Edit on GitHub
  • John Snow Labs
  • Sep 22, 2023

German BertEmbeddings Base Cased model (from PM-AI)

  • de
  • open_source
  • bert_embeddings
  • onnx

Description

Pretrained BertEmbeddings model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP. bi-encoder_msmarco_bert-base_german is a German model originally trained by PM-AI.

Download Copy S3 URI

How to use

documentAssembler = DocumentAssembler() \
    .setInputCols(["text"]) \
    .setOutputCols("document")

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

bert_loaded = BertEmbeddings.pretrained("bert_embeddings_bi_encoder_msmarco_base_german","de") \
    .setInputCols(["document", "token"]) \
    .setOutputCol("embeddings") \
    .setCaseSensitive(True)
    
pipeline = Pipeline(stages=[documentAssembler, tokenizer, bert_loaded])

data = spark.createDataFrame([["I love Spark NLP"]]).toDF("text")

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

val documentAssembler = new DocumentAssembler() 
    .setInputCols(Array("text")) 
    .setOutputCols(Array("document"))
      
val tokenizer = new Tokenizer()
    .setInputCols("document")
    .setOutputCol("token")
 
val bert_loaded = BertEmbeddings.pretrained("bert_embeddings_bi_encoder_msmarco_base_german","de") 
    .setInputCols(Array("document", "token"))
    .setOutputCol("embeddings")
    .setCaseSensitive(true)    
   
val pipeline = new Pipeline().setStages(Array(documentAssembler, tokenizer, bert_loaded))

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

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

Model Information

Model Name: bert_embeddings_bi_encoder_msmarco_base_german
Compatibility: Spark NLP 5.1.0+
License: Open Source
Edition: Official
Input Labels: [document, token]
Output Labels: [bert]
Language: de
Size: 409.7 MB
Case sensitive: true

References

  • https://huggingface.co/PM-AI/bi-encoder_msmarco_bert-base_german
  • https://github.com/UKPLab/sentence-transformers
  • https://microsoft.github.io/msmarco/#ranking
  • https://arxiv.org/abs/2108.13897
  • https://openreview.net/forum?id=wCu6T5xFjeJ
  • https://public.ukp.informatik.tu-darmstadt.de/thakur/BEIR/datasets/
  • https://github.com/beir-cellar/beir
  • https://github.com/beir-cellar/beir/blob/main/examples/retrieval/training/train_msmarco_v3_margin_MSE.py
  • https://sbert.net/datasets/msmarco-hard-negatives.jsonl.gz
  • https://github.com/beir-cellar/beir/blob/main/examples/retrieval/training/train_msmarco_v3_margin_MSE.py%5D
  • https://github.com/UKPLab/sentence-transformers/blob/master/examples/training/ms_marco/README.md
  • https://github.com/beir-cellar/beir/blob/main/examples/retrieval/training/train_msmarco_v3_margin_MSE.py
  • https://arxiv.org/abs/2104.12741
  • https://www.elastic.co/guide/en/elasticsearch/reference/current/index-modules-similarity.html#bm25
  • https://en.th-wildau.de/
  • https://senseaition.com/
  • https://www.linkedin.com/in/herrphilipps
  • https://efre.brandenburg.de/efre/de/
  • https://www.senseaition.com
  • https://www.th-wildau.de