# PyKEEN

A Python library for training and evaluating knowledge graph embedding models, with dozens of built-in models and datasets and a high-level pipeline function for benchmarking a model on a dataset in a few lines of code.

PyKEEN is a Python package for training and evaluating knowledge graph embedding models, including support for incorporating multi-modal information. A single pipeline function provides a high-level entry point that trains a chosen model, such as TransE, on a chosen dataset with a stochastic local closed-world-assumption training loop by default, and evaluates it with rank-based metrics, returning a result object with the trained model, training loop and evaluation results attached.

It ships with 37 built-in datasets, plus five inductive datasets for the inductive link-prediction setting, and 40 built-in embedding models, all sharing consistent APIs so a model, a training loop or a custom triples factory built from a user's own dataset can be swapped in without changing the rest of a pipeline.

It requires Python 3.10 or newer, installs from PyPI or directly from source on GitHub, and is released under the MIT license.

Website: https://pykeen.readthedocs.io/en/stable/