Overview

The “User Guide” section is for those who want to quickly get started and run models. An overview of the running pipeline is shown in the figure below:

XGCN data process and model running pipeline

XGCN supports taking text files of the graph and evaluation sets as input, and processing them into a standard “dataset instance” format which is used by all the downstream models. Once a dataset instance is generated, you can easily run models with XGCN’s APIs.

In the following, we’ll introduce how to use XGCN from the four aspects below:

Data Preparation: how to process your text data into our “dataset instances”.
Training and Evaluation: the model training and evaluation APIs.
Supported Models: available models and their configurations.
Usage Examples: end-to-end usage examples.