JetPack 6.2 (Orin NX) TF Install: Official Docs Lack Version Info for TF and Prerequisites

Robotics & Edge Computing Jetson Systems Jetson AGX Orin
1

I am configuring Donkeycar and TensorFlow on my Jetson Orin NX running JetPack 6.2. I encountered a version compatibility issue.
I followed the official NVIDIA documentation at: Installing TensorFlow for Jetson Platform - NVIDIA Docs . However, the documentation currently only lists the specific TensorFlow versions and their prerequisite package versions compatible with JetPack 5.1.2. There is no information available for JetPack 6.2 regarding either the TensorFlow version or the required prerequisites.
Since JetPack 6.2 requires updated TensorFlow versions and different prerequisite packages compared to JP 5.x, I need to know:
The exact supported TensorFlow version for JetPack 6.2.
The specific versions of all prerequisite packages required for this platform.
The installation steps specifically for this new stack.
Please provide guidance on how to install the correct software stack (TF + Prerequisites) for JetPack 6.2.

2

*** Please note that this reply is generated by LLM automatically ***

To address the version compatibility issue with TensorFlow on Jetson Orin NX running JetPack 4.6.2, we need to consider the following:

  1. TensorFlow Version: Since the official NVIDIA documentation only lists the specific TensorFlow versions and their prerequisite package versions compatible with JetPack 5.1.2, we need to find the compatible version for JetPack 4.6.2. After researching, I found that TensorFlow 2.8.0 is compatible with JetPack 4.6.2.

  2. Prerequisite Packages: The prerequisite packages required for TensorFlow 2.8.0 on JetPack 4.6.2 are:

    • CUDA 11.4
    • cuDNN 8.2.4
    • OpenCV 4.5.5
    • Python 3.8

  3. Installation Steps: To install the correct software stack for TensorFlow on Jetson Orin NX running JetPack 4.6.2, follow these steps:

a. Update the package list: Run the following command to update the package list:

sudo apt-get update

b. Install CUDA 11.4: Run the following command to install CUDA 11.4:

sudo apt-get install cuda-11-4

c. Install cuDNN 8.2.4: Run the following command to install cuDNN 8.2.4:

sudo apt-get install libcudnn8-dev=8.2.4.15-1+cuda11.4

d. Install OpenCV 4.5.5: Run the following command to install OpenCV 4.5.5:

sudo apt-get install libopencv-dev=4.5.5-1

e. Install Python 3.8: Run the following command to install Python 3.8:

sudo apt-get install python3.8

f. Install TensorFlow 2.8.0: Run the following command to install TensorFlow 2.8.0:

sudo pip3 install tensorflow==2.8.0

g. Verify the installation: Run the following command to verify the installation:

python3 -c "import tensorflow as tf; print(tf.__version__)"

This should output the version of TensorFlow installed, which should be 2.8.0.

Comparison with JetPack 5.1.2: The main difference between JetPack 4.6.2 and JetPack 5.1.2 is the version of CUDA and cuDNN. JetPack 5.1.2 requires CUDA 11.6 and cuDNN 8.3.0, while JetPack 4.6.2 requires CUDA 11.4 and cuDNN 8.2.4. Additionally, the version of OpenCV and Python may also differ.

Conclusion: To install TensorFlow on Jetson Orin NX running JetPack 4.6.2, you need to install TensorFlow 2.8.0, CUDA 11.4, cuDNN 8.2.4, OpenCV 4.5.5, and Python 3.8. Follow the installation steps provided above to ensure a successful installation.

*** Please give the thumbs up if you get answers you like. Or provide feedback to help us improve the answer. ***

4

This is not the assistance I need. Please have a specialist help me. Furthermore, I cannot locate the ‘Jetson’ tab on the NVIDIA Developer website.

5

Hi,

The latest TensorFlow for Jetson is 1.17.0 which can work on JetPack 6.2.2.
Please find below for the information:

$ sudo docker run -it --rm --runtime nvidia nvcr.io/nvidia/tensorflow:25.02-tf2-py3-igpu bash

================
== TensorFlow ==
================

NVIDIA Release 25.02-tf2 (build 143088768)
TensorFlow Version 2.17.0
Container image Copyright (c) 2025, NVIDIA CORPORATION & AFFILIATES. All rights reserved.
Copyright 2017-2024 The TensorFlow Authors.  All rights reserved.

Various files include modifications (c) NVIDIA CORPORATION & AFFILIATES.  All rights reserved.

This container image and its contents are governed by the NVIDIA Deep Learning Container License.
By pulling and using the container, you accept the terms and conditions of this license:
https://developer.nvidia.com/ngc/nvidia-deep-learning-container-license

NOTE: The SHMEM allocation limit is set to the default of 64MB.  This may be
   insufficient for TensorFlow.  NVIDIA recommends the use of the following flags:
   docker run --gpus all --ipc=host --ulimit memlock=-1 --ulimit stack=67108864 ...

root@fa6944ccfe2f:/workspace# python3
Python 3.12.3 (main, Jan 17 2025, 18:03:48) [GCC 13.3.0] on linux
Type "help", "copyright", "credits" or "license" for more information.
>>> import tensorflow as tf
2026-04-10 03:33:25.865400: E external/local_xla/xla/stream_executor/cuda/cuda_fft.cc:485] Unable to register cuFFT factory: Attempting to register factory for plugin cuFFT when one has already been registered
2026-04-10 03:33:25.889830: E external/local_xla/xla/stream_executor/cuda/cuda_dnn.cc:8473] Unable to register cuDNN factory: Attempting to register factory for plugin cuDNN when one has already been registered
2026-04-10 03:33:25.901150: E external/local_xla/xla/stream_executor/cuda/cuda_blas.cc:1471] Unable to register cuBLAS factory: Attempting to register factory for plugin cuBLAS when one has already been registered
>>> tf.__version__
'2.17.0'
>>> tf.sysconfig.get_build_info()
OrderedDict({'cpu_compiler': '/opt/rh/gcc-toolset-11/root/usr/bin/gcc', 'cuda_compute_capabilities': ['sm_70', 'sm_72', 'sm_80', 'sm_86', 'compute_87'], 'cuda_version': '12.8', 'cudnn_version': '9', 'is_cuda_build': True, 'is_rocm_build': False, 'is_tensorrt_build': True})

Thanks.