tensorflow

• Installation for Tensorflow and Keras environment for legacy Dell Precision Workstations and HP ZBook G-series Workstations
• Testing/tested on Dell Precision 5510 Workstation, HP ZBook G6 Workstation
• OS environments testing/tested on: Windows 10 Pro, Windows 11 Pro, LnOS Arch (Arch-based)
• wip: environment.yml provides TF and Keras versioning information and Python version needed. Runs on Python 3.8, Tensorflow 2.6.0, Keras 2.6.0, CUDA 11.8, CUDNN 8.9.7, and Windows WSL2 - Ubuntu 24.04
• Arch Linux wip
• Windows WSL2 for Linux does not have NUMA support.
• Run Tensorflow environment in a conda virtual environment.

---

Getting Started

Ubuntu/UbuntuWSL

To install Anaconda on Ubuntu/UbuntuWSL see Anaconda installation docs at Technical Documentation section. Refer back to steps below Clone repository

git clone https://github.com/rcghpge/tensorflow.git
cd tensorflow

Initialize a Conda environment (install Conda if needed - see technical documentation)

conda init
conda config --set auto_activate_base false # disables venv auto-activate 
conda create -n tfenv python=3.8
conda activate tfenv

You should be able to replicate the environment via

# List available Conda environments
conda env list

# Install Conda environment
conda env update -f environment.yml --prune

Arch Linux/ArchWSL

wip - Arch Linux installation is a little different. Currently not detecting GPU.

# Install Conda
sudo pacman -S python-conda
conda init
conda config --set auto_activate_base false
conda create -n tfenv python=3.8
conda activate tfenv

# Check Conda local environment
# List available conda environments
conda env list

# Install Conda environment
conda env update -f environment.yml --prune

You can install TensorFlow from pacman on Arch Linux and test Python versioning from latest. Currently only detects CPU - no GPU detected.

]Verify Tensorflow and Keras environment

python -c "import tensorflow as tf; print('TensorFlow Keras Version:', tf.keras.__version__)"
python -c "import tensorflow as tf; print('Num CPUs Available:', len(tf.config.list_physical_devices('CPU')))"
python -c "import tensorflow as tf; print('Num GPUs Available:', len(tf.config.list_physical_devices('GPU')))"

Test Tensorflow and Keras environment

I have provided a test/ directory with sample models to test the development environment. Run test models

python3 test/testkeras.py

---

Technical Documentation

• Ubuntu Desktop - Ubuntu Installation docs
• Arch Linux - Arch Linux Installation docs
• UbuntuWSL - UbuntuWSL Instalation docs
• LnOS Arch - LnOS Arch (Arch-based) Installation Wiki
• Conda - Main Conda Installation docs
• Conda on Arch Linux - Conda Installation docs for Arch Linux
• Anaconda - Anaconda Installation docs for Ubuntu/UbuntuWSL
• TensorFlow - TensorFlow Installation docs
• Keras - Keras Installation docs

---

If there is a stable solution for older Dell workstations feel free to contact me or send PR's to optimize this stack for legacy Dell Precision Workstation line of machines.

---