rspamd: Bayesian filter doesn't appear to be working

I am setting up rspamd. It seems to work quite effectively, however one thing I've noticed is this log entry in my rspamd.log every time I receive a message: > 2019-05-03 11:36:24 #13015(normal) ; task; bayes_classify: skip classification as ham class has not enough learns: 0, 200 required However, the number of learns (currently zero) does not appear to be going up. Is there something I should have done or configured to make it start learning?