Skip to content

PhotoFF Installation Guide (Linux & Windows)

Prerequisites

Before installing PhotoFF, ensure you have the following prerequisites:

  • Python 3.9 or newer
  • NVIDIA GPU with CUDA support
  • CUDA Toolkit 11.0 or newer – Required for compiling the CUDA components
  • CFFI – Used for interfacing between Python and the CUDA library
  • Pillow – Used for image loading, saving, and text rendering

Python Dependencies

Install the required Python packages:

pip install cffi pillow

Installing CUDA Toolkit

  1. Download the CUDA Toolkit from the NVIDIA Developer website
  2. Follow the instructions for your OS (Linux or Windows)
  3. Ensure nvcc is accessible: 
    nvcc --version

Compiling the CUDA Library

🐧 For Linux:

  1. Clone the repository: 

    git clone https://github.com/offerrall/photoff.git
cd photoff

  2. Compile the .so shared object: 

    python3 photoff_cuda_src/compile_linux.py

  3. You’ll get photoff.so.

  4. Make it available to your system:

Add to ~/.bashrc or ~/.zshrc: 

export LD_LIBRARY_PATH=/your/path/photoff:$LD_LIBRARY_PATH
Replace /your/path/photoff with the actual path where photoff.so is located. Then run: 
source ~/.bashrc
or 
source ~/.zshrc

🪟 For Windows:

  1. Clone the repository: 

    git clone https://github.com/offerrall/photoff.git
cd photoff

  2. Compile the .dll using: 

    python photoff_cuda_src/compile_windows.py

  3. Add the folder containing photoff.dll to your system PATH:

  4. Search for “Environment Variables” in the Start menu
  5. Edit the PATH variable, and add the folder path
  6. Restart your terminal or IDE

Installing the Python Package

Run this in the root of the project (after compilation):

pip install .

Verifying the Installation

Test your installation with the following Python script:

from photoff.operations.fill import fill_color
from photoff.io import save_image
from photoff.core.types import CudaImage, RGBA

# Create a 200x200 red square
img = CudaImage(200, 200)
fill_color(img, RGBA(255, 0, 0, 255))
save_image(img, "red_square.png")
img.free()

print("Installation successful!")

If you see the image red_square.png and the message “Installation successful!”, your setup is working.

Notes

  • CFFI will load the appropriate file based on your OS

License

PhotoFF is distributed under the MIT license.