MethodError: Evolution of Out Of Sample Performance Range

Specific Domains
1

Hello Folks:

I am using MLJ@.load_boston data where

feat, targ = MLJ.@load_boston
	feat = DataFrames.DataFrame(feat)
	targ = categorical(targ)

I have split the categorical.array using:

	rng = 52
	TRAIN, TEST = partition(eachindex(targ), 0.8, shuffle=true, rng=rng)

I am loading the NNC model using:

	mflux =  @load NeuralNetworkClassifier
	clf = NeuralNetworkClassifier()

Testing the machine using:

FLUX = machine(clf, feat, targ)

Fitting the machine with:

MLJ.fit!(FLUX, rows=TRAIN, verbosity=3)

When I enter the following:

r_clf = range(clf, :epochs, lower=1, upper=200, scale=:log10)

… am getting the following error:

MethodError: objects of type Int64 are not callable

When I remove the categorical(targ) casting,
I get another suite of issues.

Any ideas out there?

2

Okay – First issue addressed. Prefixing:

MLJ.range()