Add describe info

[owulveryck]

Is your feature request related to a problem? Please describe.
N/A

Describe the solution you'd like
it could help to have functions to describe the expected shapes of input and output.
For example, in the case of image classification, the input's shape is related to the size of the picture. This could allow to easily use a pre-processing of the picture, without hardcoding it.

Getting the same info for the output can help to decide (in case of a classification algorithm).

Describe alternatives you've considered
Finding them manually.

Additional context
N/A