-
## 🐛 Bug
Hi TorchMetrics Team,
In the following example, nDCG calculation using GPU tensors spent 2 times longer the time using CPU tensors and numpy array.
### To Reproduce
The codes were…
-
您好,关于Ciao这个数据集跑出来的结果,HR=0.6374,NDCG=0.4065 。能否告知一下代码调参的数值是啥,我跑不出来这个结果
-
**Is your feature request related to a problem? Please describe.**
In some extremely challenging scenarios for information retrieval, items relevant to a query follow a long tail distribution. Thus, …
-
I'm using `DenseRetrievalExactSearch` & `DenseRetrievalParallelExactSearch` with https://huggingface.co/sentence-transformers/sentence-t5-xxl on SciFact and getting different results. I'm using them i…
-
i find ndcg print from train,py,i want to know the means of ndcg
-
Our use of nDCG as a prediction accuracy metric is non-standard and should probably be dropped.
Equivalent behavior can be obtained by using a `recommend` task with unlimited size and a candidate s…
-
你是否还在苦恼大型语言模型的「幻觉」问题?是否希望构建自己网站或者自己需求专属的聊天机器人?是否正在寻找简单易用且准确性高的RAG模型?
Denser Retriever 这款工具帮到你!
- 官方文档:https://retriever.denser.ai/
- GitHub地址:https://github.com/denser-org/denser-retriever
## …
-
Hi, through [Wikipedia](https://en.wikipedia.org/wiki/Discounted_cumulative_gain) and related [papers](http://proceedings.mlr.press/v30/Wang13.pdf), I find that NDCG=DCG/IDCG. The maximum possible DCG…
-
## Goals
- collect data on HyDE effectiveness with open source LLMs
## Tested LLMs
- [x] Mistral 7B instruct v0.2
- [x] Llama2 7B chat hf
- [x] Zepyher-7B beta
## Benchmark Results
##…
-
Hi,
Just to make sure, is ranking.valid-ndcg-truncation actually reflects the 'k' of NDCG@k used to optimize on the training data ?
Thanks