Shuffle cross-validation mode for each learner type
Usage
assign_learner_cv(
learner = c("lgb", "mlp", "elnet"),
cv_mode = c("spatiotemporal", "spatial", "temporal"),
cv_rep = 100L,
num_device = ifelse(torch::cuda_device_count() > 1, 2, 1),
balance = FALSE
)Arguments
- learner
character(1). The base learner to be used. Default is "mlp". Available options are "mlp", "lgb", "elnet".
- cv_mode
character(1). The cross-validation mode to be used. Default is "spatiotemporal". Available options are "spatiotemporal", "spatial", "temporal".
- cv_rep
integer(1). The number of repetitions for each
cv_mode.- num_device
integer(1). The number of CUDA devices to be used. Each device will be assigned to each eligible learner (i.e., lgb, mlp).
- balance
logical(1). If TRUE, the number of CUDA devices will be equally distributed based on the number of eligible devices.