I am trying to figure out which scikit-learn estimators training is possible to parallelize with dask. From the dask documentation I understood that if the