reactivate AD tests: FiniteGP #314
Conversation
![github-actions[bot]](https://avatars.githubusercontent.com/in/15368?s=60&v=4){kind=small}
Co-authored-by: github-actions[bot] <41898282+github-actions[bot]@users.noreply.github.com>
…ractGPs.jl into st/ad_tests2
Co-authored-by: github-actions[bot] <41898282+github-actions[bot]@users.noreply.github.com>
Co-authored-by: github-actions[bot] <41898282+github-actions[bot]@users.noreply.github.com>
…ractGPs.jl into st/ad_tests2
test/sparse_approximations.jl
Outdated
| f = GP(SqExponentialKernel()) | ||
|
|
||
| for Σy in (Σy_Diagonal, Σy_ScalMat) | ||
| for Σy in (Σy_Diagonal, Σy_ScalMat) #, Σy_dense) |
There was a problem hiding this comment.
it's not implemented for dense 🙈
There was a problem hiding this comment.
Probably for the best for now. If the observation covariance is dense, I would imagine that performance will generally be worse than exact inference.
Codecov Report
@@ Coverage Diff @@
## master #314 +/- ##
=======================================
Coverage 97.64% 97.64%
=======================================
Files 10 10
Lines 382 382
=======================================
Hits 373 373
Misses 9 9 Continue to review full report at Codecov.
|
test/sparse_approximations.jl
Outdated
| f = GP(SqExponentialKernel()) | ||
|
|
||
| for Σy in (Σy_Diagonal, Σy_ScalMat) | ||
| for Σy in (Σy_Diagonal, Σy_ScalMat) #, Σy_dense) |
There was a problem hiding this comment.
Probably for the best for now. If the observation covariance is dense, I would imagine that performance will generally be worse than exact inference.
Follow-up to #313