human-pose-estimation-3d-0001.xml: expected to have 2 outputs

[vishnu_tagglabs]

I installed OpenVINO Toolkit. Then I downloaded and converted the human-pose-estimation-3d-0001 model using downloader.py and converter.py.

I built demo apps using Visual Studio 2017.
Then I tried running the human_pose_estimation_demo.exe like this:

human_pose_estimation_demo.exe -i "D:\Videos\pose_vid.mp4" -m "C:\Program Files (x86)\IntelSWTools\openvino_2020.3.355\deployment_tools\open_model_zoo\tools\downloader\public\human-pose-estimation-3d-0001\FP32\human-pose-estimation-3d-0001.xml" -d GPU

 Then I get this error:

InferenceEngine: 00007FFCEF541340
Parsing input parameters
[ ERROR ] C:\Program Files (x86)\IntelSWTools\openvino_2020.3.355\deployment_tools\open_model_zoo\tools\downloader\public\human-pose-estimation-3d-0001\FP16\human-pose-estimation-3d-0001.xml: expected to have 2 outputs

I converted the models using mo.py manually.

python "C:\Program Files (x86)\IntelSWTools\openvino_2020.3.355\deployment_tools\model_optimizer\mo.py"  --input_model "C:\Program Files (x86)\IntelSWTools\openvino_2020.3.355\deployment_tools\open_model_zoo\tools\downloader\public\human-pose-estimation-3d-0001\human-pose-estimation-3d-0001.onnx"

The error persists.

Please help. Thank you.

[vishnu_tagglabs · Accepted solution]

My bad. I was using the wrong model with wrong demo app. The c++ app was for 2 pose estimation while 3d pose estimation was in python demos.

[Wan_Intel]

Hi Vishnu_Tagglabs,

Thank you for reaching out to us.

Human Pose Estimation C++ Demo showcases the work of a multi-person 2D pose estimation algorithm. You can use the following pre-trained model with the demo:

·      “human-pose-estimation-0001”

In the latest OpenVINO™ Toolkit, 4 pre-trained models can be used with this demo. You can check the list of models here:

https://github.com/openvinotoolkit/open_model_zoo/blob/2021.3/demos/human_pose_estimation_demo/cpp/models.lst#L1

Regards,

Wan

[Wan_Intel]

Hi Vishnu_Tagglabs,

This thread will no longer be monitored since this issue has been resolved. If you need any additional information from Intel, please submit a new question.

Regards,

Wan