I am trying to sample from the gamma distribution and would like to “fix” the sample I get.

What I do is:

from scipy.stats import gamma
gamma.rvs(a = 20, loc = 0, scale = 1/2, size = 10)

and this gives me a different array of values, each time I run the code…

I think that using rng could “freeze” the array once sampled, but browsing the maybe(?) appropriate page HERE confuses me on how to concretely implement the rng instruction in my code lines.

Could someone explain what would be the best way?