Russian BertForTokenClassification Cased model (from IlyaGusev)

Description

Pretrained BertForTokenClassification model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP. rubertconv_toxic_editor is a Russian model originally trained by IlyaGusev.

Predicted Entities

delete, insert, equal, replace

Live Demo Open in Colab Download Copy S3 URI

How to use

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

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

tokenClassifier = BertForTokenClassification.pretrained("bert_token_classifier_rubertconv_toxic_editor","ru") \
    .setInputCols(["document", "token"]) \
    .setOutputCol("ner")

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

data = spark.createDataFrame([["PUT YOUR STRING HERE"]]).toDF("text")

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

val documentAssembler = new DocumentAssembler()
    .setInputCol("text")
    .setOutputCol("document")

val tokenizer = new Tokenizer()
    .setInputCols("document")
    .setOutputCol("token")

val tokenClassifier = BertForTokenClassification.pretrained("bert_token_classifier_rubertconv_toxic_editor","ru")
    .setInputCols(Array("document", "token"))
    .setOutputCol("ner")

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

val data = Seq("PUT YOUR STRING HERE").toDS.toDF("text")

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

Model Information

References

• https://huggingface.co/IlyaGusev/rubertconv_toxic_editor
• https://colab.research.google.com/drive/1NUSO1QGlDgD-IWXa2SpeND089eVxrCJW
• https://github.com/skoltech-nlp/russe_detox_2022/tree/main/data
• https://github.com/IlyaGusev/rudetox/blob/main/rudetox/marker/compute_tags.py
• https://github.com/IlyaGusev/rudetox/blob/main/rudetox/marker/train.py
• https://github.com/IlyaGusev/rudetox/blob/main/dvc.yaml#L367