Example TF_objec_detection_1.py problems

OpenMV Boards
1

Hi,

I’m trying to run the “tf_object_detection_1.py” example on OpenMV H7 plus firmware version 4.7.0 but i got this error :

immagine
immagine228×112 4.84 KB

I’ve saved the tflite model and lables on the SD

immagine

This is the serial monitor output:

immagine
immagine521×164 5.09 KB

Any suggestion ?

2

The default script in examples is tryin to read the model off of ROMFS. So, you need to update/reset the ROMFS on the OpenMV Cam H7 Plus.

Let me know if you want better face detection, we can now run BlazeFace at 2 FPS onboard the H7 plus which will be released soon, but, you can get pre-release firmware for it if you want.

3

Hi , yes please let me test the pre release version. Tks.

4

Hi, in the IDE do:

Tools→Install Latest Dev Release → Select the Update ROMFS checkbox and erase internal flash checkbox.

Then, you can run examples from here: openmv/scripts/examples/03-Machine-Learning/00-TensorFlow at master · openmv/openmv · GitHub

For the H7 Plus, just blaze face, tf_image_classificaion/object_detection, and YOLO_LC work. You can load the models for the other networks onto your camera, but, they will run at 0.1 FPS ish… so, very slow.

5

Kwagyeman ,

Actually I’ve the firmware 4.7.0.

Do I need install the the last development release date september 9th to have the pre release firmware?

6

Yes, v4.7.0 is from earlier this year.