为什么我跑SGL的结果会比这篇论文高这么多(alibaba数据集)

RUCAIBox / NCL

[WWW'22] Official PyTorch implementation for "Improving Graph Collaborative Filtering with Neighborhood-enriched Contrastive Learning".

119 stars 20 forks source link

为什么我跑SGL的结果会比这篇论文高这么多(alibaba数据集) #53

Open njt123123 opened 1 year ago

njt123123 commented 1 year ago

Tue 28 Nov 2023 23:22:47 INFO best valid : OrderedDict([('recall@10', 0.0544), ('recall@20', 0.08), ('recall@50', 0.1273), ('ndcg@10', 0.0297), ('ndcg@20', 0.0362), ('ndcg@50', 0.0456)]) Tue 28 Nov 2023 23:22:47 INFO test result: OrderedDict([('recall@10', 0.0533), ('recall@20', 0.0786), ('recall@50', 0.1246), ('ndcg@10', 0.0293), ('ndcg@20', 0.0357), ('ndcg@50', 0.0448)])

hyp1231 commented 1 year ago

请问这两个结果对应的运行脚本是？

njt123123 commented 1 year ago

请问这两个结果对应的运行脚本是？