[Feature Request] ParallelPostFit multiple scorers, scores with dask...

Hi,

On `[ParallelPostFit documentation](https://ml.dask.org/modules/generated/dask_ml.wrappers.ParallelPostFit.html#dask_ml.wrappers.ParallelPostFit)` it is told that `scoring` parameter is

> For evaluating multiple metrics, either give a list of (unique) strings or a dict with names as keys and callables as values.

Howevere, this isn´t implemented by now. Are you planning on implementing them?

Are you planning on make a dask implementation of sklearn metrics for large datasets?

I can make a pull request modifying the documentation or adding this things. What do you think?

Best regards

Begin with the linked ParallelPostFit documentation and its scoring parameter. Compare the documented multi-metric forms with current behavior, then determine whether the scope is documentation, ParallelPostFit support, or Dask-compatible metrics. Done requires an agreed scope plus corresponding tests or documentation, but the issue names no files or tests.

Written by the indexing model from the issue text.