Dask-ML with sklearn Random forest results in connection closure

Question

I am attempting to train a model using Dask-ML. My end goal is to make predictions on larger than memory datasets, so I am utilizing Dask\'s ParallelPostFit wrapper to train