Recommender Systems AI Workflow

NVIDIA AI Workflows

• Introduction
  
  • What are NVIDIA AI Workflows?
  • NVIDIA AI Workflow Components
  • Guides in This AI Workflow

Recommender Systems AI Workflow

• Technical Brief
  
  • Overview
  • NVIDIA AI Components
  • Additional Components

Next Item Prediction

• Deployment Guide
  
  • Overview
  • Step 1: Prerequisites
  • Step 2: Set Up Required Infrastructure
    
    • GPU-Enabled Hardware Infrastructure
    • Kubernetes Cluster
    • NVIDIA Cloud Native Service Add-On Pack
    • Workflow Components
  • Step 3: Install Workflow Components
    
    • Configure Keycloak
    • Generate an Access Token from Keycloak
    • Set Environment Variables Required For Deployment
    • Install Next Item Prediction Workflow Charts
  • Step 4: Run The Workflow
• Deployment/DevOps Customization Guide
  
  • Overview
  • MinIO Chart Overview
    
    • MinIO Chart Installation
    • Chart Values Explained (./deploy/charts/minio/values.yaml)
  • MLflow Chart Overview
    
    • MLflow Chart Installation
    • Chart Values Explained (./deploy/charts/mlflow/values.yaml)
  • Data Preparation Chart Overview
    
    • Data Preparation Chart Installation
    • Chart Values Explained (./deploy/charts/next-item-wf-data-prep/values.yaml)
  • Training Chart Overview
    
    • Training Chart Installation
    • Chart Values Explained (./deploy/charts/next-item-wf-data-train/values.yaml)
  • Inference Chart Overview
    
    • Inference Chart Installation
    • Chart Templates Explained
    • Chart Values Explained (/deploy/charts/next-item-wf-data-train/values.yaml)
• Developer Guide
  
  • Configuring Your Development Environment
    
    • Install Docker And Docker-compose
    • Download The Workflow Source Code
    • Developing Locally Using MLflow
    • Bulding The Docker Image
    • Running Scripts With Docker-compose
    • Mounting Data Into The Containers
  • Generating Sample Data For Model Training
  • Visualizing The Retraining Schedule
    
    • Test / Validation Split
  • (Synthetic) Data Preprocessing
    
    • Pre-grouping by session_id
    • Running The Data Prep Step
  • Feature Engineering & Model Training
    
    • Building the Docker Image
    • Feature Engineering With NVTabular
      
      • Input And Output Schemas
    • Model Training With Transformers4Rec
      
      • Loading Data And Feature Engineering
      • Model Definition
      • Training
  • Exporting For Serving
    
    • Debugging the model on Triton
  • Understanding Schemas
    
    • Inspecting The NVTabular Workflow Input Schema
    • Converting To Values And Offsets Format
    • Defining The HTTP Payload
  • Replacing The Sample Data With Your Training Data
    
    • Loading Data With NVTabular
      
      • Reading From Data Warehouses
  • Just Let Me Run The Code!

Appendix

• Authentication
  
  • Authentication in Microservices
  • Who Does the Job of Authentication?
  • JSON Web Tokens
  • Envoy Proxy
  • Envoy Config
• Keycloak Configuration
  
  • Keycloak Terminology
  • Steps for User Creation and Registration
  • Realm Creation
  • User Creation
  • Client Creation
  • Role Creation
  • Increase the Access Token Lifespan
  • Token Endpoint
  • Return to Workflow and Generate Access Token
• Using MLflow’s Model Registry
  
  • Setup and Further Resources

Support

• Support and Services

Notices

• Notices
  
  • Notice
  • Trademarks
  • Copyright