sklearn needs the column order to be the same as the training dataset: scikit-learn/scikit-learn#7242
So there should be an endpoint accepting an array instead of a dict.
Another option could be to somehow provide a map to sort the columns (as xgboost does with get_booster().feature_names).
sklearnneeds the column order to be the same as the training dataset: scikit-learn/scikit-learn#7242So there should be an endpoint accepting an array instead of a dict.
Another option could be to somehow provide a map to sort the columns (as xgboost does with
get_booster().feature_names).