Skip to main content
Question

Trouble running the yolo models on metis compute board

  • November 5, 2025
  • 10 replies
  • 117 views

Forum|alt.badge.img

Hey again !

Finally I was able to install the voyager-SDK on my metis compute board ! Turns out I was debugging too much with chat gpt and it really does nonsense things I should have read more thoroughly the user guide for the metis or  the github repository… 

But now I have the SDK and I can test the different models directly in the SDK :)

I successfully ran the yolov5s-v7-coco.yaml via ./inference <modelname> <media> 

But I can only run this model…

I wonder if it’s normal and the other models in the folder ax_models/ are here but not compiled and need a certain action by the user (me), or if it is a bug and I’m doing something wrong ? 

When I try to run for example the yolov7 model with the argument --no-display just to see if it works : 

./inference.py yolov7-coco media/traffic1_1080p.mp4 --no-display 

I have this error : 

2025-11-05 15:46:12.084642319 [W:onnxruntime:Default, device_discovery.cc:164 DiscoverDevicesForPlatform] GPU device discovery failed: device_discovery.cc:89 ReadFileContents Failed to open file: "/sys/class/drm/card1/device/vendor"
arm_release_ver: g13p0-01eac0, rk_so_ver: 11
INFO : yolov7-coco is being compiled for single-core cores (but can be run using 4 cores).
INFO : Deploying model yolov7-coco for 1 core. This may take a while...

INFO : ## Compiling network yolov7-coco **/home/ubuntu/voyager-sdk/ax_models/zoo/yolo/object_detection/yolov7-coco.yaml** yolov7-coco
INFO : Compile model: yolov7-coco
ERROR : HTTPSConnectionPool(host='media.axelera.ai', port=443): Max retries exceeded with url: /artifacts/model_cards/weights/yolo/object_detection/yolov7.pt (Caused by NameResolutionError("<urllib3.connection.HTTPSConnection object at 0x7f4e3f60e0>: Failed to resolve 'media.axelera.ai' ([Errno -3] Temporary failure in name resolution)"))
INFO : Compiling yolov7-coco: yolov7-coco took 8.807 seconds
ERROR : Failed to deploy network

2025-11-05 15:46:17.612793084 [W:onnxruntime:Default, device_discovery.cc:164 DiscoverDevicesForPlatform] GPU device discovery failed: device_discovery.cc:89 ReadFileContents Failed to open file: "/sys/class/drm/card1/device/vendor"

ERROR : Failed to deploy model yolov7-coco
ERROR : Command '/home/ubuntu/voyager-sdk/deploy.py --model yolov7-coco --num-cal-images 200 --aipu-cores 1 --data-root /home/ubuntu/.cache/axelera/data --pipe gst --build-root /home/ubuntu/voyager-sdk/build /home/ubuntu/voyager-sdk/ax_models/zoo/yolo/object_detection/yolov7-coco.yaml --aipu-cores 1 --metis pcie ' returned non-zero exit status 1.

Is this error normal because the models need to be compiled or configure with a certain manipulation (that I am not aware of) ?

Thank you !

Thibaut

10 replies

Spanner
Axelera Team
Forum|alt.badge.img+2
  • Axelera Team
  • November 5, 2025

Yo ​@tchretien! Nice work getting the Metis Compute Board up and running again 👍

At first glance, that error kinda looks like your Board doesn’t have internet access? Or a netword issue, anyway. Like it can’t reach media.axelera.ai for the download. I wonder if yolov5s-v7-coco was already compiled, which is why it worked?

If we can check network connectivity first, then hopefully the ./download_prebuilt.py <model_name> download script will help you get the precompiled models nice and easy.

What does running ping media.axelera.ai from the board show you?


Forum|alt.badge.img
  • Author
  • Ensign
  • November 6, 2025

Yes I think that is it…

ubuntu@antelao-3588:~/voyager-sdk$ ping media.axelera.ai
bash: ping: command not found
ubuntu@antelao-3588:~/voyager-sdk$ ./download_prebuilt.py yolov7-coco
bash: ./download_prebuilt.py: No such file or directory
 

How can I resolve this ?

 


Forum|alt.badge.img
  • Author
  • Ensign
  • November 6, 2025

I installed the librairie linked to ping and now my metis is connected to internet but still the same error for ./download_prebuilt.py : no such file or directory…

 


Spanner
Axelera Team
Forum|alt.badge.img+2
  • Axelera Team
  • November 6, 2025

Ah sorry, ​@SamP just reminded me that ./download_prebuilt.py <model_name> has been deprecated in favour of the command axdownloadmodel <model_name> for the prebuilt models from the zoo. My fault there, apologies!

Does that work instead?

 


Forum|alt.badge.img
  • Author
  • Ensign
  • November 6, 2025

It works ! thanks !


Spanner
Axelera Team
Forum|alt.badge.img+2
  • Axelera Team
  • November 6, 2025

Ah fantastic! Keep me posted on how it’s going 👍


Forum|alt.badge.img
  • Author
  • Ensign
  • November 6, 2025

New error xD 

I can’t load onnx models… 

I get this error is this normal ? 

(venv) ubuntu@antelao-3588:~/voyager-sdk$ axdownloadmodel yolo11l-coco-onnx
Error: Failed to reach https://media.axelera.ai/artifacts/prebuilt/voyager-sdk-v1.4.2/models.json: [Errno -3] Temporary failure in name resolution
(venv) ubuntu@antelao-3588:~/voyager-sdk$ axdownloadmodel yolo11m-coco-onnx
Error: Failed to reach https://media.axelera.ai/artifacts/prebuilt/voyager-sdk-v1.4.2/models.json: [Errno -3] Temporary failure in name resolution
(venv) ubuntu@antelao-3588:~/voyager-sdk$ axdownloadmodel yolo10n-coco-onnx
Error: Failed to reach https://media.axelera.ai/artifacts/prebuilt/voyager-sdk-v1.4.2/models.json: [Errno -3] Temporary failure in name resolution
(venv) ubuntu@antelao-3588:~/voyager-sdk$ axdownloadmodel yolo10s-coco-onnx
Error: Failed to reach https://media.axelera.ai/artifacts/prebuilt/voyager-sdk-v1.4.2/models.json: [Errno -3] Temporary failure in name resolution

 


Spanner
Axelera Team
Forum|alt.badge.img+2
  • Axelera Team
  • November 6, 2025

Hmm, this still kinda looks like a network issue, or potentially a firewall or similar? Do you know if there are any restrictions on your connection?


Forum|alt.badge.img
  • Author
  • Ensign
  • November 6, 2025

My bad ! 

I fixed it, I simply change my ethernet cable😂

 


Spanner
Axelera Team
Forum|alt.badge.img+2
  • Axelera Team
  • November 6, 2025

Ah, that’s great news! No worries at all, as long as we got there, that’s all that matters! Keep us posted!