Italian BertForQuestionAnswering model (from luigisaetta)

Description

Pretrained Question Answering model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP. squad_it_xxl_cased_hub1 is a Italian model originally trained by luigisaetta.

Download Copy S3 URI

How to use

documentAssembler = MultiDocumentAssembler() \
.setInputCols(["question", "context"]) \
.setOutputCols(["document_question", "document_context"])

spanClassifier = BertForQuestionAnswering.pretrained("bert_qa_squad_xxl_cased_hub1","it") \
.setInputCols(["document_question", "document_context"]) \
.setOutputCol("answer")\
.setCaseSensitive(True)

pipeline = Pipeline(stages=[documentAssembler, spanClassifier])

data = spark.createDataFrame([["Qual è il mio nome?", "Mi chiamo Clara e vivo a Berkeley."]]).toDF("question", "context")

result = pipeline.fit(data).transform(data)
val documentAssembler = new MultiDocumentAssembler() 
.setInputCols(Array("question", "context")) 
.setOutputCols(Array("document_question", "document_context"))

val spanClassifer = BertForQuestionAnswering.pretrained("bert_qa_squad_xxl_cased_hub1","it") 
.setInputCols(Array("document", "token")) 
.setOutputCol("answer")
.setCaseSensitive(true)

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

val data = Seq("Qual è il mio nome?", "Mi chiamo Clara e vivo a Berkeley.").toDF("question", "context")

val result = pipeline.fit(data).transform(data)
import nlu
nlu.load("it.answer_question.squad.bert.xxl_cased").predict("""Qual è il mio nome?|||"Mi chiamo Clara e vivo a Berkeley.""")

Model Information

Model Name: bert_qa_squad_xxl_cased_hub1
Compatibility: Spark NLP 4.0.0+
License: Open Source
Edition: Official
Input Labels: [document_question, document_context]
Output Labels: [answer]
Language: it
Size: 413.3 MB
Case sensitive: true
Max sentence length: 512

References

  • https://huggingface.co/luigisaetta/squad_it_xxl_cased_hub1
  • https://github.com/luigisaetta/nlp-qa-italian/blob/main/train_squad_it_final1.ipynb