-
![image](https://user-images.githubusercontent.com/21833174/196373242-1da2c8b2-a45b-43a5-a98f-6c632214532b.png)
可能会影响recall指标和ndcg指标计算
可能需要对这一行增加判断
-
**Is your feature request related to a problem? Please describe.**
If you have a query/doc pair that you feel isn't rateable, should we capture that? We offer a skip option, but what aobut if you ca…
-
在RO_RS这种评估设置下,group_by_user可以是True或False,实验发现其对一些指标比如NDCG会有一些影响,看了下官网,general模型通常需要按用户对数据进行分组并执行否定采样,一般按如下设置:
```
group_by_user: True
training_neg_sample_num: 1
```
那Knowledge-based模型一般怎么设置group_…
-
IN the paper the threshold said to be 3 but in fact it's 4 when run python train_implicit.py --dataset yahooR3
[see threshold value assignment in the code](https://github.com/DongHande/AutoDebias/b…
-
hi,
when would lightgbm 2.3.2 be released? as the documentation online (https://lightgbm.readthedocs.io/) is 2.3.2 with some new features but the latest python package is lightgbm 2.3.1. Thanks.
-
您好,首先很感谢贵组开源该有效率的推荐系统框架,但针对其中GPU实现top-K预测部分的优化,本人有问题期待您的帮助!
- 请问在实现过程中, re-organize操作是否必要呢?假如已经对训练集样本置为负无穷,那么直接进行排序argsort得到的就是TopK预测结果的下标了,再与ground truth进行对比(这里可以用numpy的与或操作即可)则比较直接得到最终的recall,ndc…
-
就默认数据集默认模型就可以get到。
-
When using 'full' instead of 'uni100', the program gives the following error. No other settings are changed.
Traceback (most recent call last):
File "C:\Users\Blake\OneDrive\Documents\From_Old_C…
-
**Is your feature request related to a problem? Please describe.**
I was using the newly added Tensorboard feature.
The current version has the `Valid_score`, which seems to only contain the `va…
-
hello,
`def auc(vector_predict, vector_true, device = 'cuda'):
pos_indexes = torch.where(vector_true == 1)[0].to(device)
pos_whe=(vector_true == 1).to(device)
sort_indexes = torch.arg…