Skip to main content

TensorFlow

TensorFlow is an end-to-end machine learning and deep learning framework developed and open-sourced by Google. It is widely used to build, train, and deploy machine learning models on CPUs, GPUs, and embedded platforms such as NVIDIA Jetson.

This document describes how to install and verify TensorFlow on Jetson systems.

1. System Information​

Before installing TensorFlow, verify your system and CUDA environment.

System Information Overview

Query CUDA Version​

nvcc --version

CUDA Version Output

2. Install TensorFlow​

You can install TensorFlow using either offline or online methods.

2.1 Offline Installation​

Step 1: Download TensorFlow Package​

Download the TensorFlow wheel matching your JetPack version:

https://developer.download.nvidia.com/compute/redist/jp/v61/tensorflow/

Step 2: Install the Package​

cd ~/Downloads
sudo pip3 install tensorflow-2.16.1+nv24.08-cp310-cp310-linux_aarch64.whl
note

Ensure the TensorFlow version matches your JetPack and Python version.

2.2 Online Installation​

sudo pip3 install --extra-index-url https://developer.download.nvidia.com/compute/redist/jp/v61 tensorflow==2.16.1+nv24.08

2.3 Install NumPy​

sudo pip install numpy==1.23.5
warning

Using incompatible NumPy versions may cause runtime errors.

3. Verify Installation​

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

TensorFlow Version Output

References​

Maintained by HemiHex for Jetson-based AI workflows.