Skip to content

Fix hessian bug #1232

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
merged 3 commits into from
May 7, 2020
Merged

Fix hessian bug #1232

merged 3 commits into from
May 7, 2020

Conversation

cpfiffer
Copy link
Member

Fixes #1231, which is necessary for #1230. Adds a test to make sure the Hessian is being correctly calculated with ReverseDiff and ForwardDiff.

@yebai
Copy link
Member

yebai commented Apr 26, 2020

@cpfiffer It seems some tests are broken with this PR

@cpfiffer
Copy link
Member Author

Yeah, it requires TuringLang/DynamicPPL.jl#83 to be merged & released before it should work, though I'm not sure what stuff is in the air over there. @mohamed82008 could TuringLang/DynamicPPL.jl#83 be merged and released without too much issue at DynamicPPL?

@yebai yebai force-pushed the master branch 2 times, most recently from 2661769 to 701f5d2 Compare May 4, 2020 21:46
@codecov
Copy link

codecov bot commented May 7, 2020

Codecov Report

Merging #1232 into master will not change coverage.
The diff coverage is n/a.

Impacted file tree graph

@@           Coverage Diff           @@
##           master    #1232   +/-   ##
=======================================
  Coverage   67.11%   67.11%           
=======================================
  Files          25       25           
  Lines        1341     1341           
=======================================
  Hits          900      900           
  Misses        441      441           
Impacted Files Coverage Δ
src/inference/Inference.jl 90.19% <ø> (ø)

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update 9e75144...0568fda. Read the comment docs.

@cpfiffer
Copy link
Member Author

cpfiffer commented May 7, 2020

The test failures here seem unrelated.

@devmotion
Copy link
Member

I think they're not actually. Using .~ inside a model will lead to parsing errors on Julia 1.0. Alternatively, you could use @. and escape non-broadcasted parts with $, or use some function like syntax (~.(LHS, RHS) IIRC).

@cpfiffer
Copy link
Member Author

cpfiffer commented May 7, 2020

Oh, I see what you mean. I hadn't noticed that this syntax didn't work for Julia 1.0. Let me try this and see what happens.

@cpfiffer
Copy link
Member Author

cpfiffer commented May 7, 2020

Alright, looks like tests are passing.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

Successfully merging this pull request may close these issues.

Hessian failure for logjoint
3 participants