Merge pull request GoogleCloudPlatform#1719 from nick-stroud/ml-slurm…

…-base-image

Update ml-slurm base image

examples/README.md

@@ -258,7 +258,7 @@ to 256
 ### [ml-slurm.yaml] ![core-badge]
 
 This blueprint provisions an HPC cluster running the Slurm scheduler with the
-machine learning frameworks [PyTorch] and [TensorFlow] pre-installed on every
+machine learning frameworks PyTorch and TensorFlow pre-installed on every
 VM. The cluster has 2 partitions:
 
 * [A2 family VMs][a2] with the NVIDIA A100 GPU accelerator
@@ -295,8 +295,7 @@ sbatch -N 1 torch_test.sh
 When you are done, clean up the resources in reverse order of creation:
 
 ```text
-terraform -chdir=ml-example/cluster destroy
-terraform -chdir=ml-example/primary destroy
+./ghpc destroy ml-example
 ```
 
 Finally, browse to the [Cloud Console][console-images] to delete your custom

examples/ml-slurm.yaml

@@ -134,7 +134,7 @@ deployment_groups:
       omit_external_ip: false
       source_image_project_id: [schedmd-slurm-public]
       # see latest in https://github.com/SchedMD/slurm-gcp/blob/master/docs/images.md#published-image-family
-      source_image_family: schedmd-v5-slurm-22-05-9-debian-11
+      source_image_family: slurm-gcp-5-7-debian-11
       # You can find size of source image by using following command
       # gcloud compute images describe-from-family <source_image_family> --project schedmd-slurm-public
       disk_size: $(vars.disk_size_gb)