Chinese Bert Embeddings (Bert, Whole Word Masking)

Description

Pretrained Bert Embeddings model, uploaded to Hugging Face, adapted and imported into Spark NLP. chinese-bert-wwm-ext is a Chinese model orginally trained by hfl.

Live Demo Open in Colab Download Copy S3 URI

How to use

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

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

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

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

data = spark.createDataFrame([["I love Spark NLP"]]).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_chinese_bert_wwm_ext","zh") 
.setInputCols(Array("document", "token")) 
.setOutputCol("embeddings")

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

val data = Seq("I love Spark NLP").toDF("text")

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

import nlu
nlu.load("zh.embed.chinese_bert_wwm_ext").predict("""I love Spark NLP""")

Model Information

References

• https://huggingface.co/hfl/chinese-bert-wwm-ext
• https://arxiv.org/abs/1906.08101
• https://github.com/google-research/bert
• https://github.com/ymcui/Chinese-BERT-wwm
• https://github.com/ymcui/MacBERT
• https://github.com/ymcui/Chinese-ELECTRA
• https://github.com/ymcui/Chinese-XLNet
• https://github.com/airaria/TextBrewer
• https://github.com/ymcui/HFL-Anthology
• https://arxiv.org/abs/2004.13922
• https://arxiv.org/abs/1906.08101