Reproduction of GNNAutoScale using PGL

rusty1s / pyg_autoscale

Implementation of "GNNAutoScale: Scalable and Expressive Graph Neural Networks via Historical Embeddings" in PyTorch

http://arxiv.org/abs/2106.05609

MIT License

160 stars 27 forks source link

Reproduction of GNNAutoScale using PGL #10

Open DesmonDay opened 3 years ago

DesmonDay commented 3 years ago

We are impressed by your great work in Auto-Scaling GNNs, which gives us a lot of inspiration. Therefore, we recently reproduce your GNNAutoScale framework using PGL and PaddlePaddle. And the code is available at https://github.com/PaddlePaddle/PGL/tree/main/apps/GNNAutoScale.

Please feel free to give us your valuable advice. Thank you!

rusty1s commented 3 years ago

Hi @DesmonDay, this is really great. Thank you so much! I appreciate the efforts of converting GNNAutoScale to PGL. Let me know if you need any advice in implementing the asynchronous device transfer scheme, this should give you a big additional boost in runtimes :)