Unravel_index operator missing the Boundary Check for the Input index

[Tommliu]

Description

The current unravel_index method under tensor operator didn't check whether the input index is
valid given the input shape.

To Reproduce

[lanking520]

Hi @Tommliu That seemed to be a numpy problem since you import numpy as np.

[Tommliu]

Should the unravel_index function in Mxnet follow the numpy convention? I think a boundary check is necessary when using the unravel function.

[szha]

yes it should.