Skip to main content

NVIDIA VSS (Video Search & Summarization) Creation Guide

This guide explains how to create and configure NVIDIA VSS (Video Search & Summarization) through the Cordatus Platform.
NVIDIA VSS is an advanced AI solution that enables semantic search, event detection, and video summarization over video streams or recorded video content.

VSS analyzes video data to understand objects, actions, and contextual information, allowing users to efficiently search and review large-scale video datasets.
One of the key features included in NVIDIA VSS is Event Reviewer, which allows users to detect, classify, and review specific events within video content.

See details → Application Hub Overview | Application Hub Quickstart | Standard Application Launch Guide | Container Management Guide | User Models Guide

About NVIDIA VSS​

NVIDIA VSS (Video Search & Summarization) is a powerful AI-driven platform designed for intelligent video analysis and content management.

Key capabilities include:

  • Semantic Search: Query video content using natural language
  • Event Detection: Automatically identify specific objects, actions, or behaviors
  • Video Summarization: Generate concise summaries of long video content
  • Event Reviewer: Retroactively review critical moments using Computer Vision (CV) models

The Cordatus Platform simplifies the deployment of NVIDIA VSS by providing an intuitive interface for configuring all required components, including VLM, LLM, Embedding, and Rerank models.

Steps to Launch NVIDIA VSS​

  1. From the left-side menu, click Containers.
  2. From the dropdown submenu, select Applications.
  3. Choose NVIDIA VSS from the list of applications and navigate to its Detail page.

Up to the Advanced Settings section, the flow is identical to other application creation processes in Cordatus.

Device and Version Selection​

  1. Click Start Application.
  2. In the modal window, select the device on which you want to run NVIDIA VSS and connect to it.
  3. Proceed to the Version step, where available Docker image versions are displayed.
    • If the image is already installed on your device, a downloaded icon appears.
    • If not, a to be downloaded icon will be shown.

Advanced Settings​

The final step takes you to the Advanced Settings page, where you can configure all parameters for launching your NVIDIA VSS deployment.
NVIDIA VSS deployment consists of 5 main configuration steps.

Advanced Settings Details​

VSS Settings​

In this step, you configure the parameters required to generate the docker run command for the container that will be created from the NVIDIA VSS image.

You can choose between:

  • NVIDIA VSS only
  • NVIDIA VSS with Event Reviewer enabled

Event Reviewer Configuration​

If Event Reviewer is enabled, you must provide:

  • MODEL_ROOT_DIR: Directory path for the CV model
  • GDINO_MODEL_PATH: Path to the Grounding DINO model

If you use the default CV model, set both values to none.

4.2 VLM Settings (Vision Language Model)​

You can choose one of the following options:

  • Create a New VLM Application:
    Deploy a new Vision Language Model container alongside VSS.

    When creating a new application, you can select from three model sources:

    1. Cordatus Models: Pre-tested and pre-registered models within the Cordatus system

    2. User Models: Use models you have previously defined in your Cordatus account

      💡 Note: To transfer a model from Device X to Device Y, add it to User Models first.
      Learn more: User Models and Model Transfer Guide

      Cordatus enables automatic volume configuration and model transfer between devices on the same network.
      Configure model paths via Metrics > Model Info for Huggingface, Ollama, and NVIDIA NIM.

  • Assign an Existing VLM Application:
    If you have previously created a VLM application through Cordatus on:

    • The same device
    • A different device defined in your Cordatus account on same local network

    You can directly select the existing application using this option.
    Simply choose the application you wish to use and proceed to the next step.

  • Remote VLM Application:
    This option is for users who want to use an external (remote) VLM application.

    In this step:

    • Enter the URL of the remote VLM application
    • After entering the URL, select the model you wish to use from the next input field
    • If the remote VLM application requires a Token for authentication:
      • You can select from tokens already defined in your Cordatus account
      • If the token is not yet defined, you must add it to your Cordatus account first

4.3 LLM Settings​

This section follows the same configuration flow as VLM Settings but applies to the Large Language Model (LLM) component.

You can:

  • Create a new LLM application
  • Assign an existing LLM application
  • Use a remote LLM application

4.4 Embed Settings​

This section follows the same configuration flow as VLM Settings but applies to the Embedding Model component.

You can:

  • Create a new Embedding model application
  • Assign an existing Embedding model application
  • Use a remote Embedding model application

4.5 Rerank Settings​

This section follows the same configuration flow as VLM Settings but applies to the Rerank Model component.

You can:

  • Create a new Rerank model application
  • Assign an existing Rerank model application
  • Use a remote Rerank model application

5. Launching the Application​

Once you've configured all five components (VSS, VLM, LLM, Embed, and Rerank),
the Start Environment button will become active.

  • Cordatus will request your Sudo Password to authorize container creation.
  • If the selected Docker image is not present on your device, Cordatus will prompt you to confirm whether it should be downloaded.
  • If already available, Cordatus will start all containers automatically using your defined configuration.

Cordatus will automatically create all required containers, establish network connections between components, and configure the complete NVIDIA VSS pipeline.

You can monitor container statuses under:

  • Applications > Containers, or
  • The Containers section in the main sidebar.