GitLab

Hi @mbertoni I have been reading the documentation of the estimator we use for binary classification https://www.tensorflow.org/api_docs/python/tf/contrib/estimator/binary_classification_head and I just realized that, even though it accepts a weight_column (to down weight or boost examples during training) it is not provided when creating the Adanet object.

That is a problem since sometime we have to deal with unbalance data so we need means to account for it. I think that the best option is to let Adanet_wrap look for the "weight_colum" from the TrainTest object when initialized (and if it does not exists set it as None). What do you think?