Hindi DistilBERT Embeddings (from neuralspace-reverie)

Description

Pretrained DistilBERT Embeddings model, uploaded to Hugging Face, adapted and imported into Spark NLP. indic-transformers-hi-distilbert is a Hindi model orginally trained by neuralspace-reverie.

Download

How to use

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

tokenizer = Tokenizer() \
    .setInputCols("document") \
    .setOutputCol("token")
  
embeddings = DistilBertEmbeddings.pretrained("distilbert_embeddings_indic_transformers_hi_distilbert","hi") \
    .setInputCols(["document", "token"]) \
    .setOutputCol("embeddings")
    
pipeline = Pipeline(stages=[documentAssembler, tokenizer, embeddings])

data = spark.createDataFrame([["मुझे स्पार्क एनएलपी पसंद है"]]).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 = DistilBertEmbeddings.pretrained("distilbert_embeddings_indic_transformers_hi_distilbert","hi") 
    .setInputCols(Array("document", "token")) 
    .setOutputCol("embeddings")

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

val data = Seq("मुझे स्पार्क एनएलपी पसंद है").toDF("text")

val result = pipeline.fit(data).transform(data)
import nlu
nlu.load("hi.embed.indic_transformers_hi_distilbert").predict("""मुझे स्पार्क एनएलपी पसंद है""")

Model Information

Model Name: distilbert_embeddings_indic_transformers_hi_distilbert
Compatibility: Spark NLP 3.4.2+
License: Open Source
Edition: Official
Input Labels: [sentence, token]
Output Labels: [bert]
Language: hi
Size: 247.6 MB
Case sensitive: true

References

  • https://huggingface.co/neuralspace-reverie/indic-transformers-hi-distilbert
  • https://oscar-corpus.com/