| --- |
| configs: |
| - config_name: default |
| data_files: |
| - split: train |
| path: data/train-* |
| dataset_info: |
| features: |
| - name: audio |
| dtype: audio |
| - name: transcription |
| dtype: string |
| splits: |
| - name: train |
| num_bytes: 6486606662.193651 |
| num_examples: 8874 |
| download_size: 5389628614 |
| dataset_size: 6486606662.193651 |
| --- |
| # Arabic-Diacritized-TTS Dataset |
|
|
| ## Overview |
| The **Arabic-Diacritized-TTS** dataset contains Arabic audio samples and their corresponding text with full diacritization. This dataset is designed to support research in Arabic speech processing, text-to-speech (TTS) synthesis, automatic diacritization, and other natural language processing (NLP) tasks. |
|
|
| ## Dataset Contents |
| - **Audio Samples**: High-quality Arabic speech recordings. |
| - **Text Transcriptions**: Fully diacritized Arabic text aligned with the audio. |
|
|
| ## Data Generation |
| The dataset was generated using the **TTS Arabic** model from the following repository: |
|
|
| [https://github.com/nipponjo/tts_arabic](https://github.com/nipponjo/tts_arabic) |
|
|
| ### About the Model |
| The **TTS Arabic** model is an advanced text-to-speech system for Arabic, capable of generating high-quality, natural-sounding speech with full diacritization. It is built on deep learning techniques and trained on diverse Arabic text and speech datasets. The model: |
| - Supports Modern Standard Arabic (MSA). |
| - Includes proper diacritic placement to enhance pronunciation accuracy. |
| - Is optimized for high-fidelity and intelligibility in Arabic speech synthesis. |
|
|
| ## Usage |
| This dataset can be used for: |
| - Training and evaluating Arabic **ASR (Automatic Speech Recognition)** models. |
| - Enhancing **TTS (Text-to-Speech)** systems with diacritized text. |
| - Improving **Arabic diacritization** models. |
| - Conducting **linguistic research** on Arabic phonetics and prosody. |
|
|
| ## Citation |
| If you use this dataset, please cite the source and credit the contributors. |
|
|
| --- |
|
|
| For any inquiries or contributions, feel free to reach out! |
|
|
