-
1. FT后loss值一直降不下去,参数如下,本地cpu跑的,5轮训练后差不多这样,这是什么原因呢或者有什么优化的地方
{"epoch": 4.18,"learning_rate": 1.6492693110647182e-06,"loss": 0.2706,"step": 2000}
torchrun --nproc_per_node 1 -m FlagEmbedding.baai_gene…
-
When a documents get a lot of ReadingEvents it (or parts of it) will be indexed many times, thus reducing the inverse document frequency. This should probably be fixed in DiMe's indexing.
One way is …
-
感谢贵团队的工作!
想请教一下,检索完成以后采用排序模型进行Rerank,这个Rerank的值设置为多大比较合适?0.5吗,低于0.5就是不相关,高于0.5就是相关?
-
[Youtube - Playlist](https://youtube.com/playlist?list=PLCoJWKqBHERuUmKM5OSpDc1vurIbppoDx)
Aruna Lakshmanan gave an awesome Lightning talk with tons of in-depth advice around search signals. I thoug…
-
Hi, jingtao
can you share the warmup model for the doc ranking data?
-
`get_relevant_chunks` gets sets of relevant passages via various retrieval methods (semantic/dense, sparse-embedding based, lexical/keyword, fuzzy); most of these produce scores (except fuzzy, I think…
-
I started trying to integrate the Matches API into the UnifiedHighlighter, but there's a fairly heavy impedance mismatch between the way the two of them work (eg Matches doesn't give you freqs, it's e…
-
Add a couple details of what we are doing here and why please @newgraph-lschick
push progress to repo daily or work through issues/barriers to progress here
R component
QGIS component
https://githu…
-
Hello, thanks for your great work!
I have a little question about the code implementation. In your code repo, there is a HLATR reranker (i.e. **class HLATR_reranker** in src/t5_model), which i supp…
-
Hi, thanks fo this fantastic repo and its documentation!
I have a question: I am working on a research project on fact-verification in Czech and as the first step we are trying various approaches t…