Yolo V8 Segmentation fault core dumped issue on Orin

[lenka.venkatagiribabu]

Hi all,

I tried to create the basic Yolo V8 model on Jetson AGX Orin (both 32/64 GB variants) , while running below code on Jetson AGX Orin I am getting Segmentation Fault(core dumped)…

Installed all necessary packages on Jetson AGX Orin as per requirements mentioned & same file/setup I am able to run on my Host (windows) pc…

getting issues on Jetson AGX Orin board only,

please let me know if I miss any thing here,

Thanks in Advance…

Regards,
L V Giribabu.

[AastaLLL]

Hi,

Could you share the detailed steps for setting and reproducing the error?
So we can check it further?

Thanks.

[lenka.venkatagiribabu]

Hi,

1. Flashed OS image - R35.4.1 (Jetpack 5.1.2) on Jetson AGX Orin (both 32/64 GB Variant’s)
2. deep stream version - 6.3
3. Downloaded Yolo V8 source code from git repo → GitHub - ultralytics/ultralytics: NEW - YOLOv8 🚀 in PyTorch > ONNX > OpenVINO > CoreML > TFLite

& followed steps as mentioned in that link & installed all necessary packages on Jetson AGX Orin as per

requirements mentioned & same file/setup I am able to run on my Host (windows) pc…

i) pip installed
ii)

iii) Training the model on board
shared log file as below here for your reference…

same setup / file able to run / train the model on host(windows) pc,

plz let me know if u need any info for the same…

Thanks,
L V Giribabu

[AastaLLL]

Hi,

Thanks for the details.
Will get back to you later.

[lenka.venkatagiribabu]

Hi,

Ok, Sure.

Thanks & best regards,
L V Giribabu.

[lenka.venkatagiribabu]

Hi,

Can you please give any update for the same?

Thanks in Advance.

Thanks,
L V Giribabu.

[lenka.venkatagiribabu]

Hi,

Can u please share any inputs for the same?

Thanks,
L V Giribabu.

[AastaLLL]

Hi,

Thanks for your patience.

$ pip3 install ultralytics

Above command will download a CPU PyTorch package and cause the segmentation fault error.
Please reinstall PyTorch with our GPU prebuilt to fix this issue…

$ sudo apt-get -y install autoconf bc build-essential g++-8 gcc-8 clang-8 lld-8 gettext-base gfortran-8 iputils-ping libbz2-dev libc++-dev libcgal-dev libffi-dev libfreetype6-dev libhdf5-dev libjpeg-dev liblzma-dev libncurses5-dev libncursesw5-dev libpng-dev libreadline-dev libssl-dev libsqlite3-dev libxml2-dev libxslt-dev locales moreutils openssl python-openssl rsync scons python3-pip libopenblas-dev;
$ python3 -m pip install --upgrade pip; python3 -m pip install aiohttp numpy=='1.19.4' scipy=='1.5.3' export "LD_LIBRARY_PATH=/usr/lib/llvm-8/lib:$LD_LIBRARY_PATH"; python3 -m pip install --upgrade protobuf; python3 -m pip install --no-cache https://developer.download.nvidia.cn/compute/redist/jp/v511/pytorch/torch-2.0.0+nv23.05-cp38-cp38-linux_aarch64.whl

Then try the YOLOv8 command again.

Thanks.

[lenka.venkatagiribabu]

Hi,

Sure, will check it & let u know…

Thanks for ur quick reply…

Regards,
L V giribabu

[lenka.venkatagiribabu]

Hi,

Yes, now I am able to run & create the customized model,
Thank you for ur kind support.

Regards,
L V Giribabu.

[AastaLLL]

Good to know this.
Thanks for the update.