Not able to connect to remote cluster with dask_ml.joblib being down
Open
- Dominant language
- Python
- Stars
- 951
- Forks
- 262
- PR merge metrics
- No merged PRs in 30d
Description
I'm getting an error with dask_ml version 1.2.0 -- module dask_ml.joblib not found. I've tried alternatively connecting per the suggestion here-- https://github.com/dask/dask-ml/issues/591, but instead get a TCP stream is closed error. Is there an underlying bug preventing the remote connection? Can anyone confirm connection using the below code?
'from dask.distributed import Client
import joblib
client` = Client(processes=False) # create local cluster
client = Client("scheduler-address:8786") # or connect to remote cluster'
with joblib.parallel_backend('dask'):
# Your scikit-learn code`
Contributor guide
Assessment
This issue has not been assessed yet.