Parameter with uniform distribution over integer range for hyperparameter optimization
Parameter with uniform distribution over integer range for hyperparameter optimization
comparer::par_hype -> par_integer
nameName of the parameter, must match the input to `eval_func`.
lowerLower bound of the parameter
upperUpper bound of the parameter
ggtransTransformation for ggplot, see ggplot2::scale_x_continuous()
fromraw()Function to convert from raw scale to transformed scale
R6_par_integer$fromraw(x)xValue of raw scale
toraw()Function to convert from transformed scale to raw scale
R6_par_integer$toraw(x)xValue of transformed scale
generate()Generate values in the raw space based on quantiles.
R6_par_integer$generate(q)qIn [0,1].
getseq()Get a sequence, uniform on the transformed scale
R6_par_integer$getseq(n)nNumber of points. Ignored for discrete.
isvalid()Check if input is valid for parameter
R6_par_integer$isvalid(x)xParameter value
convert_to_mopar()Convert this to a parameter for the mixopt R package.
R6_par_integer$convert_to_mopar(raw_scale = FALSE)raw_scaleShould it be on the raw scale?
new()Create a hyperparameter with uniform distribution
R6_par_integer$new(name, lower, upper)nameName of the parameter, must match the input to `eval_func`.
lowerLower bound of the parameter
upperUpper bound of the parameter
...not used,
clone()The objects of this class are cloneable with this method.
R6_par_integer$clone(deep = FALSE)deepWhether to make a deep clone.
p1 <- R6_par_integer$new('x1', 0, 2)
class(p1)
print(p1)
Run the code above in your browser using DataLab