Creates a categorical grid search definition from a vector of values.

value-vec - Vector of categorical values to search over

Returns a grid search definition map that will enumerate all values in the vector during grid search. This is a convenience wrapper around linear for categorical parameters.

Example: (categorical [:adam :sgd :rmsprop]) creates a definition that will try all three optimizers.

See also: linear, sobol-gridsearch

Creates a linear grid search definition for hyperparameter tuning.

• start - Starting value of the range
• end - Ending value of the range
• n-steps - Number of evenly-spaced steps (default: 100)
• res-dtype-or-space - Either a datatype keyword (:float64, :int32, etc.) or a vector of categorical values (default: :float64)

Returns a grid search definition map that can be used with sobol-gridsearch.

When res-dtype-or-space is a vector, values are mapped to categorical options by index. For example, [:small :medium :large] with 3 steps generates those three categorical values.

See also: categorical, sobol-gridsearch

Given an map of key->values where some of the values are gridsearch definitions produce a sequence of fully defined maps.

user> (require '[tech.v3.ml.gridsearch :as ml-gs])
nil
user> (def opt-map  {:a (ml-gs/categorical [:a :b :c])
                     :b (ml-gs/linear 0.01 1 10)
                     :c :not-searched})
user> opt-map
{:a
 {:tech.v3.ml.gridsearch/type :linear,
  :start 0.0,
  :end 2.0,
  :n-steps 3,
  :result-space [:a :b :c]}
  ...

user> (ml-gs/sobol-gridsearch opt-map)
({:a :b, :b 0.56, :c :not-searched}
 {:a :c, :b 0.22999999999999998, :c :not-searched}
 {:a :b, :b 0.78, :c :not-searched}
...