Object detection size reduction in edge impulse

Hello, I am using OpenMV M7 to detect 10 classes of objects. I m using edge impulse to train the model. I m able to train image classification model with a size in kBs. But object detection model is in MBs. Is there any way to reduce the object detection model size?

Not really. Also, we don’t support object detection yet. But, we will soon on the STM32H7.

Quick question of you don’t mind! Any ideas on when that will be available? I would LOVE to have an x,y of where something is, it would be extremely useful!!!

I’m finishing up the release of a new version of OpenMV IDE and then I will start work on objection detection.

1 Like