I should mention that the outputs without the error message seem to be right and match the paper's stochastic versions, but the output with the warning does not. In other words, it does seem that something has gone wrong.