Parallel learning using Dask

Question

Scikit-Learn already provides parallel computing on a single machine with Joblib.
But l want to use Dask how can l achieve this?

from dask.distributed import Client
client = Client()  

how do l proceed from this?

Answer 1

There is a snippet from the dask documentation, you can use both the dask syntax and sk"Scikit-learn" code

import dask_ml.joblib
from sklearn.externals.joblib import parallel_backend
with parallel_backend('dask'):
# Your normal scikit-learn code here

Additional documentation here: http://dask-ml.readthedocs.io/