wang-shuo/graphrec_pytorch

A PyTorch implementation of Graph Neural Networks for Social Recommendation (GraphRec)

graphrec_pytorch's Issues

关于collate_fn的疑问

为什么每个用户要选取truncate_len个交互的物品，每个用户选取truncate_len个朋友呢？

How to predict a list of items for specific user

Hello, I would like to generate recommendations for specific users in order to further analyze the results of the items.

Is there a function that I do not see in your code implementaion?

Or maybe there is an easy way to do that ?

Thank you in advance!

Results from the original paper

Hi, have you compared the results of your implementation and those reported in the original paper? Is there any difference?

The result of the default parameter does not correspond to the result provided in the original paper

My result is Test: MAE: 0.8483, RMSE: 1.1361. Why I cannot achieve the result in the paper?

What's the reproduced results of the proposed GreaphRec in the paper?

1

How to decide truncate length?

How to decide the truncate length for batches of data?