Sentence Embeddings - sbert medium (tuned)

Description

This model maps sentences & documents to a 512-dimensional dense vector space by using average pooling on top of BERT model. It’s also fine-tuned on the RxNorm dataset to help generalization over medication-related datasets.

Predicted Entities

Live Demo Open in Colab Download Copy S3 URI

How to use

sentence_embeddings = BertSentenceEmbeddings.pretrained("jsl_sbert_medium_rxnorm", "en", "clinical/models")\
        .setInputCols(["sentence"])\
        .setOutputCol("sbert_embeddings")

val sentence_embeddings = BertSentenceEmbeddings.pretrained('jsl_sbert_medium_rxnorm', 'en','clinical/models')
      .setInputCols("sentence")
      .setOutputCol("sbert_embeddings")

import nlu
nlu.load("en.embed_sentence.bert_medium.rxnorm").predict("""Put your text here.""")

Results

Gives a 512-dimensional vector representation of the sentence.

Model Information

Data Source

Tuned on RxNorm dataset.