Skip to content
Back to skills

Multi Gie

ASecurity

"Routes DeepStream-Yolo tasks that run multiple detectors in one

  • 247 stars
  • 0 votes
  • 0 copies
  • 3 views
  • Added September 8, 2026
devops

Security analysis

A100/100

Pro scans all 5 files and shows the line behind each finding

Scanned September 8, 2026

npx -y skills add VectorSpaceLab/AREX-Skill --skill multi-gie --agent claude-code

Installs into .claude/skills of the current project.

Are you the author of Multi Gie?

Add the live security badge to your README. It updates with every re-scan.

Security grade badge for Multi Gie
[![Security: A — Skills Directory](https://www.skillsdirectory.com/api/skills/vectorspacelab-multi-gie/badge)](https://www.skillsdirectory.com/skills/vectorspacelab-multi-gie)

More formats (shields.io, HTML) on the badges page. Keep it an A: scan every change in CI with Pro.

Download with Pro
SKILL.md
---
name: multi-gie
description: "Routes DeepStream-Yolo tasks that run multiple detectors in one
  DeepStream pipeline or use secondary GIEs."
disable-model-invocation: true
metadata:
  disco-role: operating
license: MIT
---

# Multi-GIE

Use this sub-skill when the user wants more than one YOLO detector in one DeepStream app, especially when the task involves duplicated `gieN` folders, secondary inference, or plugin-version collisions.

## Trigger phrases

- multiple GIEs
- two YOLO models in one app
- secondary GIE
- `operate-on-gie-id`
- `operate-on-class-ids`
- `YOLOLAYER_PLUGIN_VERSION`
- duplicate detector pipeline

## Include here

- Folder duplication for `gie1`, `gie2`, ...
- Changing the plugin version inside each copied `yoloPlugins.h`.
- Wiring `primary-gie` and `secondary-gieN` sections.
- Moving the generated engine file into the right `gieN/` folder.
- Troubleshooting collisions between the duplicated inference stacks.

## Exclude or route elsewhere

- Exporting the model to ONNX first: use `model-conversion`.
- Single-detector deployment: use `deployment`.
- INT8 calibration tuning: use `int8-benchmarking`.
- Skill maintenance or router import logic.

## How to use this route

1. Read `references/workflows.md` for the multi-GIE folder layout.
2. Read `references/configuration.md` before editing any duplicated config files.
3. Use `scripts/setup-multi-gie-tree.sh --count 2 --output-dir ./deepstream-yolo-multi-gie` to scaffold a self-contained multi-GIE runtime tree from the bundled assets.
4. Check `references/troubleshooting.md` if the app starts but the GIEs do not line up.

## What a future agent should be able to do here

- Explain why `deepstream-app` cannot use multiple primary GIEs without a custom code path.
- Scaffold the `gie1`, `gie2`, ... folder layout.
- Set unique plugin versions and GIE IDs consistently.
- Route secondary inference to a primary detector and/or class subset.

## Common failure signals

- `gie-unique-id` collisions
- `operate-on-gie-id` points at the wrong primary GIE
- `YOLOLAYER_PLUGIN_VERSION` was not incremented
- Engine files are left in the root folder instead of the owning `gieN/` folder

## Linked helpers

- `scripts/setup-multi-gie-tree.sh` — scaffold the duplicated folder structure from the bundled assets.
- `references/workflows.md` — step-by-step multi-GIE flow.
- `references/configuration.md` — config keys and meaning.
- `references/troubleshooting.md` — multi-GIE failure modes.

Files in this skill

  • SKILL.md2.4 KB
  • references/configuration.md1.8 KB
  • references/troubleshooting.md1.2 KB
  • references/workflows.md1.9 KB
  • scripts/setup-multi-gie-tree.sh6.3 KB

Attribution

Is this your skill, or is something wrong with this listing? Request removal or report an issue. Author removals are honored within 72 hours.

Comments

Loading comments…

Newsletter

Get the weekly list of safe new skills, and the ones we pulled.