How to implement conv:clone in pytorch

Hi,
I am reproducing a torch model to a pytorch model. but i have difficulties in how to share the parameters and the gradParameters in two layers.
Here is torch scipt,

.
Thanks for any help.