Sampling posterior predictive from shifted Lognormal

Thanks for the quick reply!

I don’t see the draw_values method. Was it added recently? (currently running PyMC3 3.6)

Edit: OK, found it. Just saw the function is not member of the Distribution class as I thought …