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).
Somewhere between closing #25 and now, I've introduced another bottleneck. On the Macbook, tidylda is still faster than textmineR, if just. But on the Ubuntu it is now slower than the Macbook. And textmineR is now also faster than tidylda on the Ubuntu. Not sure what happened.
(a) I'll look into what I've modified between #25 and now
(b) I'll look into moving forward solutions vs. trying to roll back, unless necessary.
Off the top of my head, two things it could be: cache misses, particularly for Cv and the sample function in the middle of everything could be a bottleneck.
TommyJones
commented
2 years ago
Somewhere between closing #25 and now, I've introduced another bottleneck. On the Macbook, tidylda is still faster than textmineR, if just. But on the Ubuntu it is now slower than the Macbook. And textmineR is now also faster than tidylda on the Ubuntu. Not sure what happened.
(a) I'll look into what I've modified between #25 and now (b) I'll look into moving forward solutions vs. trying to roll back, unless necessary.
Off the top of my head, two things it could be: cache misses, particularly for Cv and the sample function in the middle of everything could be a bottleneck.