Portuguese Bert Embeddings (from pierreguillou)

Description

Pretrained Bert Embeddings model, uploaded to Hugging Face, adapted and imported into Spark NLP. bert-large-cased-pt-lenerbr is a Portuguese model orginally trained by pierreguillou.

Download

How to use

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

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

data = spark.createDataFrame([["Eu amo 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_bert_large_cased_pt_lenerbr","pt") 
    .setInputCols(Array("document", "token")) 
    .setOutputCol("embeddings")

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

val data = Seq("Eu amo Spark NLP").toDF("text")

val result = pipeline.fit(data).transform(data)

Model Information

Model Name: bert_embeddings_bert_large_cased_pt_lenerbr
Compatibility: Spark NLP 3.4.2+
License: Open Source
Edition: Official
Input Labels: [sentence, token]
Output Labels: [bert]
Language: pt
Size: 1.3 GB
Case sensitive: true

References

  • https://huggingface.co/pierreguillou/bert-large-cased-pt-lenerbr
  • https://medium.com/@pierre_guillou/nlp-modelos-e-web-app-para-reconhecimento-de-entidade-nomeada-ner-no-dom%C3%ADnio-jur%C3%ADdico-b658db55edfb
  • https://github.com/piegu/language-models/blob/master/Finetuning_language_model_BERtimbau_LeNER_Br.ipynb
  • https://paperswithcode.com/sota?task=Fill+Mask&dataset=pierreguillou%2Flener_br_finetuning_language_model