Legal English Bert Embeddings (Base Uncased)

Description

Legal Pretrained Bert Embeddings model, trained with uncased text, uploaded to Hugging Face, adapted and imported into Spark NLP. legal-bert-base-uncased is a English model orginally trained by nlpaueb.

Download

How to use

documentAssembler = DocumentAssembler() \
     .setInputCol("text") \
     .setOutputCol("document")

tokenizer = Tokenizer() \
    .setInputCols("document") \
    .setOutputCol("token")
  
embeddings = BertEmbeddings.pretrained("bert_embeddings_legal_bert_base_uncased","en") \
    .setInputCols(["document", "token"]) \
    .setOutputCol("embeddings")
    
pipeline = Pipeline(stages=[documentAssembler, tokenizer, embeddings])

data = spark.createDataFrame([["I love Spark NLP"]]).toDF("text")

result = pipeline.fit(data).transform(data)
val documentAssembler = new DocumentAssembler() 
      .setInputCol("text") 
      .setOutputCol("document")
 
val tokenizer = new Tokenizer() 
    .setInputCols(Array("document"))
    .setOutputCol("token")

val embeddings = BertEmbeddings.pretrained("bert_embeddings_legal_bert_base_uncased","en") 
    .setInputCols(Array("document", "token")) 
    .setOutputCol("embeddings")

val pipeline = new Pipeline().setStages(Array(documentAssembler, tokenizer, embeddings))

val data = Seq("I love Spark NLP").toDF("text")

val result = pipeline.fit(data).transform(data)
import nlu
nlu.load("en.embed.legal_bert_base_uncased").predict("""I love Spark NLP""")

Model Information

Model Name: bert_embeddings_legal_bert_base_uncased
Compatibility: Spark NLP 3.4.2+
License: Open Source
Edition: Official
Input Labels: [sentence, token]
Output Labels: [bert]
Language: en
Size: 410.1 MB
Case sensitive: false

References

  • https://huggingface.co/nlpaueb/legal-bert-base-uncased
  • https://aclanthology.org/2020.findings-emnlp.261/
  • https://eur-lex.europa.eu/
  • https://www.legislation.gov.uk/
  • https://case.law/
  • https://www.sec.gov/edgar.shtml
  • https://archive.org/details/legal_bert_fp
  • http://nlp.cs.aueb.gr/