wrong example for guide homepage.

[Shi-Jialiang]

Sorry, but shouldn’t the example on the homepage of LightGCN's default dataset be Gowalla (not Yelp)?

[Shi-Jialiang]

[Re-bin]

Hi 👋!

Thanks for pointing that out! I will fix the typo on the homepage :)

Best regards,
Xubin

[Shi-Jialiang]

Hi,

Thank you for your response!

I have another related question: I’m using your LightGCN, and my results are recall@20: 0.2258 and ndcg@20: 0.1451. However, the results in the LightGCN paper are 0.1830 and 0.1554.

Is it acceptable for there to be some variation in these values? Thank you!

[Re-bin]

Hi,

This is because the SSLRec paper and the LightGCN paper use two different datasets (with differing numbers of users, items, and interactions), although both are processed from the Gowalla data.

Additionally, the training strategy differs from the original LightGCN code. In SSLRec, we also use validation data for early stopping to avoid overfitting.

Hope the answer is helpful to you :)

Best regards,
Xubin