Roboflowoak with empty prediction

[kamk]

I have trained a model using Roboflow work space. The model works fine with web cam web interface and with depth_sdk example:
from depthai_sdk import OakCamera
import depthai

Download & deploy a model from Roboflow Universe

https://universe.roboflow.com

with OakCamera() as oak:
color = oak.create_camera(‘color’)
model_config = {
‘source’: ‘roboflow’, # Specify that we are downloading the model from Roboflow
‘model’:‘person-snpao/1’,
‘key’:‘xx’ # FAKE Private API key, replace with your own!
}
nn = oak.create_nn(model_config, color)
oak.visualize(nn, fps=True)
oak.start(blocking=True)
When I use the code from this link Luxonis OAK | Roboflow Docs, I get empty prediction printed in the console.
Some details:
- Object Detection
- MacOS
- project ID:person-snpao
I’m using roboflowoak version 0.0.12
Any ideas?

[Lake]

Hi there! Can you try setting a very low confidence, like shown here in docs? Luxonis OAK | Roboflow Docs

This will help us understand if there is something wrong in the pipeline, or if the model simply isn’t returning high confidence predictions.

[kamk]

Hi, I still get empty predictions when confidence is set to 0.05 and I also tried 0.0001.

[kamk]

Hi, I also tried a few models from Roboflow Universe. Some models are able to give predictions while others results in empty prediction list.

[Lake]

If this method of deployment is not working, I’d highly recommend using one of the officially supported devices for running Roboflow Inference via our inference server: Local Installation - Roboflow Inference.

Could you try deploying your model with a method here, and seeing if you are able to get the expected results?

[Lake]

For further help with your Oak, I suggest posting a topic in the Luxonis Forum! https://discuss.luxonis.com/