-
I am trying to run on MIND on BERT4REC, I get the .inter file with MIND using the instructions in recbole.io.
These are my parameters
```
topk : [5,10,20,50]
epochs : 5
loss_type : BPR
metric…
-
thank you for share your nice algorithm and code
occurred error at evaluation.py line 94
TypeError: can't convert cuda:0 device type tensor to numpy. Use Tensor.cpu() to copy the tensor to host mem…
-
您好,我想确认一下rating和label在RecBole中的定义是什么?
从这个问题我了解到如果想要分析某个值,比如在这里是我的商品评分,那就将它列为label #817 ,可是这样的话rating_field的含义是什么呢?并且这位提问者最终在 #819 训练context-aware的时候`.inter`文件并没有定义任何除了user_id和item_id以外的feature,conf…
-
Hi,
Thank you for your good work!
I am trying to finetune your [`egovlp.pth`](https://drive.google.com/file/d/1-cP3Gcg0NGDcMZalgJ_615BQdbFIbcj7/view) checkpoint to obtain the numbers that are re…
-
-
我在尝试运行CDAE等方法时,因为看见底层代码实际是包含预测评分的部分的,所以尝试将指标改成RMSE之类的方法后,报错,NotImplementedError: full sort can't use IndividualEvaluator,然后将排序改为RO_RS之后仍然报错。
我想知道,如何查看这些方法在评分预测目标下的运行结果
-
**Describe the bug**
Timestamp field not loaded in `ml-1m` dataset.
Same command (model, loss_type) works for `ml-100k` so expected just changing `ml-1m` to work
**To Reproduce**
`recbole.__…
-
跑BPR跑last-fm数据集,评测用mode: uni100
Traceback (most recent call last):
File "C:/notC/TRUST/RecBole-master/run_recbole.py", line 25, in
run_recbole(model=args.model, dataset=args.dataset, config_…
-
Hi, I run the diffnetplus code on yelp data, get HR=0.3676,NDCG=0.2274, which is consistent with the performance in paper.
When I remove the user and item features by`self.fusion_user_embedding = se…
-
请问,上下文信息怎么加入,比如预测用户每一次购买的商品TopK ,用户下单的时点信息(比如星期几),位置信息(城市)
之前没搞过这个。。。请见谅
我用Wide&Deep 模型的时候,尝试过放在.inter 里面,但是在eval 的时候,这些字段没有带进来。eval的时候,interaction 只有.user的字段,缺少了上下文信息,就会导致找不到field 报错。看了下源码,取出来的…