1

I am attempting to create a Random Forest Regression model in R using parsnip tidymodels. I am following this tutorial but using regression instead of classification. I run into an issue when I run this line of code.

doParallel::registerDoParallel()

set.seed(345)
tune_res <- tune_grid(
  tune_wf,
  resamples = trees_folds,
  grid = 20
)

The error is as follows:

Error in checkForRemoteErrors(lapply(cl, recvResult)) : 
  3 nodes produced errors; first error: object '.doSnowGlobals' not found

I have looked at the following SO threada to attempt to resolve this question before asking it myself

  1. error: object '.doSnowGlobals' not found?
  2. parallel, foreach, exported object not found
  3. Error in check for remote errors (val): 5 nodes produced an error: object not found

None of which resolved my issue and none of which helped me fully understand what the issue may be.

For more context here is all my script:

# Splitting out the Data Between Training & Testing -------
set.seed(123)

# Creating Inital Split 
tree_split <- initial_split(stc_boise, strata = DELIVERY_BUS_DAYS, prop = .80)

# Creating Training Set 
tree_train <- training(tree_split)

# Creating Test Set 
tree_test <- testing(tree_split)

# Create Recipe 
tree_rec <- recipe(DELIVERY_BUS_DAYS     ~ DOW + MVNDR_NBR + BRANCH_ID +  STR_NBR + MKT_NBR + RGN_NBR + DISNACE, data = tree_train) 
tree_prep <- prep(tree_rec)
juiced <- juice(tree_prep)

# Specifying the Specs of the Engine 
tune_spec <- rand_forest(
  mtry = tune(),
  trees = 1000,
  min_n = tune()
) %>%
  set_mode("regression") %>%
  set_engine("randomForest")

# Creating the Model Workflow
tune_wf <- workflow() %>%
  add_recipe(tree_rec) %>%
  add_model(tune_spec)

# Train Hyperparameters 
set.seed(234)
trees_folds <- vfold_cv(tree_train)


# Tune the Model 
doParallel::registerDoParallel()

set.seed(345)
tune_res <- tune_grid(
  tune_wf,
  resamples = trees_folds,
  grid = 20
)

tune_res

I have installed all required packages for parallel computing such as: doParallel, doSNOW, am I missing any?

If someone can provide a solution, may you share the code that I can fix it with please? And also, possibly provide an explination for the error so I fully understand why I am facing this error>

desertnaut
  • 57,590
  • 26
  • 140
  • 166
Jaskeil
  • 1,044
  • 12
  • 33
  • 1
    I recommend you [check out this guidance on how to set up parallel processing for tune](https://tune.tidymodels.org/articles/extras/optimizations.html#parallel-processing-1). – Julia Silge Jan 10 '22 at 19:20

0 Answers0