Skip to main content

Installation

PartiNet can be installed using several methods. Choose the option that best fits your environment and requirements.

Prerequisites​

  • Python 3.8 or higher
  • CUDA-compatible GPU (recommended for optimal performance)
  • Git (for source installation)

Please note that AMD/Intel GPUs have not been tested, but may still used with PartiNet

This method gives you the latest version and full control over the installation:

# Create new python environment
conda create -n partinet python=3.9
conda activate partinet
# or using venv
python -m venv partinet-env
source partinet-env/bin/activate

# Install PartiNet
git clone git@github.com:WEHI-ResearchComputing/PartiNet.git
cd PartiNet
pip install .

Method 2: Apptainer/Singularity Container​

For users who prefer containerized environments or have limited system permissions:

Option A: Pull and store locally

apptainer pull partinet.sif oras://ghcr.io/wehi-researchcomputing/partinet:main-singularity
apptainer exec --nv --no-home -B /vast partinet.sif partinet --help

Option B: Run directly from registry

apptainer exec --nv --no-home \
    -B /vast oras://ghcr.io/wehi-researchcomputing/partinet:main-singularity \
    partinet --help

Container options explained:

  • --nv: Enables NVIDIA GPU support
  • --no-home: Prevents mounting your home directory
  • -B /vast: Mounts the /vast directory (adjust path as needed for your data directory)

Method 3: Docker Container​

For Docker users:

docker pull ghcr.io/wehi-researchcomputing/partinet:main
docker run --gpus all -v /path/to/your/data:/data \
    ghcr.io/wehi-researchcomputing/partinet:main partinet --help

Docker options explained:

  • --gpus all: Enables GPU support (requires nvidia-docker)
  • -v /path/to/your/data:/data: Mounts your data directory

Verification​

After installation, verify that PartiNet is working correctly:

partinet --help

You should see version information and available commands.

GPU Support​

PartiNet is designed to leverage GPU acceleration for optimal performance. Ensure you have:

  • NVIDIA GPU with CUDA compute capability 3.5+ (e.g., NVIDIA A30, A100, H100)
  • CUDA drivers installed
  • For containers: nvidia-docker (Docker) or --nv flag (Apptainer)

AMD and Intel GPUs have not been tested and may not support full PartiNet functionality

Model Weights​

PartiNet model weights are available on HuggingFace. Weights can be downloaded through the browser or through CLI via Git LFS

# Verify Git LFS is installed
git lfs --help
mkdir PartiNet_weights
cd PartiNet_weights
git clone git@hf.co:MihinP/PartiNet

You will see two .pt files available: denoised_micrographs.pt and raw_micrographs.pt.

Next Steps​

Once installed, proceed to Getting Started to run your first PartiNet analysis.