File size: 10,212 Bytes
89437c2
 
 
 
 
9706a73
89437c2
 
aed98c2
89437c2
 
 
aed98c2
 
89437c2
 
 
 
 
 
9706a73
89437c2
 
 
 
 
 
aed98c2
89437c2
 
 
 
 
 
 
 
 
 
 
9706a73
 
89437c2
 
 
9706a73
89437c2
 
9706a73
 
89437c2
 
9706a73
 
89437c2
 
9706a73
 
89437c2
 
9706a73
89437c2
 
aed98c2
 
89437c2
 
 
 
 
9706a73
89437c2
 
 
 
 
 
 
 
 
9706a73
89437c2
 
 
 
 
fd39d7b
1733b92
89437c2
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
---
language: 
- ar
license: apache-2.0
widget:
 - text: "الهدف من الحياة هو [MASK] ."
---

# 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.
The details are described in the paper *"[The Interplay of Variant, Size, and Task Type in Arabic Pre-trained Language Models](https://arxiv.org/abs/2103.06678)."*
We release eight models with different sizes and variants as follows:

||Model|Variant|Size|#Word|
|-|-|:-:|-:|-:|
||`bert-base-camelbert-mix`|CA,DA,MSA|167GB|17.3B|
||`bert-base-camelbert-ca`|CA|6GB|847M|
|✔|`bert-base-camelbert-da`|DA|54GB|5.8B|
||`bert-base-camelbert-msa`|MSA|107GB|12.6B|
||`bert-base-camelbert-msa-half`|MSA|53GB|6.3B|
||`bert-base-camelbert-msa-quarter`|MSA|27GB|3.1B|
||`bert-base-camelbert-msa-eighth`|MSA|14GB|1.6B|
||`bert-base-camelbert-msa-sixteenth`|MSA|6GB|746M|

This model card describes **CAMeLBERT-DA** (`bert-base-camelbert-da`), a model pre-trained on the DA dataset.

## 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](https://github.com/CAMeL-Lab/CAMeLBERT).

#### How to use
You can use this model directly with a pipeline for masked language modeling:
```python
>>> from transformers import pipeline
>>> unmasker = pipeline('fill-mask', model='CAMeL-Lab/bert-base-camelbert-da')
>>> unmasker("الهدف من الحياة هو [MASK] .")
[{'sequence': '[CLS] الهدف من الحياة هو.. [SEP]',
  'score': 0.062508225440979,
  'token': 18,
  'token_str': '.'},
 {'sequence': '[CLS] الهدف من الحياة هو الموت. [SEP]',
  'score': 0.033172328025102615,
  'token': 4295,
  'token_str': 'الموت'},
 {'sequence': '[CLS] الهدف من الحياة هو الحياة. [SEP]',
  'score': 0.029575437307357788,
  'token': 3696,
  'token_str': 'الحياة'},
 {'sequence': '[CLS] الهدف من الحياة هو الرحيل. [SEP]',
  'score': 0.02724040113389492,
  'token': 11449,
  'token_str': 'الرحيل'},
 {'sequence': '[CLS] الهدف من الحياة هو الحب. [SEP]',
  'score': 0.01564178802073002,
  'token': 3088,
  'token_str': 'الحب'}]
```

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

and in TensorFlow:
```python
from transformers import AutoTokenizer, TFAutoModel
tokenizer = AutoTokenizer.from_pretrained('CAMeL-Lab/bert-base-camelbert-da')
model = TFAutoModel.from_pretrained('CAMeL-Lab/bert-base-camelbert-da')
text = "مرحبا يا عالم."
encoded_input = tokenizer(text, return_tensors='tf')
output = model(encoded_input)
```

## Training data
- DA
  - A collection of dialectal Arabic data described in [our paper](https://arxiv.org/abs/2103.06678).

## Training procedure
We use [the original implementation](https://github.com/google-research/bert) 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](https://github.com/google-research/bert/blob/eedf5716ce1268e56f0a50264a88cafad334ac61/tokenization.py#L286-L297).
- We also remove lines without any Arabic characters.
- We then remove diacritics and kashida using [CAMeL Tools](https://github.com/CAMeL-Lab/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](https://github.com/huggingface/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](https://github.com/CAMeL-Lab/CAMeLBERT).

### Results

| Task                 | Dataset         | Variant | Mix   | CA    | DA    | MSA   | MSA-1/2 | MSA-1/4 | MSA-1/8 | MSA-1/16 |
| -------------------- | --------------- | ------- | ----- | ----- | ----- | ----- | ------- | ------- | ------- | -------- |
| NER                  | ANERcorp        | MSA     | 80.2% | 66.2% | 74.2% | 82.4% | 82.3%   | 82.0%   | 82.3%   | 80.5%    |
| POS                  | PATB (MSA)      | MSA     | 97.3% | 96.6% | 96.5% | 97.4% | 97.4%   | 97.4%   | 97.4%   | 97.4%    |
|                      | ARZTB (EGY)     | DA      | 90.1% | 88.6% | 89.4% | 90.8% | 90.3%   | 90.5%   | 90.5%   | 90.4%    |
|                      | Gumar (GLF)     | DA      | 97.3% | 96.5% | 97.0% | 97.1% | 97.0%   | 97.0%   | 97.1%   | 97.0%    |
| SA                   | ASTD            | MSA     | 76.3% | 69.4% | 74.6% | 76.9% | 76.0%   | 76.8%   | 76.7%   | 75.3%    |
|                      | ArSAS           | MSA     | 92.7% | 89.4% | 91.8% | 93.0% | 92.6%   | 92.5%   | 92.5%   | 92.3%    |
|                      | SemEval         | MSA     | 69.0% | 58.5% | 68.4% | 72.1% | 70.7%   | 72.8%   | 71.6%   | 71.2%    |
| DID                  | MADAR-26        | DA      | 62.9% | 61.9% | 61.8% | 62.6% | 62.0%   | 62.8%   | 62.0%   | 62.2%    |
|                      | MADAR-6         | DA      | 92.5% | 91.5% | 92.2% | 91.9% | 91.8%   | 92.2%   | 92.1%   | 92.0%    |
|                      | MADAR-Twitter-5 | MSA     | 75.7% | 71.4% | 74.2% | 77.6% | 78.5%   | 77.3%   | 77.7%   | 76.2%    |
|                      | NADI            | DA      | 24.7% | 17.3% | 20.1% | 24.9% | 24.6%   | 24.6%   | 24.9%   | 23.8%    |
| Poetry               | APCD            | CA      | 79.8% | 80.9% | 79.6% | 79.7% | 79.9%   | 80.0%   | 79.7%   | 79.8%    |

### Results (Average)
|                      | Variant | Mix   | CA    | DA    | MSA   | MSA-1/2 | MSA-1/4 | MSA-1/8 | MSA-1/16 |
| -------------------- | ------- | ----- | ----- | ----- | ----- | ------- | ------- | ------- | -------- |
| Variant-wise-average<sup>[[1]](#footnote-1)</sup> | MSA     | 81.9% | 75.3% | 79.9% | 83.2% | 82.9%   | 83.1%   | 83.0%   | 82.1%    |
|                      | DA      | 73.5% | 71.1% | 72.1% | 73.5% | 73.1%   | 73.4%   | 73.3%   | 73.1%    |
|                      | CA      | 79.8% | 80.9% | 79.6% | 79.7% | 79.9%   | 80.0%   | 79.7%   | 79.8%    |
| Macro-Average        | ALL     | 78.2% | 74.0% | 76.6% | 78.9% | 78.6%   | 78.8%   | 78.7%   | 78.2%    |

<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

```bibtex
@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.",
}
```