[Detector Support]: Model usage issues #20476
Replies: 1 comment
-
|
Based on the error message in your logs, the issue is with the model configuration for your custom SSDLite MobileNet v2 model(1). The OpenVINO detector expects specific model output formats, and your model doesn't match the expected SSD format(1). The error "SSD models must only have 1 output. Found 4" and "SSD model output doesn't match. Found [1,100,4]" indicates your model has a different output structure than expected(1). For the OpenVINO detector with the default SSDLite MobileNet v2 model, you should use the pre-configured model that comes with Frigate(1). Here's the correct configuration: detectors:
ov:
type: openvino
device: CPU
model:
width: 300
height: 300
input_tensor: nhwc
input_pixel_format: bgr
path: /openvino-model/ssdlite_mobilenet_v2.xml
labelmap_path: /openvino-model/coco_91cl_bkgr.txtThe key differences from your configuration:
The default OpenVINO model is provided in the container and comes from Intel's Open Model Zoo, converted to an FP16 precision IR model(1). If you want to use your own custom model, you'll need to ensure it matches the expected SSD output format or consider using a different model type like YOLO-NAS, YOLOX, or other supported models(1). Remove the custom model volume mount from your Docker Compose file and use the default paths, then restart Frigate. 📚 Sources: Have another question? Just tag @inkeep. |
Beta Was this translation helpful? Give feedback.
Uh oh!
There was an error while loading. Please reload this page.
-
Describe the problem you are having
I deployed frigate based on openvino and ssdlite_mobilenet_v2.xml, but the model could not run normally.
Version
0.16-0
Frigate config file
docker-compose file or Docker CLI command
Relevant Frigate log output
Install method
Docker Compose
Object Detector
OpenVino
Screenshots of the Frigate UI's System metrics pages
Any other information that may be helpful
No response
Beta Was this translation helpful? Give feedback.
All reactions