search

EdjeElectronics / TensorFlow-Lite-Object-Detection-on-Android-and-Raspberry-Pi

A tutorial showing how to train, convert, and run TensorFlow Lite object detection models on Android devices, the Raspberry Pi, and more!
Apache License 2.0
1.5k stars 684 forks source link

using object detection with TPU has some error(Segmentation fault). #180

Closed sangyo1 closed 6 months ago

sangyo1 commented 6 months ago

I was watching the youtube video and run the detection model on my rpasberry pi with TPU. https://www.youtube.com/watch?v=qJMwNHQNOVU&ab_channel=EdjeElectronics

I was able to run without TPU, it ran with 8FPS with my webcam

python3 TFLite_detection_webcam.py --modeldir=Sample_TFLite_model

However, when I ran with TPU

python3 TFLite_detection_webcam.py --modeldir=Sample_TFLite_model --edgetpu

It kept gave me "Segmentation fault" message. What is wrong and how do I fix it?

SamuelKollar commented 6 months ago

Hey, I get the same error did you find any fix for it?

SamuelKollar commented 6 months ago

Resolved this issue, I just installed the Edge TPU runtime library from https://github.com/feranick/libedgetpu/releases/tag/v16.0TF2.13.1-2 for the correct version of my RaspberryPi OS and it resolved the issue.

sangyo1 commented 6 months ago

Resolved this issue, I just installed the Edge TPU runtime library from https://github.com/feranick/libedgetpu/releases/tag/v16.0TF2.13.1-2 for the correct version of my RaspberryPi OS and it resolved the issue.

Could you tell me more specific? libedgetpu-dev_16.0tf2.13.1-2.bookworm_arm64.deb I download above file in my rapsberry pi, and sudo dpkg -i the package. I ran python3 TFLite_detection_webcam.py --modeldir=Sample_TFLite_model --edgetpu I still get the same error message. Do you have any link that how to solve this problem?

SamuelKollar commented 6 months ago

I specifically installed libedgetpu1-std_16.0tf2.13.1-2.ubuntu22.04_arm64.deb form the link above, as it is the standard runtime and it is specific for the OS that is running on my pi4, I also installed python 3.11 specific wheel from https://pypi.org/project/tensorflow/2.12.0/#files.

After that I just ran the command for TFLite_detection_webcam.py and it ran with no issues.