Using tensorflow lite model generated on edge impulse

#1

I want to download the tensorflow lite model output for the data trained using edge impulse and use it as a normal tensorflow lite model but i have been facing segmentation errors whenever I try to use such model however if i download any model from the internet, my code works

#2

Hi @Vic_fretz,

Can you share your results, your example code? Can you also specify which learn block you’re using?

EDIT: Can you share your impulse setup and project id?

#3

@rjames sorry for the late response i will look at my code and forward it to you but as i have said in my ealrier question, the model i download from edge impulse resukts in segmentation error

#4

@vic_fretz,

Yes it’ll be helpful to provide more information or share snippets of code.
Are you using the download TFlite file from Dashboard? What kind of model is it?
How are you testing? Python? C++?

#5

I am using the download tflite file from the dshaboard.

It is an image classification model and the code is in python

#6

Hi @Vic_fretz,

Have you resolved this issue?
If not, can you share more information on how you’re creating the input data to feed to the NN?
Or a snippet of the code? We run the TFlite model in Studio when we profile the model so I expect this issue to be user error. But the more information you provide the better we can help.

// Raul