Description
This model has been trained on a curated list of clinical and biomedical datasets. It has been fine-tuned for Few-Shot Assertion but can also be utilized for other purposes, such as Classification and Retrieval-Augmented Generation (RAG).
How to use
document_assembler = DocumentAssembler()\
.setInputCol("text")\
.setOutputCol("document")
e5_embeddings = E5Embeddings.pretrained("e5_base_v2_embeddings_medical_assertion_jsl", "en", "clinical/models")\
.setInputCols(["document"])\
.setOutputCol("e5_embeddings")
pipeline = Pipeline().setStages([document_assembler, e5_embeddings])
text = [
["I feel a bit drowsy after taking an insulin."],
["Peter Parker is a nice guy and lives in New York"]
]
data = spark.createDataFrame(text).toDF("text")
result = pipeline.fit(data).transform(data)
val document_assembler = new DocumentAssembler()
.setInputCol("text")
.setOutputCol("document")
val e5_embeddings = E5Embeddings.pretrained("e5_base_v2_embeddings_medical_assertion_jsl", "en", "clinical/models")
.setInputCols(Array("document"))
.setOutputCol("e5_embeddings")
val pipeline = new Pipeline().setStages(Array(document_assembler, e5_embeddings))
val result = pipeline.fit(data).transform(data)
Results
+--------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------+
| embeddings|
+--------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------+
|[{sentence_embeddings, 0, 43, I feel a bit drowsy after taking an insulin., {sentence -> 0}, [-0.03397934, -0.001984031, 0.0041575357, 0.005777784, 0.008281726, -0.0050704903, 0.0367161, 0.01730193...|
|[{sentence_embeddings, 0, 47, Peter Parker is a nice guy and lives in New York, {sentence -> 0}, [0.107267894, 0.05047583, 0.025502523, 0.047211427, -0.014814768, -0.01552255, 0.033394374, 0.026739...|
+--------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------+
Model Information
| Model Name: | e5_base_v2_embeddings_medical_assertion_jsl |
| Compatibility: | Healthcare NLP 5.3.3+ |
| License: | Licensed |
| Edition: | Official |
| Input Labels: | [document] |
| Output Labels: | [embeddings] |
| Language: | en |
| Size: | 393.8 MB |
| Case sensitive: | false |