special.softmax() should support pd.DataFrame as input #12164
Description
Is your feature request related to a problem? Please describe.
when the input df is of type pandas.core.frame.DataFrame, softmax(df) doesn't work:
gives this error
C:\ProgramData\Anaconda3\lib\site-packages\pandas\core\ops\__init__.py in _align_method_FRAME(left, right, axis)
662 else:
663 raise ValueError(
--> 664 "Unable to coerce to DataFrame, shape "
665 f"must be {left.shape}: given {right.shape}"
666 )
Describe the solution you'd like
should be able to run softmax(df)
Describe alternatives you've considered
currently have to either run:
softmax(df.values)
or
softmax(np.array(df))
Additional context (e.g. screenshots)