ismailelsayedeltanja/Egyptian-dialect-8k
Egyptian Dialect Speech Dataset (8k) Dataset Description This dataset contains around 8,000 cleaned speech samples in the Egyptian Arabic dialect. It is optimized for speech recognition (ASR), text-to-speech (TTS), and audio processing tasks tailored specifically for Egyptian dialect applications. Primary Language: Egyptian Arabic (ar-EG) Total Samples: ~8,000 audio clips with corresponding text transcriptions. Source: Cleaned and processed from… See the full description on the dataset page: https://huggingface.co/datasets/ismailelsayedeltanja/Egyptian-dialect-8k.
language:
- ar license: mit task_categories:
- automatic-speech-recognition
- text-to-speech
- audio-classification tags:
- arabic
- egyptian-dialect
- audio
- speech size_categories:
- 1K<n<10K ---
Egyptian Dialect Speech Dataset (8k)
Dataset Description
This dataset contains around 8,000 cleaned speech samples in the Egyptian Arabic dialect. It is optimized for speech recognition (ASR), text-to-speech (TTS), and audio processing tasks tailored specifically for Egyptian dialect applications.
- Primary Language: Egyptian Arabic (
ar-EG) - Total Samples: ~8,000 audio clips with corresponding text transcriptions.
- Source: Cleaned and processed from
oddadmix/da7ee7_sep_cleaned-8k.
Quick Start & Usage
Using the datasets library (Recommended)
You can load this dataset with a single line of code using Hugging Face's datasets library:
from datasets import load_dataset
# Load the full dataset
import pandas as pd
from datasets import load_dataset, Audio
from IPython.display import display, HTML, Audio as AudioPlayer
OUTPUT_DIR = "/gdrive/MyDrive/TTS-Spark"
dataset = load_dataset("ismailelsayedeltanja/Egyptian-dialect-8k", split="train")
dataset = dataset.cast_column("audio", Audio(sampling_rate=16000))
dataset = dataset.filter(
lambda x: bool(x.get("text")) and x.get("audio") is not None)
shuffled_dataset = dataset.shuffle(seed=101)
dataset = shuffled_dataset.take(430)
print(f"Data_: \n {dataset}")
display(HTML("<h2 style='color: #4f46e5; font-family: sans-serif; text-align: right;'> Showing_Sample_from_Raw_data </h2>"))
html_style = """
<style>
.audio-table {
border-collapse: separate;border-spacing: 0;margin: 15px 0;font-size: 1em;
font-family: 'Segoe UI', Tahoma, Geneva, Verdana, sans-serif;width: 100%;
direction: rtl;text-align: right;box-shadow: 0 4px 12px rgba(79, 70, 229, 0.12);
border-radius: 12px;overflow: hidden;border: 1px solid #e0e7ff;
}
.audio-table th {
background: linear-gradient(135deg, #6366f1 0%, #4f46e5 100%);color: #ffffff;
padding: 14px 18px;font-weight: 600;letter-spacing: 0.5px;
}
.audio-table td {padding: 12px 18px;border-bottom: 1px solid #f1f5f9;color: #1e293b;
background-color: #ffffff;
}
.audio-table tr:hover td {background-color: #f5f3ff;color: #4338ca;
}
.sample-tag {background-color: #e0e7ff;color: #3730a3;
padding: 4px 10px;border-radius: 6px;font-weight: bold;
}
</style>
"""
display(HTML(html_style))
for i in range(1):
sample = dataset[i]
text = sample["text"]
audio_data = sample["audio"]["array"]
sr = sample["audio"]["sampling_rate"]
display(HTML(f"""
<table class="audio-table">
<tr>
<th style="width: 15%;">Sample</th>
<th style="width: 85%;">Texe</th>
</tr>
<tr>
<td><span class="sample-tag">#{i+1}</span></td>
<td style="font-size: 1.05em; font-weight: 500;">{text}</td>
</tr>
</table>
"""))
display(AudioPlayer(audio_data, rate=sr))
## Stream the Dataset (Without Downloading Entirely)
from datasets import load_dataset
# Stream dataset
dataset = load_dataset("ismailelsayedeltanja/Egyptian-dialect-8k", streaming=True)
for sample in dataset['train']:
print(sample['text']) # Replace 'text' with the actual text column name if different
break
## Data Structure
Each entry in the dataset typically contains:
audio: Audio object containing array (waveform), path, and sampling_rate.
text (or transcription): The corresponding text script in Egyptian Arabic.
## Citation & Acknowledgments
If you use this dataset in your research or project, please credit the original source:
Source repository: oddadmix/da7ee7_sep_cleaned-8k
Re-uploaded and formatted by: @ismailelsayedeltanja