CAMeLBERT-CA NER 模型

模型描述

CAMeLBERT-CA NER 模型是通过对 CAMeLBERT Classical Arabic (CA) 模型进行微调来构建的命名实体识别（NER）模型。对于微调，我们使用了 ANERcorp 数据集。关于我们使用的微调过程和超参数的详细信息可以在我们的论文 *" The Interplay of Variant, Size, and Task Type in Arabic Pre-trained Language Models "中找到。

• 我们的微调代码可以在 here 找到。

预期用途

您可以直接将 CAMeLBERT-CA NER 模型作为我们的 CAMeL Tools NER 组件的一部分使用（推荐），也可以作为 transformers pipeline 的一部分使用。

如何使用

要将模型与 CAMeL Tools NER 组件一起使用：

>>> from camel_tools.ner import NERecognizer
>>> from camel_tools.tokenizers.word import simple_word_tokenize
>>> ner = NERecognizer('CAMeL-Lab/bert-base-arabic-camelbert-ca-ner')
>>> sentence = simple_word_tokenize('إمارة أبوظبي هي إحدى إمارات دولة الإمارات العربية المتحدة السبع')
>>> ner.predict_sentence(sentence)
>>> ['O', 'B-LOC', 'O', 'O', 'O', 'O', 'B-LOC', 'I-LOC', 'I-LOC', 'O']

您还可以直接使用 transformers pipeline 来使用 NER 模型：

>>> from transformers import pipeline
>>> ner = pipeline('ner', model='CAMeL-Lab/bert-base-arabic-camelbert-ca-ner')
>>> ner("إمارة أبوظبي هي إحدى إمارات دولة الإمارات العربية المتحدة السبع")
[{'word': 'أبوظبي',
  'score': 0.9895730018615723,
  'entity': 'B-LOC',
  'index': 2,
  'start': 6,
  'end': 12},
 {'word': 'الإمارات',
  'score': 0.8156259655952454,
  'entity': 'B-LOC',
  'index': 8,
  'start': 33,
  'end': 41},
 {'word': 'العربية',
  'score': 0.890906810760498,
  'entity': 'I-LOC',
  'index': 9,
  'start': 42,
  'end': 49},
 {'word': 'المتحدة',
  'score': 0.8169114589691162,
  'entity': 'I-LOC',
  'index': 10,
  'start': 50,
  'end': 57}]

注意：要下载我们的模型，您需要 transformers>=3.5.0。否则，您可以手动下载模型。

引用

@inproceedings{inoue-etal-2021-interplay,
    title = "The Interplay of Variant, Size, and Task Type in {A}rabic Pre-trained Language Models",
    author = "Inoue, Go  and
      Alhafni, Bashar  and
      Baimukan, Nurpeiis  and
      Bouamor, Houda  and
      Habash, Nizar",
    booktitle = "Proceedings of the Sixth Arabic Natural Language Processing Workshop",
    month = apr,
    year = "2021",
    address = "Kyiv, Ukraine (Online)",
    publisher = "Association for Computational Linguistics",
    abstract = "In this paper, we explore the effects of language variants, data sizes, and fine-tuning task types in Arabic pre-trained language models. To do so, we build three pre-trained language models across three variants of Arabic: Modern Standard Arabic (MSA), dialectal Arabic, and classical Arabic, in addition to a fourth language model which is pre-trained on a da of the three. We also examine the importance of pre-training data size by building additional models that are pre-trained on a scaled-down set of the MSA variant. We compare our different models to each other, as well as to eight publicly available models by fine-tuning them on five NLP tasks spanning 12 datasets. Our results suggest that the variant proximity of pre-training data to fine-tuning data is more important than the pre-training data size. We exploit this insight in defining an optimized system selection model for the studied tasks.",
}