Unable to run Tensorflow image classification example

OpenMV Products
1

Hi,

I am trying to run the mobilenet example Examples > Machine Learning > Tensorflow > tf_image_classification.py on a H7 Plus

No matter which of the provided tflite files I try and use I get an error:

Traceback (most recent call last):
File “”, line 57, in
OSError: tensorflow/lite/micro/kernels/softmax_common.cc:78 input->type != kTfLiteFloat32 (UINT8 != FLOAT32)
Node SOFTMAX (number 65f) failed to prepare with status 1
AllocateTensors() failed!

How can I make this work?

Thanks

2

Hi, we basically have dropped support for that model as it’s not really useful for anything. We should remove the example.

3

Cheers fella… just wanted something to play with when I first got the camera, setup with Edge Impulse now…

Andy

4

There’s a built-in FOMO face detector, that should be working.

5

Hmmm…I’m (still) getting the same error, nearly a year later (with most recent IDE).

6

Yeah, we aren’t supporting mobilenet. We’re pretty much tied to whatever edge impulse is supporting.