tensorflow - how do I represent an input tensor with multiple data types? -
i have far used tensor flow input tensor of 1 data type, example, float32.
i want have input can represented table rows , column such:
column0 column1 column2 column3 ... columnn person0 catx float float cat0 person1 caty float float cat1 person2 catz float float cat2 . . . personn cataa float float cat00
a mixture fo floats , categorical values 1 hot encode. cat placeholder strings drawn restricted domain of values such as, say, unique identifiers of machine parts. categorical columns this. how create input tensor these sorts of mixed inputs? possible 1 hot encodings way well?
Comments
Post a Comment