Gujarati RoBERTa Embeddings (from surajp)

Description

Pretrained RoBERTa Embeddings model, uploaded to Hugging Face, adapted and imported into Spark NLP. RoBERTa-hindi-guj-san is a Gujarati model orginally trained by surajp.

Download

How to use

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

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

data = spark.createDataFrame([["મને સ્પાર્ક એનએલપી ગમે છે"]]).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_RoBERTa_hindi_guj_san","gu") 
    .setInputCols(Array("document", "token")) 
    .setOutputCol("embeddings")

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

val data = Seq("મને સ્પાર્ક એનએલપી ગમે છે").toDF("text")

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

Model Information

Model Name: roberta_embeddings_RoBERTa_hindi_guj_san
Compatibility: Spark NLP 3.4.2+
License: Open Source
Edition: Official
Input Labels: [sentence, token]
Output Labels: [bert]
Language: gu
Size: 252.1 MB
Case sensitive: true

References

  • https://huggingface.co/surajp/RoBERTa-hindi-guj-san
  • https://github.com/goru001/inltk
  • https://www.kaggle.com/disisbig/hindi-wikipedia-articles-172k
  • https://www.kaggle.com/disisbig/gujarati-wikipedia-articles
  • https://www.kaggle.com/disisbig/sanskrit-wikipedia-articles
  • https://twitter.com/parmarsuraj99
  • https://www.linkedin.com/in/parmarsuraj99/