Bangla Bert Embeddings

Description

Pretrained Bert Embeddings model, uploaded to Hugging Face, adapted and imported into Spark NLP. bangla-bert-base is a Bangla model orginally trained by sagorsarker.

Download

How to use

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

tokenizer = Tokenizer() \
.setInputCols("document") \
.setOutputCol("token")

embeddings = BertEmbeddings.pretrained("bert_embeddings_bangla_bert_base","bn") \
.setInputCols(["document", "token"]) \
.setOutputCol("embeddings")

pipeline = Pipeline(stages=[documentAssembler, tokenizer, embeddings])

data = spark.createDataFrame([["আমি স্পার্ক এনএলপি ভালোবাসি"]]).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_bangla_bert_base","bn") 
.setInputCols(Array("document", "token")) 
.setOutputCol("embeddings")

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

val data = Seq("আমি স্পার্ক এনএলপি ভালোবাসি").toDF("text")

val result = pipeline.fit(data).transform(data)
import nlu
nlu.load("bn.embed.bangala_bert").predict("""আমি স্পার্ক এনএলপি ভালোবাসি""")

Model Information

Model Name: bert_embeddings_bangla_bert_base
Compatibility: Spark NLP 3.4.2+
License: Open Source
Edition: Official
Input Labels: [sentence, token]
Output Labels: [bert]
Language: bn
Size: 617.6 MB
Case sensitive: true

References

  • https://huggingface.co/sagorsarker/bangla-bert-base
  • https://github.com/sagorbrur/bangla-bert
  • https://arxiv.org/abs/1810.04805
  • https://github.com/google-research/bert
  • https://oscar-corpus.com/
  • https://dumps.wikimedia.org/bnwiki/latest/
  • https://github.com/sagorbrur/bnlp
  • https://github.com/sagorbrur/bangla-bert
  • https://github.com/google-research/bert
  • https://twitter.com/mapmeld
  • https://github.com/rezacsedu/Classification_Benchmarks_Benglai_NLP
  • https://github.com/sagorbrur/bangla-bert/blob/master/notebook/bangla-bert-evaluation-classification-task.ipynb
  • https://github.com/sagorbrur/bangla-bert/tree/master/evaluations/wikiann
  • https://arxiv.org/abs/2012.14353
  • https://arxiv.org/abs/2104.08613
  • https://arxiv.org/abs/2107.03844
  • https://arxiv.org/abs/2101.00204
  • https://github.com/sagorbrur
  • https://www.tensorflow.org/tfrc
  • https://github.com/google-research/bert