.5.1 (#928).
9+
10+
#### New Features
11+
* Add `HigherOrderGP` composite Bayesian Optimization tutorial notebook (#864).
12+
* Add Multi-Task Bayesian Optimziation tutorial (#867).
13+
* New multi-objective test problems from (#876).
14+
* Add `PenalizedMCObjective` and `L1PenaltyObjective` (#913).
15+
* Add a `ProximalAcquisitionFunction` for regularizing new candidates towards previously generated ones (#919, #924).
16+
* Add a `Power` outcome transform (#925).
17+
18+
#### Bug Fixes
19+
* Batch mode fix for `HigherOrderGP` initialization (#856).
20+
* Improve `CategoricalKernel` precision (#857).
21+
* Fix an issue with `qMultiFidelityKnowledgeGradient.evaluate` (#858).
22+
* Fix an issue with transforms with `HigherOrderGP`. (#889)
23+
* Fix initial candidate generation when parameter constraints are on different device (#897).
24+
* Fix bad in-place op in `_generate_unfixed_lin_constraints` (#901).
25+
* Fix an input transform bug in `fantasize` call (#902).
26+
* Fix outcome transform bug in `batched_to_model_list` (#917).
27+
28+
#### Other Changes
29+
* Make variance optional for `TransformedPosterior.mean` (#855).
30+
* Support transforms in `DeterministicModel` (#869).
31+
* Support `batch_shape` in `RandomFourierFeatures` (#877).
32+
* Add a `maximize` flag to `PosteriorMean` (#881).
33+
* Ignore categorical dimensions when validating training inputs in `MixedSingleTaskGP` (#882).
34+
* Refactor `HigherOrderGPPosterior` for memory efficiency (#883).
35+
* Support negative weights for minimization objectives in `get_chebyshev_scalarization` (#884).
36+
* Move `train_inputs` transforms to `model.train/eval` calls (#894).
37+
538
## [0.5.0] - Jun 29, 2021
639

740
#### Compatibility

0 commit comments

Comments
 (0)