Linking error while building tensorflow

Autonomous Machines Jetson & Embedded Systems Jetson Orin NX
1

I’m encountering a linking error while building TensorFlow 2.16.1 using Bazel 6.5.0 on the Orin NX(Jetpack 6) platform. but the build fails at the linking stage. I have attached a screenshot for reference.
Any suggestions on how to resolve it?

Image
Image1920×1440 315 KB

3

Hi,

We have TensorFlow v2.16 prebuilt in the below link.
You don’t need to build it from the source.

https://developer.download.nvidia.cn/compute/redist/jp/v60/tensorflow/

Thanks.

4

Hi Team,

Thanks for the clarification.
Can we use the mentioned whl file with our source application code and can we build and create a binary out of it using the bazel.

please reply.

Thanks,
Swasthika

5

Hi Team,

We could see the attached whl files and look like they are compiled for python environment.

But the use case for is like, we have an application (C++ code) that is wrote on bazel environment. Can we use the whl files to extract the Tensorflow and add our source code that has to be built in the bazel environment and generate the binaries? please clarify.

Thanks.

6

Hi Team,

Gentle reminder!!!

Is there any update on this topic. Please reply.

Cheers

7

Hi,

Suppose you need the C++ header to build your custom app.
This is not included in the package shared above.

To build TF from the source, please find the below topic for a community sharing:

Thanks

9

This topic was automatically closed 14 days after the last reply. New replies are no longer allowed.