Sam Daulton comments

Results 64 comments of


                                            Sam Daulton

Fix output tensor dimension in degenerate `optimize_acqf` call

Thanks for the PR! It looks like a test is failing. Would you mind fixing that please?

[FEATURE REQUEST]: StratifiedStandardize for multi-output models

Hi @Hrovatin, > This would be needed for MultiTaskGP which requires stratified standardisation and can work with multiple outputs. StratifiedStandardize works with `MultiTaskGP` if you include the task values in...

[FEATURE REQUEST]: StratifiedStandardize for multi-output models

> Did you ever observe that doing startified output scaling could in fact be problematic when tasks do not cover the same regions of the output space? I haven't observed...

Probabilistic Reparameterization Part 2

> I have included the tests for the *PRInputTransform with the other PR tests. Should they be moved to test/models/transform/input.py? It seems fine to leave them with the other PR...

NP Regression Model w/ LIG Acquisition

Hi @eibarolle! Thanks for the PR! I'll review it shortly.

NP Regression Model w/ LIG Acquisition

Looks like there is a merge conflict in docs/acquisition.md and docs/models.md too

NP Regression Model w/ LIG Acquisition

Thanks! Running tests now

NP Regression Model w/ LIG Acquisition

Looks like some more lines require test coverage, and it looks like there is some non-unix new-line character in the code files. ``` You must use ONLY Unix linebreaks \n...

NP Regression Model w/ LIG Acquisition

@eibarolle looks like there are some lines still missing test coverage. Could you add tests for those gaps please? Thanks!

[Bug] `KroneckerMultiTaskGP` incompatible with `batch_cross_validation` or batched models

Hi @slishak-PX, Thanks for flagging this. The docstring for `batch_cross_validation` says that MultiTaskGPs are not supported, but we should raise an `UnsupportedError` too. I'll put up a PR to add...