Installation

The packages requires at least Python 3.12. To install the package via pip, simply run:

pip install glasses-detector

Or, to install it from source, run:

git clone https://github.com/mantasu/glasses-detector
cd glasses-detector && pip install .

Tip

You may want to set up PyTorch (use Nightly version for compatibility with Python \(\ge\) 3.12) in advance to enable GPU support for your device. Note that CUDA is backwards compatible, thus even if you have the newest version of CUDA Toolkit, PyTorch GPU acceleration should work just fine.

Note

By default, the required models will be automatically downloaded and saved under Torch Hub directory, which by default is ~/.cache/torch/hub/checkpoints. For more information and how to change it, see Torch Hub documentation.