-
Hi， Dear：
Thanks for your open source code！
Recently I read about hybrid search with infinity，best hybrid search method is：_**BM25 + Dense vector + Sparse vector + colBert**_
Can qdrant…
-
### Bug Description
I couldn't install llama-index-retrievers-bm25 on Python 3.11
.9 with Ubuntu 22.0
`m/.venv/include -I/usr/include/python3.11 -c src/Stemmer.c -o build/temp.linux-x86_64-cpython…
-
Running your code on the test200 benchmark gives me an error
dietz@jelly:~/datasci/code2/team1$ python ef_ranking_document_generate.py ../all.test200.cbor.outlines ...test200.cbor.paragraphs outpu…
-
### Describe your problem
What is the retrieve logic after adding questions and keyword extraction? The questions, keywords, and original content are all in one chunk. Is there a multi-channel retrie…
-
### Question Validation
- [x] I have searched both the documentation and discord for an answer.
### Question
Hi, I have stored a collection of nodes in a PGVector DB, I'd like to be able to retriev…
-
### Is your feature request related to a problem?
Related to [RFC](https://github.com/opensearch-project/neural-search/issues/126). The current problem with the RFC is that when we are combining scor…
-
## Result
| | | X95 | | WSS@95 | |
|-----------------------|--------------|:------:|:---:|:------:|:----:|
| Dataset | Treatment | …
-
### Description
The system shall build an inverted index on the sparse vector embedding that are loaded by the user into the system, in an efficient manner.
### Acceptance criteria
1. The time re…
-
우선, 이 글을 쓰는 이유는 모든 파트가 할 것이 많지만 그 중, Retriever를 조금 건드려 봤습니다.
그래서 아직 역할이 안 정해졌지만 Retriever하시는 분들에게 조금이라도 편하고 빠르게 진행할 수 있도록 시도해봤습니다.
(아직 코드에 허점이 있을 것 같지만 성과를 보였기에 말씀드리겠습니다.)
흐름은 rank_bm25의 github …
-
新版langchain-chatchat的本地知识库搜索算法有问题，同样的搜索条件，我用老版的余弦相似度算法拿到的切片相似度还可以，但是新版bm25+向量库检索的方式，感觉搜索到的切片相关度比较低，而且只有bm25有效果，将ensemble切换到vectorstore，直接搜索不出来
xqxls updated
3 months ago