"Routes DeepStream-Yolo tasks that run multiple detectors in one
Scanned 9/8/2026
Install to Claude Code
npx -y skills add VectorSpaceLab/AREX-Skill --skill multi-gie --agent claude-codeInstalls into .claude/skills of the current project.
Are you the author of Multi Gie?
Add the live security badge to your README — it updates automatically with every re-scan.
[](https://www.skillsdirectory.com/skills/vectorspacelab-multi-gie)More formats (shields.io, HTML) on the badges page.
---
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.
Is this your skill, or is something wrong with this listing? Request removal or report an issue. Author removals are honored within 72 hours.
No comments yet. Be the first to comment!