Dutch RoBERTa Embeddings (from DTAI-KULeuven)

Description

Pretrained RoBERTa Embeddings model, uploaded to Hugging Face, adapted and imported into Spark NLP. robbertje-1-gb-bort is a Dutch model orginally trained by DTAI-KULeuven.

Download

How to use

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

tokenizer = Tokenizer() \
    .setInputCols("document") \
    .setOutputCol("token")
  
embeddings = RoBertaEmbeddings.pretrained("roberta_embeddings_robbertje_1_gb_bort","nl") \
    .setInputCols(["document", "token"]) \
    .setOutputCol("embeddings")
    
pipeline = Pipeline(stages=[documentAssembler, tokenizer, embeddings])

data = spark.createDataFrame([["Ik hou van vonk 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 = RoBertaEmbeddings.pretrained("roberta_embeddings_robbertje_1_gb_bort","nl") 
    .setInputCols(Array("document", "token")) 
    .setOutputCol("embeddings")

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

val data = Seq("Ik hou van vonk nlp").toDF("text")

val result = pipeline.fit(data).transform(data)
import nlu
nlu.load("nl.embed.robbertje_1_gb_bort").predict("""Ik hou van vonk nlp""")

Model Information

Model Name: roberta_embeddings_robbertje_1_gb_bort
Compatibility: Spark NLP 3.4.2+
License: Open Source
Edition: Official
Input Labels: [sentence, token]
Output Labels: [bert]
Language: nl
Size: 172.8 MB
Case sensitive: true

References

  • https://huggingface.co/DTAI-KULeuven/robbertje-1-gb-bort
  • http://github.com/iPieter/robbert
  • http://github.com/iPieter/robbertje
  • https://www.clinjournal.org/clinj/article/view/131
  • https://www.clin31.ugent.be
  • https://arxiv.org/abs/2101.05716