Jeremy Coyle

Reduce model storage size

7

A lot of models store copies of the data and other large objects. See discussion here: http://www.win-vector.com/blog/2014/05/trimming-the-fat-from-glm-models-in-r/ related to `glm`. Our model fit objects should store only what's necessary for...

WIP PR that adds `Lrnr_base$reduce_fit` to drop `fit_object` components not necessary for prediction, reducing memory demands. In response to https://github.com/tlverse/sl3/issues/12 * Implements `Lrnr_base$reduce_fit`, which drops out `fit_object` components * Adds...

Currently, each h2o learner re-uploads data to h2o. We should refactor so it's only done once per task

helper functions e.g. confusion matrix sens/spec etc etc

Instead, each takes a separate seed argument. Currently, users can specify these manually as parameters: ``` lrnr_ranger

Something similar to the `magrittr` `%>%` for chaining learners

Saving/Loading Lrnr_dbarts does not work

1

From the help on dbarts::bart: > Saving > saveing and loading fitted BART objects for use with predict requires that R's serialization mechanism be able to access the underlying trees,...

Currently nonexistent. One implementation exists here: https://github.com/jeremyrcoyle/sl3/blob/tmle-demo-fixes/R/sl3_Task.R#L275, however this is done on-the-fly every time `task$X` is called. This results in repeated computation for the indicators when fitting lots of models...

We should write a minimal set of tests any new learner must pass before being integrated

Some kind of 'cross product' operator

1

Given two sets of learners this would create a stack of pipelines where each learner in the stack is a pipeline with a unique combination of a stage 1 and...