[BUG] [DLStreamer] - Documentation instructions for DLStreamer deployment in WSL environment

[jangorecki1]

Bug Description

1. Missing information in Installation Documentation
   Problem: Lack of comprehensive installation guide for users starting from scratch on Intel processor laptops (e.g. Intel(R) Core(TM) Ultra 7 265U).
   Expected: Clear step-by-step installation documentation

2. WSL2 /dev/dri Access Issue
   Problem: Missing /dev/dri device access in WSL2 environment.
   Reference: This appears to be related to microsoft/WSL#5822
   Suggested Fix: Add instructions for enabling GPU access in WSL2 to the installation documentation.

3. Repository Access Error: orbidden IP: Err:7 https://repositories.intel.com/graphics/ubuntu noble InRelease 403 Forbidden [IP: 18.66.233.105 443]

4. Model Dependencies Issues
   Problem: After DLStreamer installation appears successful, model-related dependencies fail:
   4a. YOLO11 Model Not Found
   Issue: YOLO11 model cannot be located
   Suspected cause: Directory/path configuration problems

4b. gvainference Plugin Not Detected
Issue: gvainference plugin is not recognized
Suspected cause: GStreamer path dependencies not properly configured

System Setup

HW/SW environment:
Processor: Intel(R) Core(TM) Ultra 7 265U (2.10 GHz)

Name DriverVersion

---

Intel(R) Graphics 32.0.101.8136

WSL version: 2.5.10.0
Kernel version: 6.6.87.2-1
WSLg version: 1.0.66
MSRDC version: 1.2.6074
Direct3D version: 1.611.1-81528511
DXCore version: 10.0.26100.1-240331-1435.ge-release
Windows version: 10.0.26100.6584

Reproducible Steps

scratch installation

Root Cause Analysis

Documentation, Verification

[jgespino]

@jangorecki1 Thank you for providing the feedback, I went through the setup guide for WSL and I also experienced the same issues you mentioned.

Model Dependencies Issues
Problem: After DLStreamer installation appears successful, model-related dependencies fail:
4a. YOLO11 Model Not Found
Issue: YOLO11 model cannot be located
Suspected cause: Directory/path configuration problems

For this to work, you need to manually download the model as mentioned in the other documentation. This should be included in this version of the documentation as well.

mkdir -p /home/${USER}/models
export MODELS_PATH=/home/${USER}/models
/opt/intel/dlstreamer/samples/download_public_models.sh yolo11s coco128

4b. gvainference Plugin Not Detected
Issue: gvainference plugin is not recognized
Suspected cause: GStreamer path dependencies not properly configured

I suspect the GPU was not properly setup, are you able to run the hello script with CPU?

/opt/intel/dlstreamer/scripts/hello_dlstreamer.sh --device=CPU

If the pipeline does not run, try running the pre-requisites script mentioned here.

[jgespino]

@kblaszczak-intel FYI

[dmichalo]

Hello,
documentation update in progress: #1069

[dmichalo]

Documentation fix has been merged to main branch.

[jangorecki1]

After installing new wsl, new ubuntu 24.04 and new drivers, step 1 on Linux WSL:
~$ ls -ltrah /dev/dri
ls: cannot access '/dev/dri': No such file or directory

Step2 is unclear what to do next, as everything was done : Step 2: If you don’t have renderD12* device, please update GPU drivers and update WSL

For step2 consider this fix: microsoft/WSL#5822

Step6: /opt/intel/dlstreamer/scripts/hello_dlstreamer.sh => following error
$ /opt/intel/dlstreamer/scripts/hello_dlstreamer.sh

Intel(R) DL Streamer 2025.1.2 has been installed successfully. You are ready to use it.

Model yolo11s which you want to use cannot be found at !
Please run the script /opt/intel/dlstreamer/samples/download_public_models.sh yolo11s coco128 to download the model.
If the model has already been downloaded, specify the path to its location (for instance: export MODELS_PATH=/home/user/you_path).

Suggestions:
-add instructions how to add Yolo11
-python packages may be missing
-ensure that below command runs on GPU and NPU
/opt/intel/dlstreamer/scripts/hello_dlstreamer.sh --device=CPU