Description
This model was imported from Hugging Face (link) and it’s been trained on Quora Question Pairs dataset, leveraging Distil-BERT embeddings and DistilBertForSequenceClassification for text classification purposes. As an input, it requires two questions separated by a space.
Predicted Entities
non_duplicated, duplicated
How to use
document_assembler = DocumentAssembler() \
.setInputCol('text') \
.setOutputCol('document')
tokenizer = Tokenizer() \
.setInputCols(['document']) \
.setOutputCol('token')
sequenceClassifier = DistilBertForSequenceClassification.pretrained("distilbert_base_sequence_classifier_qqp", "en")\
.setInputCols(["document",'token'])\
.setOutputCol("class")
pipeline = Pipeline(stages=[document_assembler, tokenizer, sequenceClassifier])
light_pipeline = LightPipeline(pipeline.fit(spark.createDataFrame([['']]).toDF("text")))
result1 = light_pipeline.annotate("Do we have to go there? Are you a doctor?")
result2 = light_pipeline.annotate("Do you want to eat something? Are you hungry?")
val document_assembler = DocumentAssembler()
.setInputCol("text")
.setOutputCol("document")
val tokenizer = Tokenizer()
.setInputCols(Array("document"))
.setOutputCol("token")
val sequenceClassifier = DistilBertForSequenceClassification.pretrained("distilbert_base_sequence_classifier_qqp", "en")
.setInputCols(Array("document", "token"))
.setOutputCol("class")
val pipeline = new Pipeline().setStages(Array(document_assembler, tokenizer, sequenceClassifier))
val example1 = Seq.empty["Do we have to go there? Are you a doctor?"].toDS.toDF("text")
val example2 = Seq.empty["Do you want to eat something? Are you hungry?"].toDS.toDF("text")
val result1 = pipeline.fit(example1).transform(example1)
val result2 = pipeline.fit(example2).transform(example2)
import nlu
nlu.load("en.classify.qqp.distil_bert.base").predict("""Do you want to eat something? Are you hungry?""")
Results
['non_duplicated']
['duplicated']
Model Information
| Model Name: | distilbert_base_sequence_classifier_qqp |
| Compatibility: | Spark NLP 3.4.0+ |
| License: | Open Source |
| Edition: | Official |
| Input Labels: | [sentence, token] |
| Output Labels: | [class] |
| Language: | en |
| Size: | 249.8 MB |
| Case sensitive: | true |
| Max sentence length: | 256 |
References
https://quoradata.quora.com/First-Quora-Dataset-Release-Question-Pairs