Installation
Nkululeko requires Python 3.9 or higher. The easiest and safest way to install Nkululeko is using a virtual environment, either with venv or conda.
Using Conda
$ conda create -n nkululeko python=3.9
$ conda activate nkululeko
$ pip install nkululeko
Using venv
$ python3 -m venv nkululeko
$ source nkululeko/bin/activate # On Windows: nkululeko\Scripts\activate
$ pip install nkululeko
Current version: 0.94.1
Optional Dependencies
Nkululeko supports optional dependencies through extras. You can install them using the following syntax:
# Install with PyTorch support (default PyTorch)
$ pip install nkululeko[torch]
# Install with CPU-only PyTorch
$ pip install nkululeko[torch-cpu]
# Or manually:
$ pip install torch torchvision torchaudio --index-url https://download.pytorch.org/whl/cpu
# Install with Spotlight support
$ pip install PyYAML # Install PyYAML first to avoid dependency issues
$ pip install nkululeko[spotlight]
# Install with TensorFlow support
$ pip install nkululeko[tensorflow]
# Install all optional dependencies
$ pip install nkululeko[all]
For specific model support:
# For SQUIM model (requires nightly PyTorch)
$ pip install nkululeko[torch-nightly]
# Or manually:
$ pip uninstall -y torch torchvision torchaudio
$ pip install --pre torch torchvision torchaudio --extra-index-url https://download.pytorch.org/whl/nightly/cpu
Development Installation
For the development version, install directly from the source:
# Clone the repository
$ git clone https://github.com/felixbur/nkululeko.git
$ cd nkululeko
# Install in editable mode with all dependencies
$ pip install -e ".[all]"
# Or with specific extras
$ pip install -e ".[torch,spotlight]"
Verify Installation
Check that Nkululeko is installed correctly:
# Via pip
$ pip list | grep nkululeko
# Inside Python
>>> import nkululeko
>>> nkululeko.__version__
>>> nkululeko.__file__ # Shows installation path
If you see the version of Nkululeko (e.g., 0.94.1), you are ready to go.