YiyanXu / DiffRec

Diffusion Recommender Model
168 stars 23 forks source link

ratio of ml-1m_clean #15

Open xiezhuoxuan opened 8 months ago

xiezhuoxuan commented 8 months ago
  1. 论文4.1.1第二段提到了“splits the sorted interactions into training, validation, and testing sets with the ratio of 7:1:2”,我在下载到的ml-1m_clean数据集中发现train,valid,test的数据条数分别为403277,110722,57532,这是7:2:1的比例
  2. DiffRec/L-DiffRec/main.py中第306行调用evaluate的第三个参数是否应为mask_train而不是mask_tv
  3. 参考DiffRec/L-DiffRec/inference.py中ml-1m_clean的参数设置,发现测试集的Recall和NDCG指标明显高于验证集,这是否是由于划分数据集时按时间排序(论文4.1.1第二段)导致训练、测试、验证集不满足独立同分布性质