This PR adds some code to handle the 5-input case for the CNN. The assumption is that only one
of those inputs will be perturbed by the XAI method.
The uses gives the 4 constant inputs to the wrapper function, and tells it where the fifth input belongs.
The fifth input, i.e. the one that will be perturbed, is given to dianna.explain_image. The wrapper
function is given as preprocess_function to dianna.explain_image and handles reconstructing the
proper input shape.
This PR adds some code to handle the 5-input case for the CNN. The assumption is that only one of those inputs will be perturbed by the XAI method. The uses gives the 4 constant inputs to the wrapper function, and tells it where the fifth input belongs. The fifth input, i.e. the one that will be perturbed, is given to
dianna.explain_image
. The wrapper function is given as preprocess_function todianna.explain_image
and handles reconstructing the proper input shape.