skills.rest
olibartfast
olibartfastCOMMUNITY
Explore all skills by this author →
WebsiteGitHub

deepstream-sop

Automate SOP compliance monitoring in industrial video feeds with DeepStream pipelines.

Updated Nov 21, 2025
Search Tags:#gpu#sop#triton#vlm#video-analytics#deepstream
One-click install
npx skills add https://github.com/olibartfast/deep-infer --skill deepstream-sop-olibartfast
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: deepstream-sop
Source: https://github.com/olibartfast/deep-infer/tree/main/skills/deepstream-sop
Command: npx skills add https://github.com/olibartfast/deep-infer --skill deepstream-sop-olibartfast

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill requires qwen-vl-utils, ipdb, playsound, confluent-kafka, onnx, onnxscript, onnxruntime, and includes references (resource) components.

What problem does it solve?

This skill automates the complex task of monitoring industrial assembly lines for Standard Operating Procedure (SOP) compliance, reducing the need for manual video review and ensuring consistent process adherence.

Core Features & Use Cases

  • Temporal Action Detection: Uses GEBD and DDM models to segment video into meaningful chunks based on event boundaries.
  • VLM-based Classification: Leverages embedded vLLM (Cosmos Reason) to classify actions and detect missing or out-of-order steps.
  • Use Case: A factory manager can deploy this service to monitor a work-cell camera, automatically flagging when an operator skips a critical safety step or performs assembly tasks in the wrong sequence.

Quick Start

Use the deepstream-sop skill to build and deploy the SOP inference microservice for monitoring the assembly line camera feed.

Frequently Asked Questions about deepstream-sop

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I automate SOP compliance monitoring for industrial video feeds?▼

Automate SOP compliance monitoring by integrating DeepStream pipelines with VLM-based action classification to detect missing or out-of-order assembly steps in real-time video feeds.

Do I need Docker and the NVIDIA DeepStream SDK to run real-time temporal action detection?▼

Yes, executing the multi-stage SOP processing pipeline requires Docker, the NVIDIA DeepStream SDK, and the Triton Inference Server to deploy GPU-accelerated microservices for temporal action detection.

How does VLM-based action classification detect out-of-order steps in an assembly line?▼

VLM-based action classification uses GEBD and DDM models to segment video into event boundaries, then leverages an embedded vLLM to classify actions and identify missing or out-of-sequence steps.

Can I use Triton Inference Server with DeepStream for sequence validation?▼

Yes, Triton Inference Server works with DeepStream pipelines to deploy GPU-accelerated microservices that perform real-time sequence validation and flag operators skipping critical safety steps.

How do I report SOP compliance violations via Kafka events?▼

Report SOP compliance violations by configuring the DeepStream pipeline to emit Kafka-based events when the VLM-based action classification detects an operator skipping a critical step or performing tasks out of sequence.

Related Skills

Explore More Software Engineering Skills→
NVIDIANVIDIA
189

deepstream-run-mv3dt

Operate the DeepStream Multi-View 3D Tracking reference application with pipeline configuration and execution workflows.

Official
Advanced
GaneshMadarasuGaneshMadarasu

drift-detector

Detect data and concept drift in Kafka anomaly scores using ADWIN and DDM detectors.

Community
Advanced
doanchienthangdevdoanchienthangdev
4

Model Serving

Deploy machine learning models with TensorFlow Serving, TorchServe, or Triton Inference Server.

Community
Advanced
NVIDIANVIDIA
49

deepstream-sop

Build and operate GPU-accelerated SOP video inference services with DeepStream and Triton.

Official
Advanced