How-to Guides

This section collects guides for Vision AI sample applications.

• Run Multiple Apps
  • Prerequisites
  • Overview
  • Set up the Applications
  • Deploy the Applications
    • Deploy all the application instances
  • Start AI Pipelines
    • Start pipeline for all instances
    • Start pipeline for a particular instance only
    • Start pipeline for a particular instance from a custom payload.json
  • Monitor Applications
    • Check Pipeline Status
    • View Container Logs
  • Stop Applications
    • Stop Pipeline Instances
    • Stop Docker Containers
• Manage Pipelines
  • Create Pipelines
  • Start the Pipeline
  • Get Statistics of the Running Pipelines
  • Stop the Pipeline
  • Additional Usage
    • Frame Batching
• Run Multiple AI Pipelines
  • Steps
• Use GPU For Inference
  • Prerequisites
    • Provide GPU access to the container
    • Hardware specific encoder/decoders
    • GPU specific element properties
    • Selecting the GPU render device of your choice if there is more than one GPU device on the system
  • Tutorial on how to use GPU specific pipelines
  • Deploying with Helm
    • Intel® GPU K8S Extension
• Use NPU For Inference
  • Prerequisites
    • Provide NPU access to the container
    • Hardware specific encoder/decoders
    • NPU specific element properties
  • Tutorial on how to use NPU specific pipelines
    • Steps
  • Deploying with Helm
    • Intel NPU K8S Extension
• Change Video Source
• Scale Video Resolution
• Publish Frames to S3 Storage
  • Steps
• Publish Frames to OPC UA Server
• Enable MLOps
  • Contents
    • Prerequisites
    • Steps
  • Additional resources
    • Downloading models from Geti™ Server
• Start MQTT Publisher
• Run Benchmarks
  • Contents
    • Prerequisites
    • Benchmark Script Usage
    • Configuration
    • Recommended Pipeline Parameters
    • Steps to run benchmarks
    • Understanding Results
    • Troubleshooting
• Use your own model and Video
  • File Location
  • Docker compose deployment
  • Helm chart deployment
• View Telemetry Data
  • End the demonstration
• Run in an Air-Gapped System
  • Prerequisites
  • Configure for Air-Gapped System
  • Start the Application
  • Stop the Application