Skip to content

Installation

Prerequisites for Conda package

  • Linux, Windows, macOS (not all features are available on macOS)
  • (Optional) Nvidia GPU with official Nvidia drivers installed for GPU acceleration

Install Mamba

The easiest way to install PlantSeg is by using the conda (Anaconda) or mamba (Miniforge) package manager. We recommend using mamba because it is faster and usually more consistent than conda.

To download Miniforge open a terminal and type:

curl -L -O "https://github.com/conda-forge/miniforge/releases/latest/download/Miniforge3-$(uname)-$(uname -m).sh"
bash Miniforge3-$(uname)-$(uname -m).sh

Then install by typing:

bash Miniforge3-$(uname)-$(uname -m).sh

and follow the installation instructions. Please refer to the Miniforge repo for more information, troubleshooting and uninstallation instructions. The miniforge installation file Miniforge3-*.sh can be deleted now.

The first step required to use the pipeline is installing mamba. The installation can be done by downloading the installer from the Miniforge repo. There you can find the download links for the latest version of Miniforge, troubleshooting and uninstallation instructions.

Install PlantSeg using Mamba

PlantSeg can be installed directly by executing in the terminal (or PowerShell on Windows). For conda users, the command is identical, just replace mamba with conda.

  • NVIDIA GPU version, CUDA=12.x

    mamba create -n plant-seg -c pytorch -c nvidia -c conda-forge pytorch pytorch-cuda=12.1 plant-seg --no-channel-priority

  • NVIDIA GPU version, CUDA=11.x

    mamba create -n plant-seg -c pytorch -c nvidia -c conda-forge pytorch pytorch-cuda=11.8 plant-seg --no-channel-priority

  • CPU version

    mamba create -n plant-seg -c pytorch -c nvidia -c conda-forge pytorch cpuonly plant-seg --no-channel-priority

  • NVIDIA GPU version, CUDA=12.x

    mamba create -n plant-seg -c pytorch -c nvidia -c conda-forge pytorch pytorch-cuda=12.1 nifty=1.2.1=*_4 plant-seg --no-channel-priority

  • NVIDIA GPU version, CUDA=11.x

    mamba create -n plant-seg -c pytorch -c nvidia -c conda-forge pytorch pytorch-cuda=11.8 nifty=1.2.1=*_4 plant-seg --no-channel-priority

  • CPU version

    mamba create -n plant-seg -c pytorch -c nvidia -c conda-forge pytorch cpuonly nifty=1.2.1=*_4 plant-seg --no-channel-priority

  • Apple silicon version

    mamba create -n plant-seg -c pytorch -c conda-forge python=3.11 pytorch::pytorch plant-seg --no-channel-priority

If you used older versions of PlantSeg, please delete the old config files in ~/.plantseg_models/configs/ after installing new PlantSeg.

The above command will create new conda environment plant-seg together with all required dependencies.

Please refer to the PyTorch website for more information on the available versions of PyTorch and the required CUDA version. The GPU version of Pytorch will also work on CPU only machines but has a much larger installation on disk.

Optional dependencies

Certain compressed TIFF files (e.g., Zlib, ZSTD, LZMA formats) require additional codecs to be processed correctly by PlantSeg. To handle such files, install the imagecodecs package:

conda activate plant-seg
pip install imagecodecs

If you plan to use SimpleITK-based watershed segmentation, you will need to install SimpleITK as an additional dependency:

conda activate plant-seg
pip install SimpleITK