Introduction

AutoGL (i.e. Auto Graph Learning) is an automatic machine learning (AutoML) toolkit specified for graph datasets & tasks.

It will automatically handle all the stages involved in graph learning problems, including dataset download & management, data preprocessing and feature engineering, model selection and training, hyper-parameter tuning and ensemble, which will reduce human labors and biases in the machine learning loop by a large scale. This toolkit also serves as a platform for users to implement and test their own auto or graph learning methods. The workflow below gives the overall framework of AutoGL.

Actively under development by THUMNLab
Feel free to open issues or contact us at autogl@tsinghua.edu.cn if you have any comments or suggestions!

from autogl.datasets import build_datasets_from_name from autogl.solver import AutoNodeClassifier cora = build_datasets_from_name('cora') solver = AutoNodeClassifier( graphModelList=['gcn', 'gat'] ) solver.fit(cora) prediction = solver.predict(cora)

Installation

Please first install the requirements of AutoGL and then install with the following command.


Install from pip

pip install auto-graph-learning

Install from source

git clone https://github.com/THUMNLab/AutoGL
python setup.py install

Requirements
Python >= 3.6.0
PyTorch >= 1.5.1
# Follow PyTorch to install
PyTorch-Geometric >= 1.5.1
# Follow PyTorch-Geometric to install

Incoming Features

  • Neural Architecture Search
  • Large-scale graph datasets support
  • More graph tasks (e.g. Link prediction, Heterogeneous graph tasks, Spatial & Temporal tasks)
  • Graph Boosting & Bagging
  • More graph library backend support (e.g. Deep Graph Library)
  • Related Links

    Here is our IJCAI tutorial on AutoML & Multimedia.

    View Tutorial >

    Contact Us

    Please contact us through autogl@tsinghua.edu.cn or xin_wang@tsinghua.edu.cn

    Send E-mail >

    Github

    Visit our github to get access to codes and ask questions.

    View Github >