tanya-jp / G-Mixup

Re-implementation of G-Mixup: Graph Data Augmentation for Graph Classification

Repository from Github https://github.comtanya-jp/G-MixupRepository from Github https://github.comtanya-jp/G-Mixup

To run this code, upload G-Mixup.ipynb file on your Google Colab account. For speeding up the process, choose T4 GPU as your runtime type. To change your runtime type, you need to choose Change runtime type from Runtime menu.

In the first cell you can define the hyper parameters, such as number of hidden layers and number of epochs. There are five different acceptable datasets, including IMDB-BINARY, IMDB-MULTI, REDDIT-BINARY, REDDIT-MULTI-5K, and REDDIT-MULTI-12K. If you are using the free version of Google Colab, the provided RAM cannot run this code for REDDIT-MULTI-5K and REDDIT-MULTI-12K.

After defining everything, you just need to choose Run all from Runtime menu. You can see the results from the two last cells of code.

About

Re-implementation of G-Mixup: Graph Data Augmentation for Graph Classification


Languages

Language:Jupyter Notebook 52.6%Language:TeX 47.4%