The input and output data type for the Function Approximation must be real-valued and is controlled by the TT_DATA template parameter.
In AIE, this parameter can be configured as int16, int32, or float. In AIE-ML, this parameter can be configured as int16, int32, float, or bfloat16.