sarman/dgx-spark-playbooks
1
0
Fork 0
mirror of https://github.com/NVIDIA/dgx-spark-playbooks.git synced 2026-04-22 18:13:52 +00:00
Code Issues Actions Packages Projects Releases Wiki Activity
316b9a41fa
dgx-spark-playbooks/nvidia/vss
History
GitLab CI 9cfb6e1735 chore: Regenerate all playbooks
2025-10-07 17:35:19 +00:00
..
README.md chore: Regenerate all playbooks 2025-10-07 17:35:19 +00:00

README.md

Video Search and Summarization

Run the VSS Blueprint on your Spark

Table of Contents

Overview

Basic idea

Deploy NVIDIA's Video Search and Summarization (VSS) AI Blueprint to build intelligent video analytics systems that combine vision language models, large language models, and retrieval-augmented generation. The system transforms raw video content into real-time actionable insights with video summarization, Q&A, and real-time alerts. You'll set up either a completely local Event Reviewer deployment or a hybrid deployment using remote model endpoints.

What you'll accomplish

You will deploy NVIDIA's VSS AI Blueprint on NVIDIA Spark hardware with Blackwell architecture, choosing between two deployment scenarios: VSS Event Reviewer (completely local with VLM pipeline) or Standard VSS (hybrid deployment with remote LLM/embedding endpoints). This includes setting up Alert Bridge, VLM Pipeline, Alert Inspector UI, Video Storage Toolkit, and optional DeepStream CV pipeline for automated video analysis and event review.

What to know before starting

  • Working with NVIDIA Docker containers and container registries
  • Setting up Docker Compose environments with shared networks
  • Managing environment variables and authentication tokens
  • Working with NVIDIA DeepStream and computer vision pipelines
  • Basic understanding of video processing and analysis workflows

Prerequisites

  • NVIDIA Spark device with ARM64 architecture and Blackwell GPU
  • FastOS 1.81.38 or compatible ARM64 system
  • Driver version 580.82.09 installed: nvidia-smi | grep "Driver Version"
  • CUDA version 13.0 installed: nvcc --version
  • Docker installed and running: docker --version && docker compose version
  • Access to NVIDIA Container Registry with NGC API Key
  • [Optional] NVIDIA API Key for remote model endpoints (hybrid deployment only)
  • Sufficient storage space for video processing (>10GB recommended in /tmp/)

Ancillary files

Time & risk

Duration: 30-45 minutes for initial setup, additional time for video processing validation

Risks:

  • Container startup can be resource-intensive and time-consuming with large model downloads
  • Network configuration conflicts if shared network already exists
  • Remote API endpoints may have rate limits or connectivity issues (hybrid deployment)

Rollback: Stop all containers with docker compose down, remove shared network with docker network rm vss-shared-network, and clean up temporary media directories.

Instructions

Step 1. Verify environment requirements

Check that your system meets the hardware and software prerequisites.

## Verify driver version
nvidia-smi | grep "Driver Version"
## Expected output: Driver Version: 580.82.09

## Verify CUDA version
nvcc --version
## Expected output: release 13.0

## Verify Docker is running
docker --version && docker compose version

Step 2. Clone the VSS repository

Clone the Video Search and Summarization repository from NVIDIA's public GitHub.

## Clone the VSS AI Blueprint repository
git clone https://github.com/NVIDIA-AI-Blueprints/video-search-and-summarization.git
cd video-search-and-summarization

Step 3. Run the cache cleaner script

Start the system cache cleaner to optimize memory usage during container operations.

## In another terminal, start the cache cleaner script.
## Alternatively, append " &" to the end of the command to run it in the background.
sudo sh deploy/scripts/sys_cache_cleaner.sh

Step 4. Set up Docker shared network

Create a Docker network that will be shared between VSS services and CV pipeline containers.

## Create shared network (may require sudo depending on Docker configuration)
docker network create vss-shared-network

Warning: If the network already exists, you may see an error. Remove it first with docker network rm vss-shared-network if needed.

Step 5. Authenticate with NVIDIA Container Registry

Log in to NVIDIA's container registry using your NGC API Key.

## Log in to NVIDIA Container Registry
docker login nvcr.io
## Username: $oauthtoken
## Password: <PASTE_NGC_API_KEY_HERE>

Step 6. Choose deployment scenario

Choose between two deployment options based on your requirements:

Deployment Scenario VLM (Cosmos-Reason1-7B) LLM (Llama 3.1 70B) Embedding/Reranker CV Pipeline
VSS Event Reviewer Local Not Used Not Used Local
Standard VSS (Hybrid) Local Remote Remote Optional

Proceed with Option A for Event Reviewer or Option B for Standard VSS.

Step 7. Option A

VSS Event Reviewer (Completely Local)

7.1 Navigate to Event Reviewer directory

Change to the directory containing the Event Reviewer Docker Compose configuration.

cd deploy/docker/event_reviewer/

7.2 Configure NGC API Key

Update the environment file with your NGC API Key. You can do this by editing the .env file directly, or by running the following command:

## Edit the .env file and update NGC_API_KEY
echo "NGC_API_KEY=<YOUR_NGC_API_KEY>" >> .env

7.3 Update the VSS Image path

Update VSS_IMAGE to nvcr.io/nvidia/blueprint/vss-engine-sbsa:2.4.0 in .env.

## Edit the .env file and update VSS_IMAGE
echo "VSS_IMAGE=nvcr.io/nvidia/blueprint/vss-engine-sbsa:2.4.0" >> .env

7.4 Start VSS Event Reviewer services

Launch the complete VSS Event Reviewer stack including Alert Bridge, VLM Pipeline, Alert Inspector UI, and Video Storage Toolkit.

## Start VSS Event Reviewer with ARM64 and SBSA optimizations
IS_SBSA=1 IS_AARCH64=1 ALERT_REVIEW_MEDIA_BASE_DIR=/tmp/alert-media-dir docker compose up

Note: This step will take several minutes as containers are pulled and services initialize. The VSS backend requires additional startup time.

7.5 Navigate to CV Event Detector directory

In a new terminal session, navigate to the computer vision event detector configuration.

cd video-search-and-summarization/examples/cv-event-detector

7.6 Update the NV_CV_EVENT_DETECTOR_IMAGE Image path

Update NV_CV_EVENT_DETECTOR_IMAGE to nvcr.io/nvidia/blueprint/nv-cv-event-detector-sbsa:2.4.0 in .env.

## Edit the .env file and update NV_CV_EVENT_DETECTOR_IMAGE
echo "NV_CV_EVENT_DETECTOR_IMAGE=nvcr.io/nvidia/blueprint/nv-cv-event-detector-sbsa:2.4.0" >> .env

7.7 Start DeepStream CV pipeline

Launch the DeepStream computer vision pipeline and CV UI services.

## Start CV pipeline with ARM64 and SBSA optimizations
IS_SBSA=1 IS_AARCH64=1 ALERT_REVIEW_MEDIA_BASE_DIR=/tmp/alert-media-dir docker compose up

7.8 Wait for service initialization

Allow time for all containers to fully initialize before accessing the user interfaces.

## Monitor container status
docker ps
## Verify all containers show "Up" status and VSS backend logs show ready state

7.9 Validate Event Reviewer deployment

Access the web interfaces to confirm successful deployment and functionality.

## Test CV UI accessibility (replace <NODE_IP> with your system's IP)
curl -I http://<NODE_IP>:7862
## Expected: HTTP 200 response

## Test Alert Inspector UI accessibility
curl -I http://<NODE_IP>:7860
## Expected: HTTP 200 response

Open these URLs in your browser:

  • http://<NODE_IP>:7862 - CV UI to launch and monitor CV pipeline
  • http://<NODE_IP>:7860 - Alert Inspector UI to view clips and review VLM results

Step 8. Option B

Standard VSS (Hybrid Deployment)

In this hybrid deployment, we would use NIMs from build.nvidia.com. Alternatively, you can configure your own hosted endpoints by following the instructions in the VSS remote deployment guide.

8.1 Get NVIDIA API Key

8.2 Navigate to remote LLM deployment directory

cd deploy/docker/remote_llm_deployment/

8.3 Configure environment variables

Update the environment file with your API keys and deployment preferences. You can do this by editing the .env file directly, or by running the following commands:

## Edit .env file with required keys
echo "NVIDIA_API_KEY=<YOUR_NVIDIA_API_KEY>" >> .env
echo "NGC_API_KEY=<YOUR_NGC_API_KEY>" >> .env
echo "DISABLE_CV_PIPELINE=true" >> .env  # Set to false to enable CV
echo "INSTALL_PROPRIETARY_CODECS=false" >> .env  # Set to true to enable CV

8.4 Update the VSS Image path

Update VIA_IMAGE to nvcr.io/nvidia/blueprint/vss-engine-sbsa:2.4.0 in .env.

## Edit the .env file and update VIA_IMAGE
echo "VIA_IMAGE=nvcr.io/nvidia/blueprint/vss-engine-sbsa:2.4.0" >> .env

8.5 Review model configuration

Verify that the config.yaml file contains the correct remote endpoints. For NIMs, it should be set to https://integrate.api.nvidia.com/v1 .

## Check model server endpoints in config.yaml
cat config.yaml | grep -A 10 "model"

8.6 Launch Standard VSS deployment

## Start Standard VSS with hybrid deployment
docker compose up

8.7 Validate Standard VSS deployment

Access the VSS UI to confirm successful deployment.

## Test VSS UI accessibility (replace <NODE_IP> with your system's IP)
curl -I http://<NODE_IP>:9100
## Expected: HTTP 200 response

Open http://<NODE_IP>:9100 in your browser to access the VSS interface.

Step 9. Test video processing workflow

Run a basic test to verify the video analysis pipeline is functioning based on your deployment.

For Event Reviewer deployment

Follow the steps here to access and use the Event Reviewer workflow.

  • Access CV UI at http://<NODE_IP>:7862 to upload and process videos
  • Monitor results in Alert Inspector UI at http://<NODE_IP>:7860

For Standard VSS deployment

Follow the steps here to navigate VSS UI - File Summarization, Q&A, and Alerts.

  • Access VSS interface at http://<NODE_IP>:9100
  • Upload videos and test summarization features

Step 10. Troubleshooting

Symptom Cause Fix
Container fails to start with "pull access denied" Missing or incorrect nvcr.io credentials Re-run docker login nvcr.io with valid credentials
Network creation fails Existing network with same name Run docker network rm vss-shared-network then recreate
Services fail to communicate Incorrect environment variables Verify IS_SBSA=1 IS_AARCH64=1 are set correctly
Web interfaces not accessible Services still starting or port conflicts Wait 2-3 minutes, check docker ps for container status

Step 11. Cleanup and rollback

To completely remove the VSS deployment and free up system resources:

Warning: This will destroy all processed video data and analysis results.

## For Event Reviewer deployment
cd deploy/docker/event_reviewer/
IS_SBSA=1 IS_AARCH64=1 ALERT_REVIEW_MEDIA_BASE_DIR=/tmp/alert-media-dir docker compose down
cd ../../examples/cv-event-detector/
IS_SBSA=1 IS_AARCH64=1 ALERT_REVIEW_MEDIA_BASE_DIR=/tmp/alert-media-dir docker compose down

## For Standard VSS deployment
cd deploy/docker/remote_llm_deployment/
docker compose down

## Remove shared network (if using Event Reviewer)
docker network rm vss-shared-network

## Clean up temporary media files and stop cache cleaner
rm -rf /tmp/alert-media-dir
sudo pkill -f sys_cache_cleaner.sh

Step 12. Next steps

With VSS deployed, you can now:

Event Reviewer deployment:

  • Upload video files through the CV UI at port 7862
  • Monitor automated event detection and reviewing
  • Review analysis results in the Alert Inspector UI at port 7860
  • Configure custom event detection rules and thresholds

Standard VSS deployment:

  • Access full VSS capabilities at port 9100
  • Test video summarization and Q&A features
  • Configure knowledge graphs and graph databases
  • Integrate with existing video processing workflows