File size: 2,014 Bytes
52ffcc4 d3536c0 52ffcc4 d3536c0 |
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 |
---
dataset_info:
features:
- name: text
dtype: string
splits:
- name: train
num_bytes: 1997024229
num_examples: 2138341
download_size: 942907181
dataset_size: 1997024229
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
language:
- ar
size_categories:
- 1M<n<10M
---
# Tashkeel Corpus
## Summary
The **Tashkeel-Corpus** is a dataset of Modern Standard Arabic (MSA) texts enriched with diacritics (Tashkeel). Diacritics are critical in Arabic script as they provide essential phonetic information that can disambiguate meanings and improve the accuracy of natural language processing (NLP) applications. This corpus serves as a valuable resource for tasks such as diacritization, speech synthesis, speech recognition, and Arabic language modeling.
The diacritics in this corpus were automatically added using the [BERTUnfactoredDisambiguator](https://arxiv.org/pdf/2211.16807) from the [CAMeL Tools](https://github.com/CAMeL-Lab/camel_tools) library.
## Supported Tasks and Leaderboards
* **Diacritization:** Enhance unvocalized Arabic text with appropriate diacritical marks.
* **Speech Synthesis (Text-to-Speech):** Improve pronunciation accuracy in TTS systems.
* **Speech Recognition (Automatic Speech Recognition):** Aid ASR systems in better understanding spoken Arabic.
* **Arabic Language Modeling:** Develop models that understand or generate fully vocalized Arabic text.
## Languages
* **Arabic** (ar)
## Dataset Structure
### Data Instances
Each data instance in the Tashkeel-Corpus is a segment of Arabic text with full diacritical annotation. An example entry might look like:
```
{
"text": "اَلْعَرَبِيَّةُ لُغَةٌ جَمِيلَةٌ"
}
```
### Data Fields
**text:** A string containing the fully diacritized Arabic text.
### Data Splits
The dataset is organized as a single collection of diacritized texts. Users may split the data into training, validation, and test sets according to their specific needs. |