Seems like the sklearn to flow serialization does not like it when a np.int64 is passed through:

  File "/home/vanrijn/projects/openml-python/openml/flows/sklearn_converter.py", line 81, in sklearn_to_flow
    raise TypeError(o, type(o))
TypeError: (2, <class 'numpy.int64'>)

Apparently, that case is not caught by
elif isinstance(o, (bool, int, float, six.string_types)) or o is None: