This repository contains the source code for the trainable transform aspect of the Graph Fractional Fourier Transform: A Unified Theory paper published in IEEE Transactions on Signal Processing. It allows various methods to obtain the ordinary graph Fourier transform (GFT), and based on the obtained GFT, the graph fractional Fourier transform (GFRFT) is computed. The transform order is trainable, and the transform can be applied to any graph signal. This repository serves as a standalone Python package for the trainable GFRFT; please also refer to the original repository of the paper at koc-lab/gfrft-unified.