qpm

A simple GUI to perform bacteria segmentation and qpm reconstruction.

To launch the GUI

Install git and uv and then run:

uvx git+https://github.com/fdrgsp/qpm

GPU & CUDA on Windows

To use the GPU version of qpm on Windows, you need to have a compatible NVIDIA GPU and the corresponding CUDA toolkit installed.

NOTE: NVIDIA Drivers should be already installed.

Option 1: Using uvx (recommended)

This is the easiest way to install qpm on Windows with CUDA GPU support:

uvx --index https://download.pytorch.org/whl/cu126 git+https://github.com/fdrgsp/qpm[cuda]

NOTE: change cu126 to match your CUDA version (to get the correct URL, check the PyTorch Get Started page).

Option 2: Manual installation

You need to go through the following steps only once, just the first time you install qpm:

• download the repository (zip file)

• extract the folder (it should contain src, .gitignore, pyproject.toml and README.md)

• in your terminal, cd into this folder (e.g. cd Path/to/qpm)

• uv venv qpm-env

• qpm-env\Scripts\activate

• uv pip install .

• update the torch and torchvision packages to the version compatible with your CUDA version, e.g. for CUDA 12.6:

  uv pip install -U torch torchvision --index-url https://download.pytorch.org/whl/cu126

  NOTE: to get the correct URL, check the PyTorch Get Started page.

• launch the GUI with: qpm

Once you have done this once, in the future, to run the qpm GUI, you can simply use your terminal and run:

• cd Path/to/qpm
• qpm-env\Scripts\activate
• qpm