Skip to content

Latest commit

 

History

History
83 lines (60 loc) · 3.7 KB

pre-steps.md

File metadata and controls

83 lines (60 loc) · 3.7 KB
Raw

Setup

For the lab it is assumed that you are familiar with Vertex AI. Vertex AI is Google Cloud's Machine Learning Development and Model Management platform. Please read the documentation to learn about the full feature set.

These steps are for environment setup to get your code up and running.

  • Step 1: Create a User Managed Workbench

    Follow the steps in the following guide to setup the workbench Choose a default image. You do not need a large image as this will act as client to your code.

  • Step 2: Open the JupyterLab

    jupyter lab

  • Step 3: Open a Terminal window in the workbench terminal

  • Step 4: Download the zip of the repository

    Note: If you plan to use the CI/CD and build. Do not clone the repo, but download as zip from the repository download

  • Step 5: Upload the zip to your notebook

  • Step 6: Setup the Source repository

    • Navigate to the terminal window in Step 4 and execute following commands
      • pip3 install --upgrade google-cloud-aiplatform
    • Set the following environment variables
      • PROJECT_ID = "[YOUR PROJECT]" # @param {type:"string"}
      • INITIALS = "[YOUR INITIALS]"
      • REPO_NAME = "[YOUR REPO NAME]" #@param {type:"string"}
      • WORKING_DIR="[YOUR WORKING DIR]
      • gcloud source repos create $REPO_NAME
    • Browse the Google Cloud Source Repo to check the repo is created. Alternatively you can call
      • gcloud source repos list
      • mkdir $WORKING_DIR
      • cd $WORKING_DIR
      • gcloud source repos clone $REPO_NAME
    • Set the Git config informatio * git config --global user.email "[email protected]" * git config --global user.name "Your Name"

  • Step 7: Unzip the MLOpsCICD-master.zip file

    • cd
    • mkdir tmp
    • unzip MLOpsCICD-master.zip -d ~/tmp

  • Step 8: Copy the content of the ~/tmp/MLOpsCICD-master

    • cp -r ~/tmp/MLOpsCICD-master $WORKING_DIR/$REPO_NAME/

  • Step 9: Add the code to the Google Cloud Source Repo

    • cd $WORKING_DIR/$REPO_NAME/
    • git add . # Do not forget the dot.
    • git commit -m "Your message"
    • git push
    • Check in the cloud source repo if the code has been checked in

  • Step 10: Setup Cloud Build

    • Follow the following steps to Enable and setup Cloud Build

  • Step 11: Execute a Manual Build

    • The following step may error out as it uses my default bucket and region. Please update the IrisPipelineTemplate notebook, with the right
      • PROJECT_ID ="demogct"
      • BUCKET_LOC = "gs://demogct/vipipelines/"
      • PIPELINE_NAME="iris-vertex-pipeline"
      • REGION = "us-central1"

    This set up defaults to us-central1. If you want a different region please update the aiplatform initialization

    • aip.init(project=PROJECT_ID, staging_bucket=BUCKET_LOC) --> aip.init(project=PROJECT_ID,location=REGION staging_bucket=BUCKET_LOC)

  • gcloud builds submit --config cloudbuild.yaml --timeout=1000

  • Step 12: Setup Cloud Triggers

  • Step 13: Test Build Triggers

    • Make a change to the code and push the code to the code repository
    • Check the Build History page which should have updated with a new build

Return