SLMLAH/P2_Dataset_Collection

P2 Dataset Collection

Dataset Summary

This is a small-size synthesised Table dataset in Arabic.

Supported Tasks and Leaderboards

• Tasks: table Understanding in Arabic
• Applicable Use Cases: [Any table related training for tables in Arabic]

Languages

• Tables are in arabic and the captions as well.

Dataset Structure

DatasetDict({
    'train': Dataset({
        features: ['image_id', 'caption', 'image'],
        num_rows: [Your number of rows]
    })
})

Data Instances

Example of a data instance:

{
    'image_id': '1.png',
    'caption': 'سجل إجازات الموظفين السنوي يوضح عدد الأيام المستحقة، المأخوذة، والمتبقية لكل موظف، مما يعكس إدارة دقيقة للإجازات.',
    'image': 1.png]
}

Data Fields

• image_id: Unique identifier for each image
• caption: Text description of the image
• image: Binary data of the image

Dataset Creation

Source Data

[The data is AI-Generated using GPT-4o and Claude3.5]

Annotations

[AI-Generated, then manually curated.]

Additional Information

Dataset Curators

[Data for IVAL/MBZUAI]