Rerun Latent GP with method="svd"
for MvNormals
#784
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Outdated suggestions cannot be applied.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.
Why & How
MvNormal was switched from SVD to Cholesky decomposition by default. This is brittle for many GPs, therefore PyMC > 5.22.0 will defaults Latent GP conditionals to
method="svd"
(via pymc-devs/pymc#7755).The example notebook includes manually created
MvNormal
s that also needmethod="svd"
to work.I also switched from numpyro back to the default sampler, because I couldn't install numpyro on my Windows.
See pymc-devs/pymc#7754.
📚 Documentation preview 📚: https://pymc-examples--784.org.readthedocs.build/en/784/