Negative rank on 2-d tensors (matrices)

[Bihaqo]

Example:

A = np.array(
[[  1.,   3.],
 [  2.,   4.]])
print tt.tensor(A)

This is a 2-dimensional tensor
r(0)=1, n(0)=2
r(1)=-101, n(1)=2
r(2)=1

[prokopevaleksey]

For matrixes use tt.matrix()

[Bihaqo]

tt.matrix(tt.tensor(np.random.rand(4, 4)))

This is a 2-dimensional matrix
r(0)=1, n(0)=2, m(0)=2
r(1)=-19, n(1)=2, m(1)=2
r(2)=1

[oseledets]

Fixed (finally) pull the latest version