While applying the ml_sampler estimator function, I was using an array of ints (0/1 values) as the is_positive label which would give me wildly incorrect estimates on a known population. As soon as I cast the lables as type bool, the estimator worked as expected. This is just a heads up that an array of 0/1 int values will not work in the estimator function, it must be explicitly as type bool to work.
While applying the ml_sampler estimator function, I was using an array of ints (0/1 values) as the is_positive label which would give me wildly incorrect estimates on a known population. As soon as I cast the lables as type bool, the estimator worked as expected. This is just a heads up that an array of 0/1 int values will not work in the estimator function, it must be explicitly as type bool to work.