Various fixes to Inferentia lab

Author: niallthomson

website/docs/aiml/inferentia/compile.md

@@ -77,14 +77,14 @@ $ kubectl logs -l app.kubernetes.io/instance=karpenter -n kube-system -f | jq
 }
 ```
 
-The Pod should be scheduled on the node provisioned by Karpenter. Check if the Pod is in it's ready state:
+The Pod should be scheduled on the node provisioned by Karpenter. Check if the Pod is in its ready state:
 
 ```bash timeout=600
 $ kubectl -n aiml wait --for=condition=Ready --timeout=10m pod/compiler
 ```
 
 :::warning
-This command can take up to 10 min.
+This command can take up to 10 minutes.
 :::
 
 Next, copy the code for compiling a model on to the Pod and run it:

website/docs/aiml/inferentia/inference.md

@@ -15,7 +15,7 @@ $ echo $AIML_DL_INF_IMAGE
 
 This is a different image than we used for training and has been optimized for inference.
 
-Now we can deploy a Pod for inference. This is the the manifest file for running the inference Pod:
+Now we can deploy a Pod for inference. This is the manifest file for running the inference Pod:
 
 ::yaml{file="manifests/modules/aiml/inferentia/inference/inference.yaml" paths="spec.nodeSelector,spec.containers.0.resources.limits"}
 
@@ -63,7 +63,7 @@ $ kubectl logs -l app.kubernetes.io/instance=karpenter -n kube-system -f | jq
 ...
 ```
 
-The inference Pod should be scheduled on the node provisioned by Karpenter. Check if the Pod is in it's ready state:
+The inference Pod should be scheduled on the node provisioned by Karpenter. Check if the Pod is in its ready state:
 
 :::note
 It can take up to 12 minutes to provision the node, add it to the EKS cluster, and start the pod.
@@ -96,7 +96,7 @@ This output shows the capacity this node has:
 }
 ```
 
-We can see that this node as a `aws.amazon.com/neuron` of 1. Karpenter provisioned this node for us as that's how many neuron the Pod requested.
+We can see that this node has an `aws.amazon.com/neuron` of 1. Karpenter provisioned this node for us as that's how many Neuron cores the Pod requested.
 
 ### Run inference
 

website/docs/aiml/inferentia/karpenter.md

@@ -9,7 +9,7 @@ In this section we will configure Karpenter to allow the creation of Inferentia
 You can learn more about Karpenter in the [Karpenter module](../../fundamentals/compute/karpenter/index.md) that's provided in this workshop.
 :::
 
-Karpenter has been installed in our EKS cluster, and runs as a deployment:
+Karpenter has been installed in our EKS cluster, and runs as a Deployment:
 
 ```bash
 $ kubectl get deployment -n kube-system
@@ -32,4 +32,4 @@ $ kubectl kustomize ~/environment/eks-workshop/modules/aiml/inferentia/nodepool
   | envsubst | kubectl apply -f-
 ```
 
-Now the NodePool is ready for the creation for our training and inference Pods.
+Now the NodePool is ready for the creation of our training and inference Pods.

website/docs/aiml/inferentia/wrapup.md

@@ -1,11 +1,11 @@
 ---
 title: "Real world implementation"
-sidebar_position: 40
+sidebar_position: 50
 ---
 
-In the previous sections we've seen how we can use Amazon EKS to build models for AWS Inferentia and deploy models on EKS using Inferentia nodes. In both these examples we've executed Python code inside our containers from our command-line. In a real world scenario we do not want to run these commands manually, but rather have the container execute the commands.
+In the previous sections we've seen how we can use Amazon EKS to train models for AWS Inferentia and deploy models on EKS using Inferentia nodes. In both these examples we've executed Python code inside our containers from our command line. In a real world scenario we do not want to run these commands manually, but rather have the container execute the commands.
 
-For building the model we would want use the DLC container as our base image and add our Python code to it. We would then store this container image in our container repository like Amazon ECR. We would use a Kubernetes Job to run this container image on EKS and store the generated model to S3.
+For training the model we would want to use the DLC container as our base image and add our Python code to it. We would then store this container image in our container repository like Amazon ECR. We would use a Kubernetes Job to run this container image on EKS and store the generated model to S3.
 
 ![Build Model](./assets/CreateModel.webp)