How to mention no grad in the following situation?

I have a tensor x and a parameter tensor R. I want to store the matrix multiplication of x and R into z and also make sure that z does not require any gradients. How do I do that? Please help

z = x @ self.R

Wrapping the operation into a no_grad() context would work:

R = nn.Parameter(torch.randn(1, 1))
x = torch.randn(1, 1)

with torch.no_grad():
    z = x @ R
> None
1 Like