how were the m1-1m.random constructed? about llmrank

[hyp1231]

Hi, thanks for your attention! The items stored in ml-1m.random are randomly selected candidate item IDs, or in other words, negative instances. These candidate item IDs will be fused with ground truth items further in Trainer. The code to produce this file has been uploaded, and how this file is used has been described in [data preparation]. Please feel free to leave comments if you have any questions.

您好，感谢关注！ml-1m.random中存储的商品 ID 是随机选择的，它们会被看成负样本。这些候选商品会在 Trainer 中进一步与正样本融合成为完整的候选集，以用于排序。生成该文件的代码已上传，且该文件是如何被使用的已在[data preparation]中描述。这个说明可能比较简略，如果有什么没写清楚的地方，欢迎随时在这里留言，谢谢！

[remember00000]

同时我查看了random中和原始数据中重合数量，也就是ground truth的数量，发现位置和数量并没有规律或者一致性，想咨询这块的作用和构造方式，谢谢🙏

[hyp1231]

Due to the lack of updates on this issue for a long time, it is temporarily closed. If you have any questions, feel free to reopen it at any time.