align_tensor.init() can now accept any type of float or int (python or numpy).
Authored by: bugman 2009-07-13
Parent: [r9207]
Child: [r9209]
