UKPLab / sentence-transformers

State-of-the-Art Text Embeddings
https://www.sbert.net
Apache License 2.0
15.19k stars 2.47k forks source link

Terminology Question: What model am I using? #1706

Open VirajVaitha123 opened 2 years ago

VirajVaitha123 commented 2 years ago

Hi there,

I'm a little confused with the terminology. I am using the model named 'all-mpnet-base-v2'.

When speaking to my team is it fair to say:

  1. We are using S-BERT, a sentence transformer trained on over 1B sentence pairs.

Or

  1. We decided to go with all-mpnet-base-v2 instead of S-BERT. This is a sentence-transformer based on BERT architecture, and trained on 1 Billion sentence pairs, that shown to have a high performance on the S-BERT website.

Why can't I see a model named S-BERT anywhere?

Thanks,

Viraj

nreimers commented 2 years ago

SBERT comes from the original Sentence-Bert paper.

I would recommend: We use all-mpnet-base-v2 from sentence-transformers, a model...

Then it is clear which model you used.