search

TommyJones / tidylda

Implements an algorithim for Latent Dirichlet Allocation using style conventions from the [tidyverse](https://style.tidyverse.org/) and [tidymodels](https://tidymodels.github.io/model-implementation-principles/index.html).
Other
41 stars 3 forks source link

Revert Gibbs sampler back to sequential #53

Closed TommyJones closed 2 years ago

TommyJones commented 2 years ago

Doing this b/c (a) doing parallel sampling in a way that respects the user's seed and conforms to CRAN policies was more trouble than it was worth because (b) parallel sampling (which was "approximate Gibbs" a-la Newman et al., 2009) significantly decreased goodness of fit by several metrics I usually rely on.

codecov[bot] commented 2 years ago

Codecov Report

Merging #53 (270cd08) into main (047ba67) will increase coverage by 0.06%. The diff coverage is 100.00%.

:exclamation: Current head 270cd08 differs from pull request most recent head 265d950. Consider uploading reports for the commit 265d950 to get more accurate results Impacted file tree graph

@@            Coverage Diff             @@
##             main      #53      +/-   ##
==========================================
+ Coverage   95.07%   95.13%   +0.06%     
==========================================
  Files          11       11              
  Lines        1117     1111       -6     
==========================================
- Hits         1062     1057       -5     
+ Misses         55       54       -1
Impacted Files Coverage Δ
R/utils.R 90.54% <100.00%> (+0.39%) :arrow_up:
src/lda_gibbs2.cpp 98.69% <100.00%> (-0.04%) :arrow_down:

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 047ba67...265d950. Read the comment docs.