Merge pull request #124 from AI-Hypercomputer/ninggu-patch-README-llama3.1-405b

Update llama3.1-405b 32 node README.md

Author: tonyjohnchen

training/a4x/llama3-1-405b/nemo-pretraining-gke/32node-FP8CS-GBS2048/recipe/README.md

@@ -83,7 +83,7 @@ gcloud container clusters get-credentials $CLUSTER_NAME --region $CLUSTER_REGION
 
 ### Configure and submit a pretraining job
 
-#### Using 16 node (64 gpus) fp8 precision
+#### Using 32 node (128 gpus) fp8 precision
 To execute the job with the default settings, run the following command from
 your client: