mkl_fft.irfftn returns incorrect shape

[vtavana]

For the following case, the output array shape from mkl_fft.irfftn and NumPy is different.

import numpy, mkl_fft
from numpy.testing import assert_allclose

a = numpy.array([[[5, 7, 6, 5],
		  [4, 6, 4, 8],
		  [9, 3, 7, 5]],
		 [[5, 9, 0, 0],
		  [0, 8, 7, 8],
		  [9, 7, 4, 7]]], dtype=numpy.float32)

print(a.shape)
# (2, 3, 4)

r1 = mkl_fft.irfftn(a, s=(1, 4, 10), axes=None)
r2 = numpy.fft.irfftn(a, s=(1, 4, 10), axes=None)

print(r1.shape, r2.shape)
# (2, 4, 10) (1, 4, 10)