Can I duplicate parameter tensor whose values are the same across duplicates while training?

So let’s say I have an input vector of size 100. Specifically, the input vector is size [50 + 50]. Then I want to create a parameter tensor of size [50 x 2] that is applied to the first 50 inputs and second 50 inputs. such that we get a vector of size [2 + 2] afterward. I want the first 50 inputs and last 50 inputs of the input vector to be multiplied by the same [50 x 2] parameter tensor and that this remains true while parameter updating. Can you “duplicate” a parameter tensor like so or is there a way to program this? I’m not sure if the “.clone()” function does what I want…

Thanks!