search

divelab / DIG

A library for graph deep learning research
https://diveintographs.readthedocs.io/
GNU General Public License v3.0
1.82k stars 281 forks source link

Graph augmentation #185

Closed anmol-anand closed 1 year ago

anmol-anand commented 1 year ago

The GraphAug model has three parts: (1) Reward Generator: A GNN model that rewards label invariant augmentations. (2) Generator: A model that can generate label invariant graph augmentations. (3) Testing: Testing improvement in classification accuracy after using the augmented dataset.

anmol-anand commented 1 year ago

Hi Youzhi, i have addressed previous comments and added the generator code. The review is back with you. Thanks!

anmol-anand commented 1 year ago

Hi Youzhi, thanks for the feedback. I have addressed these comments. The review is back with you.

anmol-anand commented 1 year ago

Hi Youzhi,

I have added the code for augmented classification. Can you please have a look.

Thank you. Anmol

anmol-anand commented 1 year ago

Hi Youzhi,

  1. I have removed the code for GINE model.
  2. Earlier the example python scripts were runnable from the root directory DIG, now they are runnable from examples/auggraph/GraphAug

The review is back with you.

anmol-anand commented 1 year ago

Hi Youzhi, I have added sphinx documentation for the auggraph module. Can you please review once.

anmol-anand commented 1 year ago

Hi Youzhi, Thanks for the quick review. I have updated the documentation as you suggested.

anmol-anand commented 1 year ago

Thanks for the review Youzhi!