#Frigate+ and openvino

1 messages · Page 1 of 1 (latest)

rough grove
#

I'm considering purchasing Frigate+ but i had some doubts I wanted to clear first.

I'm currently running the included SSDLite MobileNet v2 openvino model, with an inference speed of 8.51ms according to frigate metrics. Should I expect to have similar performance with the yolonas Frigate+ openvino models? Something else I noticed is that to use Frigate+ models the configuration looks like this:

model:
  path: plus://<your_model_id>

I was wondering how this works, does this download the model and store it locally?

I know this is not exactly related to home assistant but thought this would be a good place to ask.

#

CPU is i7 8700 and I'm using the iGPU

finite sentinel
#

yolonas is a more complex model, but its more accurate

#

inference times will probably be closer to 20ms

#

the models are downloaded into the model_cache directory