Python Tip: Hardware-Agnostic pip freeze

Avoid hardware-specific dependencies in requirements.txt

Generate hardware-agnostic requirements.txt files without NVIDIA CUDA dependencies.

Author

Shep Bryan IV

Python Tip: Hardware-Agnostic `pip freeze`

Working with PyTorch? Avoid hardware-specific dependencies like NVIDIA CUDA from your requirements.txt. Here’s a one-liner to make it hardware-agnostic.

The Problem

We often use pip freeze to generate a requirements.txt file, but it includes all installed packages, including hardware-specific ones like NVIDIA CUDA libraries.

This can cause compatibility issues for users without the same hardware setup.

# This includes hardware-specific packages
!pip freeze > requirements.txt

The Solution

Use this command to generate a requirements.txt that includes PyTorch but excludes hardware-dependent packages:

# This excludes hardware-specific dependencies
!pip list --not-required --format=freeze > requirements.txt

Wrap-Up

Now you can share your requirements without worrying about hardware compatibility issues.

Follow me for more tips.
Shep Bryan IV