CAMeLBERT: A collection of pre-trained models for Arabic NLP tasks

Model description

CAMeLBERT is a collection of BERT models pre-trained on Arabic texts with different sizes and variants. We release pre-trained language models for Modern Standard Arabic (MSA), dialectal Arabic (DA), and classical Arabic (CA), in addition to a model pre-trained on a mix of the three. We also provide additional models that are pre-trained on a scaled-down set of the MSA variant (half, quarter, eighth, and sixteenth). The details are described in the paper "The Interplay of Variant, Size, and Task Type in Arabic Pre-trained Language Models."

This model card describes CAMeLBERT-Mix (bert-base-arabic-camelbert-mix), a model pre-trained on a mixture of these variants: MSA, DA, and CA.

Intended uses

You can use the released model for either masked language modeling or next sentence prediction. However, it is mostly intended to be fine-tuned on an NLP task, such as NER, POS tagging, sentiment analysis, dialect identification, and poetry classification. We release our fine-tuninig code here.

How to use

You can use this model directly with a pipeline for masked language modeling:

>>> from transformers import pipeline
>>> unmasker = pipeline('fill-mask', model='CAMeL-Lab/bert-base-arabic-camelbert-mix')
>>> unmasker("الهدف من الحياة هو [MASK] .")
[{'sequence': '[CLS] الهدف من الحياة هو النجاح. [SEP]',
  'score': 0.10861027985811234,
  'token': 6232,
  'token_str': 'النجاح'},
 {'sequence': '[CLS] الهدف من الحياة هو.. [SEP]',
  'score': 0.07626965641975403,
  'token': 18,
  'token_str': '.'},
 {'sequence': '[CLS] الهدف من الحياة هو الحياة. [SEP]',
  'score': 0.05131986364722252,
  'token': 3696,
  'token_str': 'الحياة'},
 {'sequence': '[CLS] الهدف من الحياة هو الموت. [SEP]',
  'score': 0.03734956309199333,
  'token': 4295,
  'token_str': 'الموت'},
 {'sequence': '[CLS] الهدف من الحياة هو العمل. [SEP]',
  'score': 0.027189988642930984,
  'token': 2854,
  'token_str': 'العمل'}]

Note: to download our models, you would need transformers>=3.5.0. Otherwise, you could download the models manually.

Here is how to use this model to get the features of a given text in PyTorch:

from transformers import AutoTokenizer, AutoModel
tokenizer = AutoTokenizer.from_pretrained('CAMeL-Lab/bert-base-arabic-camelbert-mix')
model = AutoModel.from_pretrained('CAMeL-Lab/bert-base-arabic-camelbert-mix')
text = "مرحبا يا عالم."
encoded_input = tokenizer(text, return_tensors='pt')
output = model(**encoded_input)

and in TensorFlow:

from transformers import AutoTokenizer, TFAutoModel
tokenizer = AutoTokenizer.from_pretrained('CAMeL-Lab/bert-base-arabic-camelbert-mix')
model = TFAutoModel.from_pretrained('CAMeL-Lab/bert-base-arabic-camelbert-mix')
text = "مرحبا يا عالم."
encoded_input = tokenizer(text, return_tensors='tf')
output = model(encoded_input)

Training data

• MSA (Modern Standard Arabic)
  • The Arabic Gigaword Fifth Edition
  • Abu El-Khair Corpus
  • OSIAN corpus
  • Arabic Wikipedia
  • The unshuffled version of the Arabic OSCAR corpus
• DA (dialectal Arabic)
  • A collection of dialectal Arabic data described in our paper.
• CA (classical Arabic)
  • OpenITI (Version 2020.1.2)

Training procedure

We use the original implementation released by Google for pre-training. We follow the original English BERT model's hyperparameters for pre-training, unless otherwise specified.

Preprocessing

• After extracting the raw text from each corpus, we apply the following pre-processing.
• We first remove invalid characters and normalize white spaces using the utilities provided by the original BERT implementation.
• We also remove lines without any Arabic characters.
• We then remove diacritics and kashida using CAMeL Tools.
• Finally, we split each line into sentences with a heuristics-based sentence segmenter.
• We train a WordPiece tokenizer on the entire dataset (167 GB text) with a vocabulary size of 30,000 using HuggingFace's tokenizers.
• We do not lowercase letters nor strip accents.

Pre-training

• The model was trained on a single cloud TPU (v3-8) for one million steps in total.
• The first 90,000 steps were trained with a batch size of 1,024 and the rest was trained with a batch size of 256.
• The sequence length was limited to 128 tokens for 90% of the steps and 512 for the remaining 10%.
• We use whole word masking and a duplicate factor of 10.
• We set max predictions per sequence to 20 for the dataset with max sequence length of 128 tokens and 80 for the dataset with max sequence length of 512 tokens.
• We use a random seed of 12345, masked language model probability of 0.15, and short sequence probability of 0.1.
• The optimizer used is Adam with a learning rate of 1e-4, \(\beta_{1} = 0.9\) and \(\beta_{2} = 0.999\), a weight decay of 0.01, learning rate warmup for 10,000 steps and linear decay of the learning rate after.

Evaluation results

• We evaluate our pre-trained language models on five NLP tasks: NER, POS tagging, sentiment analysis, dialect identification, and poetry classification.
• We fine-tune and evaluate the models using 12 dataset.
• We used Hugging Face's transformers to fine-tune our CAMeLBERT models.
• We used transformers v3.1.0 along with PyTorch v1.5.1.
• The fine-tuning was done by adding a fully connected linear layer to the last hidden state.
• We use \(F_{1}\) score as a metric for all tasks.
• Code used for fine-tuning is available here.

Results

Results (Average)

<a name="footnote-1">[1]</a>: Variant-wise-average refers to average over a group of tasks in the same language variant.

Acknowledgements

This research was supported with Cloud TPUs from Google’s TensorFlow Research Cloud (TFRC).

Citation

@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 mix 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.",
}