| tests | 
|
|---|
Deploy Ray clusters on NERSC supercomputers.
- Free software: GNU Lesser General Public License v3 (LGPLv3)
from nersc_cluster_deploy import deploy_ray_cluster
#Create Ray Cluster with dashboard
rayCluster = deploy_ray_cluster(
slurm_options = '-q regular -A elvis -t 01:00:00 -N 3'
job_setup = 'module load pytorch',
)
>>> Ray cluster running at x.x.x.x:6379
>>> Dashboard avaliable at https://jupyter.nersc.gov/user/elivs/proxy/localhost:8265/ with metrics
#Connect to cluster
import ray
ray.init(address='auto')You can install the latest version with:
pip install git+https://github.com/asnaylor/nersc_cluster_deploy.git@v0.4.0
Visit the documentation README for more information.
For examples see the tutorial repo.