Deploying Image Classification model on Raspberry pi 4

Help
#1

Question/Issue: I am trying to deploy the image classification model on Raspberry pi 4 model B using the following command:
$ edge-impulse-linux-runner
But it is showing the error
Failed to run impulse Error: Unsupported architecture “aarch64”, only armv7l or aarch64 supported
for now.
I have installed a 32 bit Raspbian OS with aarch64 architecture.
Kindly guide me through this.
Thanks in advance!

Project ID: 152977

Context/Use case: Blur vs Clear Image Classification

#2

Hi @Dhruhi

Welcome to the forum, It seems like you are trying to run 64 bit on a 32 bit install? You wont be able to do that.

There was a similar post to this a few months ago if you want to check here. You will need to stick with the same architecture:

Best

Eoin

#3

Hey,
Thank you for the suggestion, I have now installed the 64 bit OS on the raspberry pi, I am able to connect the edge impulse account to raspberry pi but when now I run the command,

 $ edge-impulse-linux-runner

It successfully downloads the model, but when the image classification starts, it throws another error saying, “Failed to run impulse Error: Cannot find any webcams”
While connecting edge impulse with the raspberry pi it was not detecting the camera connected to the Raspberry Pi (I am using Raspberry Pi Camera V2.).
Kindly help me through!
Thanks in advance!

#4

Hi,

You may need to go into raspberry pi config and enable the web camera.

You can find this in the terminal with:

sudo raspi-config

Best

Eoin

#5

On a Raspberry Pi 4, even if you install the 32-bit Raspbian image, the 64 bit kernel will load by default.

You can disable this behavior by adding the following line to your /boot/config.txt:

arm_64bit=0

If you run ‘uname -a’ after making this change, you should see the 32 bit kernel in use.

Linux raspberrypi 6.1.21-v7l+ #1642 SMP Mon Apr 3 17:22:30 BST 2023 armv7l GNU/Linux

1 Like