Task Description

For SHAPGradient and SHAPKernel:

• describe the input and output type and the required shape of the network/predictor
• describe if predictor returns 2D tensor -> classification scenario else if
  returns 1D tensor -> regression scenario
• update the docstrings for issue#I4UXXV
• enhance the details of explainer's outputs descriptions
• enhance the descriptions of the 'features' argument, tell user it is okay to use training (or part of) feature tensor/array for it

Task Goal

Sub Task