Python: different results from random forest after fixing the random state

Question

I have the following code. I set the random state already. Everytime I do cross validation, it gives me a new set of optimal parameters. This just doesn\'t make sense to me.