Closed devmotion closed 4 years ago
Merging #8 into master will decrease coverage by
2.34%
. The diff coverage is100.00%
.
@@ Coverage Diff @@
## master #8 +/- ##
==========================================
- Coverage 74.64% 72.30% -2.35%
==========================================
Files 5 5
Lines 71 65 -6
==========================================
- Hits 53 47 -6
Misses 18 18
Impacted Files | Coverage Δ | |
---|---|---|
src/EllipticalSliceSampling.jl | 100.00% <ø> (ø) |
|
src/interface.jl | 0.00% <ø> (-42.86%) |
:arrow_down: |
src/abstractmcmc.jl | 100.00% <100.00%> (ø) |
|
src/model.jl | 63.15% <100.00%> (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 a2ec96c...7cf6032. Read the comment docs.
I guess it would be good to document and test parallel sampling, and reexport MCMCThreads
and MCMCDistributed
from AbstractMCMC.
OK, I have a working branch locally with https://github.com/TuringLang/AbstractMCMC.jl/pull/45 that fixes these issues. Let's merge this for now and wait until we reach some consensus on the AbstractMCMC PR.
This PR updates EllipticalSliceSampling to AbstractMCMC 2.0. With the switch to the
sample
interface we will get parallel sampling for free.