Issue with Tensorflow

octoberxviirobot · June 7, 2021, 10:21pm

Hi all,

I’m having an issue with TensorFlow on the Jetson AGX Javier. I followed the following link procedure on downloading TensorFlow and even updated gast and kera_processing to the correct versions since the ones on the list are outdated: Installing TensorFlow for Jetson Platform :: NVIDIA Deep Learning Frameworks Documentation

This is the error I am getting when running my program:

Would someone be able to provide some clarification on the resolution to this Warning?

AastaLLL · June 8, 2021, 4:58am

Hi,

It seems that you meet a similar issue as below.
Based on the comment, would you mind to shift the import order of OpenCV and TensorFlow to see it it works?
https://github.com/opencv/opencv/issues/14884#issuecomment-599852128

Thanks.

octoberxviirobot · June 8, 2021, 8:02am

Hi AastaLLL,

I really appreciate your comment thank you. I did not have OpenCV imported, but by further researching the answer you provided I was able to find a solution on the NVIDIA forums and StackOverflow. Here are the links for anyone curious about it in the future:

python 3.x - ImportError: /usr/lib/aarch64-linux-gnu/libgomp.so.1: cannot allocate memory in static TLS block - Stack Overflow
Error: ImportError: /usr/lib/aarch64-linux-gnu/libgomp.so.1: cannot allocate memory in static TLS block (i looked through available threads already)

The answer was just to go to the terminal and type:
export LD_PRELOAD=/usr/lib/aarch64-linux-gnu/libgomp.so.1

That removed the error.