rstudio-conf-2020 / applied-ml Goto Github PK

Some participants on windows machines seemed to be confused by the {doMC} based parallel computing in code.R.
Might be nice to have a crossplatform backend in code.R to avoid that source of confusion, though that would depend on tidymodels/tune#157.
Don't know enough about the internals of tune or parallel on windows, so I can't solve this myself 😞.

These slides have problematic text running off the bottom of the slide...

• P5: While We Wait, Can I Interest You in Parallelism?
• P5: Running in Parallel with {tune}
• P5: MARS in via {parsnip} and {tune}
• P5: Example of Logging
• P6: Class Prediction Metrics
  • Column 2
• P6: Changing the Probability Threshold
  • Column 1
• P6: Feature hashing
  • Column 2
• P6: Growing phase
  • Image labels are below slide line