Is there a reason, that in Tensorflow tensor-computing (math) functions, such as reduce_max, tanh, relu, etc. are only part of the tensorf
reduce_max
tanh
relu