Embeddings Healthcare

Description

Word Embeddings lookup annotator that maps tokens to vectors.

Data Source

Trained on PubMed + ICD10 + UMLS + MIMIC III corpora https://www.nlm.nih.gov/databases/download/pubmed_medline.html

Copy S3 URI

How to use

model = WordEmbeddingsModel.pretrained("embeddings_healthcare","en","clinical/models")\
	.setInputCols(["document","token"])\
	.setOutputCol("word_embeddings")
val model = WordEmbeddingsModel.pretrained("embeddings_healthcare","en","clinical/models")
	.setInputCols("document","token")
	.setOutputCol("word_embeddings")
import nlu
nlu.load("en.embed.glove.healthcare").predict("""Put your text here.""")

Results

Word2Vec feature vectors based on embeddings_healthcare.

Model Information

Name: embeddings_healthcare
Type: WordEmbeddingsModel
Compatibility: Spark NLP 2.4.4+
License: Licensed
Edition: Official
Input labels: [document, token]
Output labels: [word_embeddings]
Language: en
Dimension: 400.0