<ul><li>Torch-Choice is an open-source library for choice modeling with Python and PyTorch, providing a ChoiceDataset data structure for efficient database management.</li><li>The package includes models like multinomial logit and nested logit, supporting regularization during model estimation.</li><li>It allows GPU utilization for scaling to large datasets with computational efficiency and offers flexibility in model initialization.</li><li>The package's computational efficiencies are compared with mlogit in R, showcasing scalability on large-scale datasets.</li></ul>

Torch-Choice: A PyTorch Package for Large-Scale Choice Modeling with Python

Discover more