-
Thanks for your early directions about bi_encoder training, Now i have complete these process.
I review the sample generation on Cross-Encoder of ms marco dataset. it use simple random sample as nega…
-
According to https://sbert.net/examples/applications/retrieve_rerank/README.html#re-ranker-cross-encoder
the cross encoder "outputs a single score between 0 and 1". I do get these results with some u…
niebb updated
3 years ago
-
我使用的环境是
pytorch 1.4.0
transformers 2.8.0
参照着文档https://github.com/thunlp/OpenMatch/blob/master/docs/experiments-msmarco.md 中的训练命令
```
CUDA_VISIBLE_DEVICES=0 \
python train.py\
-task r…
-
Hi,
1/ Is it possible to fine tune a multilingual bi-encoder on specific domain data using unsupervised synthethic query generation? If yes, are performance comparable to fine-tuning with supervised …
-
Impressive work, it's efficient and potent. Here's a suggestion.
The search is the critical component! It's the bottleneck for answering all queries, given you already possess a robust corpus.
C…
-
**Describe the bug**
E5-small/base/large v1/v2
Hi unilm team,
Thank you so much for the great project! We try to replace sentence-transformer with E5. When using the official example scripts wi…
-
pt.text.sliding() uses iterrows() this could probably be faster.
-
Hi, Thank you for amazing framework. I'm planning to use this framework for question answer system.
Can you please help me how can I train model with my own domain data? I know that nboost doesn't s…
-
[@scottgit wrote](https://github.com/alerque/stack-verse-mapper/issues/9#issuecomment-172342050):
> How does this algorithm handle multiple references? So lets say a person put the Romans 7:3-6 and M…
-
Hi,
I am a creator of the [ANTIQUE](https://ciir.cs.umass.edu/downloads/nfL6/) dataset -- a passage retrieval dataset for non-factoid questions answering. Please find the paper that explains the da…