snap-stanford / ogb

Benchmark datasets, data loaders, and evaluators for graph machine learning
https://ogb.stanford.edu
MIT License
1.89k stars 397 forks source link

ogbn-products training and testing split #287

Closed JiuhaiChen closed 2 years ago

JiuhaiChen commented 2 years ago

Hi, OGB team:

I was wondering if you have the source code for ogbn-products training and testing split, leaderboard says you use the sales ranking (popularity) to split nodes into training/validation/test sets. Where can i get the sales ranking?

Thanks!

weihua916 commented 2 years ago

We do not provide the sales ranking, but we just used the meta data provided in the original Amazon dataset.

HaniItani commented 1 year ago

Hello @weihua916,

Do you, by any chance, have the raw text Amazon dataset from which ogbn-products was generated? In particular, I'm interested to map graph node features to text descriptions and labels to product text labels. I'm looking for something similar to the raw text files you provide for ogbn-arxiv and ogbn-papers100M. I did not find it in the dataset linked in Cluster-GCN. I also tried to download Amazon3M from here, but the mapping is not straightforward.

Any help or guidance will be much appreciated. Best regards, Hani

bigheiniu commented 1 year ago

Hi Hani, @HaniItani You can refer https://github.com/amzn/pecos/tree/mainline/examples/giant-xrt for the text datasets.