Fix failing LinRegHMM test #378
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
Changes: - Replace `jnp.ones` input with `jr.normal`. - Reduce size of hidden state to 3. - Remove unused `datetime` import and commented lines. Fundamentally the problem is that the solve step can be unstable. This does not resolve that but instead chooses a set up which is less vulnerable to the instability.
|
Sorry to miss this review request, @gileshd. Thanks for fixing the test. |
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Outdated suggestions cannot be applied.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.
Summary
This PR changes the parameters of the
LinearRegressionHMMtest to prevent it failing.Details
At present the
LinearRegressionHMMtest_sample_and_fittest is failing. Themonotonically_increasingassertion is failing as the log likelihoods quickly becomejnp.nan.I think this is fundamentally a question of the stability of the linear solve. I have made some changes to the inputs which are hopefully more robust to solving.
Changes:
jnp.onesinput withjr.normal.There are likely more principled ways of choosing a more robust test case but in the interest of getting the test to pass I think it is helpful to make a simple change quickly.
Additional changes:
datetimeimport and commented lines.