How it works...

All arithmetic operations of matrices like add, sub, p, multiply (elementwise multiplication), mod, and cross require that the two tensor matrices should be of the same data type. In case this is not so they will produce an error. We can use tf.cast() to convert Tensors from one data type to another.