Using DirectorKernel with CrossValidation crashes in python

[dhelekal]

OS family: Linux
Python 3.5 / Shogun 6.1.3 (Both conda)

Attempting to run CrossValidation on an mkl-svm containing a DirectorKernel throws:
RecursionError: maximum recursion depth exceeded while calling a Python object

Minimal working example:

from shogun import *
from numpy import *

class DirectorTestKernel(DirectorKernel):
    def __init__(self):
        DirectorKernel.__init__(self, True)
    def kernel_function(self, idx_a, idx_b):
        seq1 = self.get_lhs().get_feature_vector(idx_a)
        seq2 = self.get_rhs().get_feature_vector(idx_b)
        return dot(seq1, seq2)

feats_train=RealFeatures(identity(10))  
labels=BinaryLabels(array([1,1,1,1,1,-1,-1,-1,-1,-1]))   

ker1 = DirectorTestKernel()
ker2 = DirectorTestKernel()

#ker1 = GaussianKernel(1.0)  Test purposes, works with gaussiankernels
#ker2 = GaussianKernel(2.0) Test purposes, works with gaussiankernels

cker = CombinedKernel()

cker.append_kernel(ker1)
cker.append_kernel(ker2)

cft = CombinedFeatures()

cft.append_feature_obj(feats_train)
cft.append_feature_obj(feats_train)

cker.init(cft,cft)

libsvm = LibSVM()
svm = MKLClassification(libsvm)
svm.set_interleaved_optimization_enabled(False)
svm.set_kernel(cker)
svm.set_C_mkl(1000)
svm.set_mkl_norm(1)

Finally, running x-validation causes a crash:

splitting_strategy = StratifiedCrossValidationSplitting(labels, 2)
evaluation_criterium = AccuracyMeasure()
cross = CrossValidation(svm, cft, labels, splitting_strategy, evaluation_criterium)
cross.set_autolock(False)
cross.set_num_runs(2)

result = CrossValidationResult()
result = CrossValidationResult.obtain_from_generic(cross.evaluate())

[karlnapf]

Is this still a problem with your other fix in?

[stale]

This issue has been automatically marked as stale because it has not had recent activity. It will be closed if no further activity occurs. Thank you for your contributions.

[gf712]

user hasn't replied in two years, closing for now