How to change the behavior of an activation function in PyTorch?

I need to change the characteristics of the sigmoid function, which is usually done via a parameter called the temperature.
Is there a way to do this in PyTorch? I noticed that the sigmoid has just one value as input.

Hi there,
Sigmoid is defined as
f(x) = A/(B+exp(-Cx)) in the most general case.
Modifying C is equivalent to premultiply the input.
Modifying A is equivalent to postmultiply the output

I would also add another parameter, beta, to shift the input, that is, to move it away from the center that appears at input 0.5. Something like

pred = torch.sigmoid(alpha*pred - beta)

will do.