opendatahub-io · openshift-merge-robot · May 25, 2023 · May 19, 2023 · May 24, 2023
diff --git a/opendatahub/docs/get-started-odh-modelserving.md b/opendatahub/docs/get-started-odh-modelserving.md
@@ -1,29 +1,30 @@
-# Getting Started
+# Getting Started with ODH ModelMesh Serving
 
-## Prerequisites
+The provided deploy script allows you to quickly run OpenDataHub ModelMesh Serving with a provisioned `etcd` server. This deploy script is intended for experimentation or development purposes only. It is not intended for production purposes.
 
-- **OpenShift cluster** - A OpenShift cluster is required. You will need `cluster-admin` or `dedicated-admin` authority in order to complete all of the prescribed steps.
+## Prerequisites
 
-- **OpenShift cli** - The installation will occur via the terminal using [oc](https://docs.openshift.com/container-platform/4.11/cli_reference/openshift_cli/getting-started-cli.html#cli-installing-cli_cli-developer-commands).
+- You have `cluster-admin` access to an OpenShift cluster.
 
-- **Model storage** - The model files have to be stored in a compatible form of remote storage or on a Kubernetes Persistent Volume. 
+- You have installed the **OpenShift CLI** as described in [Installing the OpenShift CLI by downloading the binary](https://docs.openshift.com/container-platform/4.11/cli_reference/openshift_cli/getting-started-cli.html#cli-installing-cli_cli-developer-commands).
 
-We provide an deploy script to quickly run OpenDataHub ModelMesh Serving with a provisioned etcd server. This may be useful for experimentation or development but should not be used in production.
+- Your model files are stored in a compatible form of remote storage or on a Kubernetes persistent volume. 
 
 ## Namespace Scope
 
-ModelMesh Serving can be used in either cluster scope or namespace scope mode but OpenDataHub ModelServing only support namespace scope mode.
-
-- **Namespace scope mode** - All of its components must exist within a single namespace and only one instance of ModelMesh Serving can be installed per namespace. Multiple ModelMesh Serving instances can be installed in separate namespaces within the cluster.
+While ModelMesh Serving is available in either cluster scope or namespace scope mode, OpenDataHub ModelServing only supports namespace scope mode.
 
+With Namespace scope, you can configure more than one ModelMesh Serving instance on a cluster. However, you must configure each instance of ModelMesh Serving in a separate namespace and all of the ModelMesh Serving instance components must exist within that single namespace. 
 
 ## Deployed Components
 
-|            | Type             | Pod                        | Count   | Default CPU request/limit per-pod | Default mem request/limit per-pod          |
-| ---------- | ---------------- | -------------------------- | ------- | --------------------------------- | ------------------------------------------ |
-| 1          | Controller       | Modelmesh Controller pod  | 3       | 50m / 1                           | 96Mi / 2Gi                               |
-| 2          | Object Storage   | MinIO pod (optional)       | 1       | 0m / 0m                       | 0Mi / 0Mi                              |
-| 3          | Metastore        | ETCD pod                   | 1       | 200m / 300m                       | 100Mi / 200Mi                              |
-| 4          | Built-in Runtime | The OVMS Runtime Pods      | 0 \(\*) | 500m / 5     | 1Gi / 1Gi |
-| 5          | ODH Model Controller   | ODH Model Controller pod        | 3       | 10m / 500m                       | 64Mi / 2Gi                              |
-| **totals** |                  |                            | 3       | 880m / 9.4                        | 1.58Gi / 13.2Gi                             |
+The following table describes the components deployed for each ModelMesh Serving instance.
+
+| Component Type             | Pod Name                       | Number of Pods | Default CPU Request/Limit per Pod | Default Memory request/Limit per Pod          |
+| ---------------- | -------------------------- | ------- | --------------------------------- | ------------------------------------------ |
+| Controller       | Modelmesh Controller pod  | 3       | 50m / 1                           | 96Mi / 2Gi                               |
+| Object Storage   | MinIO pod (optional)       | 1       | 0m / 0m                       | 0Mi / 0Mi                              |
+| Metastore        | ETCD pod                   | 1       | 200m / 300m                       | 100Mi / 200Mi                              |
+| Built-in Runtime | The OVMS Runtime Pods      | 0 \(\*) | 500m / 5     | 1Gi / 1Gi |
+| ODH Model Controller   | ODH Model Controller pod        | 3       | 10m / 500m                       | 64Mi / 2Gi                              |
+| **Totals**                   |                            | 3       | 880m / 9.4                        | 1.58Gi / 13.2Gi                             |
diff --git a/opendatahub/quickstart/README.md b/opendatahub/quickstart/README.md
@@ -1,28 +1,30 @@
-# Quick Start 
+# Overview of the OpenDataHub's ModelServing Quick Starts
 
-This documentation aims to assist you in utilizing OpenDataHub's ModelServing effectively. It focuses on explaining new features introduced in the OpenDataHub modelmesh or odh-model-controller and provides relevant examples to enhance comprehension.
+The purpose of these quick starts is help you learn how to use OpenDataHub's ModelServing. They describe features in the OpenDataHub ModelMesh or odh-model-controller and provide relevant examples.
 
-Please note that while these documents have been verified for accuracy at the time of their creation, there is a possibility that manifests or scripts may become outdated and not function as intended over time.
+**Note:** These quick starts have been verified for accuracy at the time of their creation, but the manifests or scripts might become outdated and not function as originally intended.
 
-Our objective is to provide comprehensive and professional guidance to ensure seamless utilization of OpenDataHub's ModelServing.
+## List of quick starts
 
-The folder structure for the Quickstart is as follows:
+- [Sample Model Deployment](./basic/README.md)
+- [Sample Model Deployment and Autoscaler](./hpa/README.md)
+- [Sample Model Deployment by using a Persistent Volume Claim](./pvc/README.md) 
+
+## Quick start files
+
+Each quick start folder contains the following files:
 ~~~
-|-- basic
-     |-- deploy.sh   # Script to deploy ODH Modelmesh and Quickstart objects
-     |-- clean.sh    # Script to delete all Quickstart objects
-     |-- README.md   # Documentation providing an explanation of the Quickstart
+|-- deploy.sh   # Script to deploy the OpenDataHub ModelMesh and all quick start objects
+|-- clean.sh    # Script to delete all quick start objects
+|-- README.md   # Documentation that describes how to run the quick start
 ~~~
 
-## Requirement
+## Requirements for running the quick starts
+
 - OpenShift Cluster 4.11+
 - Default StorageClass
 - OpenShift CLI 4.11+
-- At least 8 vCPU and 16 GB memory. For more details, please see [here](../docs/get-started-odh-modelserving.md).
-- User have cluster-admin role.
+- At least 8 vCPU and 16 GB memory. For more details, see [Getting Started with ODH ModelMesh Serving](../docs/get-started-odh-modelserving.md).
+- You must have `cluster-admin` access to the OpenShift cluster.
 
-## Quick Start List
 
-- [Sample Model Deployment](./basic/README.md)
-- [Autoscaler Feature](./hpa/README.md)
-- [PVC Feature](./pvc/README.md) 
diff --git a/opendatahub/quickstart/basic/README.md b/opendatahub/quickstart/basic/README.md
@@ -1,22 +1,12 @@
-# Quick Start Guide - Sample Model Deployment
+# Quick Start - Sample Model Deployment
 
-This Quick Start Guide provides instructions for deploying the OpenDataHub Modelserving component along with NFS Provisioner, Minio, and a sample PVC. Additionally, it automatically deploys a sample model and allows you to test OpenDataHub Modelserving using its inference service.
+Welcome to the quick start for deploying a sample model and testing OpenDataHub ModelServing by using its provided inference service.
 
-## Prerequisites
-
-You can check prerequisites from [this doc](../README.md)
-
-## Install OpenDataHub ModelServing
-
-Please refer to [this doc](../docs/modelmesh-install.md)
+## Description of the inference service manifest YAML files
 
-## Deploy a Sample Model
-
-The `deploy.sh` also deployed a sample model so you don't need to create additional inference service. However, to help you understand, here's a brief description of the inference service manifests yaml.
+There are two inference service manifest YAML files that this quick start uses to specify a model path: `storageUri` or `storagePath`
 
-These are the yaml files that quick start used. There are 2 ways to specify a model path: `storageUri`, `storeagePath`
-
-- **storeagePath**
+- **storagePath**
 ~~~
 apiVersion: serving.kserve.io/v1beta1
 kind: InferenceService
@@ -35,7 +25,7 @@ spec:
         path: onnx/mnist.onnx
 ~~~
 
-- **storeageUri**
+- **storageUri**
 ~~~
 apiVersion: serving.kserve.io/v1beta1
 kind: InferenceService
@@ -52,42 +42,70 @@ spec:
       storageUri: s3://modelmesh-example-models/onnx/mnist.onnx
 ~~~
 
-### How to check model deployment status
-You can check if your model is ready by checking the inference service, where you can get the `gRPC` URL to access your model.
+## Prerequisites
+
+- Verify that you meet the requirements for running the quick starts listed in [Overview of the OpenDataHub's ModelServing Quick Starts](../README.md).
+- Install OpenDataHub ModelServing as described in [Installing OpenDataHub ModelServing](../common_docs/modelmesh-install.md).
+
+## Deploy a Sample Model
+
+Deploy the sample model:
+~~~
+./deploy.sh
+~~~
+
+## Check the model deployment status
+
+1. Check whether your model is ready by getting the OpenDataHub ModelServing's inference service:
 ~~~
 $ oc get isvc -n modelmesh-serving
+~~~
+You should see a result similar to the following:
+~~~
 NAME                 URL                                               READY   PREV   LATEST   PREVROLLEDOUTREVISION   LATESTREADYREVISION   AGE
 example-onnx-mnist   grpc://modelmesh-serving.modelmesh-serving:8033   True                                                                  4m
 ~~~
+Note that this result includes the `gRPC` URL that you can use to access the model.
 
-For the `HTTP` URL, you can check routes.
+2. To obtain the `HTTP` URL for the model, use the command to get routes:
 ~~~
 $ oc get routes
+~~~
+You should see a result similar to the following:
+~~~
 NAME                 HOST/PORT                                                                       PATH                            SERVICES            PORT   TERMINATION     WILDCARD
 example-onnx-mnist   example-onnx-mnist-modelmesh-serving.apps.jlee-test.l9ew.p1.openshiftapps.com   /v2/models/example-onnx-mnist   modelmesh-serving   8008   edge/Redirect   None
 ~~~
-## Perform an inference request
 
-Now that a model is loaded and available, you can then perform inference. OpenDataHub ModelServing has another controller `odh-model-controller` that is responsible for creating OpenShift Route for the model. You can manage the feature with the annotation `enable-route: "true"` in the ServingRuntime. Plus, the controller also manages authentication with the annotation  `enable-auth: "true"` in the ServingRuntime. (By default, both features are disabled but for this quick start, `enable-route` was set to `true`)
+## Perform inference requests
+
+After the model has deployed, you can perform inference requests. OpenDataHub ModelServing includes the `odh-model-controller` controller that is responsible for creating an OpenShift Route for the model and for authentication. These features are set with the `enable-route` and `enable-auth` ServingRuntime annotations. By default, both features are disabled (the annotations are set to `false`), but for this quick start, `enable-route` is set to `true`.
 
-**Curl Test with no authentication enabled**
+The following `curl` examples demonstrate how to perform inference requests.
+
+**Curl test without authentication enabled**
 ~~~
 export HOST_URL=$(oc get route example-onnx-mnist -ojsonpath='{.spec.host}' -n ${TEST_MM_NS})
-export HOST_PATH=$(oc get route example-onnx-mnist  -ojsonpath='{.spec.path}' -n ${TEST_MM_NS})
+export HOST_PATH=$(oc get route example-onnx-mnist -ojsonpath='{.spec.path}' -n ${TEST_MM_NS})
 
-curl   --silent --location --fail --show-error --insecure https://${HOST_URL}${HOST_PATH}/infer -d  @${COMMON_MANIFESTS_DIR}/input-onnx.json
+curl --silent --location --fail --show-error --insecure https://${HOST_URL}${HOST_PATH}/infer -d  @${COMMON_MANIFESTS_DIR}/input-onnx.json
 
 {"model_name":"example-onnx-mnist__isvc-b29c3d91f3","model_version":"1","outputs":[{"name":"Plus214_Output_0","datatype":"FP32","shape":[1,10],"data":[-8.233053,-7.7497034,-3.4236815,12.3630295,-12.079103,17.266596,-10.570976,0.7130762,3.321715,1.3621228]}]}
 ~~~
 
-**Note**: If Authentication is enabled, the route port should be `8080`.
+**Note**: If authentication is enabled, the route port should be `8080`.
+
 ~~~
 $ oc get route
+~~~
+You should see a result similar to the following:
+~~~
 NAME                 HOST/PORT                                                                       PATH                            SERVICES            PORT   TERMINATION     WILDCARD
 example-onnx-mnist   example-onnx-mnist-modelmesh-serving.apps.jlee-test.l9ew.p1.openshiftapps.com   /v2/models/example-onnx-mnist   modelmesh-serving   8008   edge/Redirect   None
 ~~~
 
-**gRPC Curl Test using port-forward**
+**gRPC Curl test by using port-forward**
+
 ~~~
 oc port-forward --address 0.0.0.0 service/modelmesh-serving 8033 -n ${TEST_MM_NS} 
 
@@ -103,10 +121,9 @@ grpcurl \
 cd -    
 ~~~
 
-**Curl Test with authentication enabled**
-
-Here we also show the case of enabling authentication for testing purposes. This can be done by simply setting the annotation 'enable-auth` in the servingrungime to true. When this feature is enabled, the token of the user who has access to this route should be sent along with it.
+**Curl test with authentication enabled**
 
+You can enable authentication for testing purposes by setting the `enable-auth` annotation in the ServingRuntime to `true`. When you enable authentication, you should also send the token of the user with access to the route.
 ~~~
 # Enable Auth for OVMS ServingRuntime
 oc apply -f  ${COMMON_MANIFESTS_DIR}/sa_user.yaml -n ${TEST_MM_NS}
@@ -121,13 +138,16 @@ curl  -H "Authorization: Bearer ${Token}" --silent --location --fail --show-erro
 {"model_name":"example-onnx-mnist__isvc-b29c3d91f3","model_version":"1","outputs":[{"name":"Plus214_Output_0","datatype":"FP32","shape":[1,10],"data":[-8.233053,-7.7497034,-3.4236815,12.3630295,-12.079103,17.266596,-10.570976,0.7130762,3.321715,1.3621228]}]}
 ~~~
 
-**Note**: If Authentication is enabled, the route port should be `8443`.
+**Note**: If authentication is enabled, the route port should be `8443`.
 ~~~
 $ oc get route
+~~~
+You should see a result similar to the following:
+~~~
 NAME                 HOST/PORT                                                                       PATH                            SERVICES            PORT   TERMINATION          WILDCARD
 example-onnx-mnist   example-onnx-mnist-modelmesh-serving.apps.jlee-test.l9ew.p1.openshiftapps.com   /v2/models/example-onnx-mnist   modelmesh-serving   8443   reencrypt/Redirect   None
 ~~~
 
 ## Cleanup
 
-Please refer to [this doc](../docs/modelmesh-cleanup.md)
+Follow the steps in [Cleaning up an OpenDataHub ModelServing installation](../common_docs/modelmesh-cleanup.md).
diff --git a/opendatahub/quickstart/common_docs/modelmesh-cleanup.md b/opendatahub/quickstart/common_docs/modelmesh-cleanup.md
@@ -1,15 +1,13 @@
-# Cleanup 
+# Cleaning up an OpenDataHub ModelServing installation
 
-## Cleanup Quick Start
+Run one of the following commands to cleanup up the Quick Start files:
 
-This will delete modelmesh and modelmesh test namespace(minio,pvc)
+- If you want to try another quick start, run this command to delete the `modelmesh` and modelmesh test namespaces (`minio` and `pvc`):
 ~~~
 ./cleanup.sh
 ~~~
 
-## Cleanup all components 
-
-This will delete modelmesh, modelmesh test namespace(minio,pvc) and nfs provisioner. If you want to try other quick start guide, please use above command.
+- If you are done with all of the quickstarts, run this command to deletes the `modelmesh` and modelmesh test namespaces (`minio` and `pvc`) and NFS provisioner:
 ~~~
 C_FULL=true ./cleanup.sh
 ~~~