Can I use DASK to parallelize computation of scikit-learn models on a single computer that do not have n_par parameter?

Question

Can I use DASK to parallelize computation on a single computer of scikit-learn models that do not have n_par parameter? I specifically interested to parallelize MLPRegresso