Update README.md
Browse files
README.md
CHANGED
@@ -447,6 +447,7 @@ model-index:
|
|
447 |
# nb-bert-base-matryoshka
|
448 |
|
449 |
This is a [sentence-transformers](https://www.SBERT.net) model finetuned from [NbAiLab/nb-bert-base](https://huggingface.co/NbAiLab/nb-bert-base) on the utdanning_pair, [ltg/norquad](https://huggingface.co/datasets/ltg/norquad) and [NbAiLab/mnli-norwegian](https://huggingface.co/datasets/NbAiLab/mnli-norwegian) datasets. It maps sentences & paragraphs to a 768-dimensional dense vector space and can be used for semantic textual similarity, semantic search, paraphrase mining, text classification, clustering, and more.
|
|
|
450 |
|
451 |
## Model Details
|
452 |
|
|
|
447 |
# nb-bert-base-matryoshka
|
448 |
|
449 |
This is a [sentence-transformers](https://www.SBERT.net) model finetuned from [NbAiLab/nb-bert-base](https://huggingface.co/NbAiLab/nb-bert-base) on the utdanning_pair, [ltg/norquad](https://huggingface.co/datasets/ltg/norquad) and [NbAiLab/mnli-norwegian](https://huggingface.co/datasets/NbAiLab/mnli-norwegian) datasets. It maps sentences & paragraphs to a 768-dimensional dense vector space and can be used for semantic textual similarity, semantic search, paraphrase mining, text classification, clustering, and more.
|
450 |
+
As with the BGE architecture and Artic-embed I use the final hidden state of the [CLS] token as the embedding vector, instead of a mean pooling strategy.
|
451 |
|
452 |
## Model Details
|
453 |
|