metadata

task_categories:
  - image-to-text
language:
  - ar
pretty_name: ArabicConceptualCaptions3M
size_categories:
  - 1M<n<10M

Arabic Translated Conceptual Captions Dataset

Overview

This dataset consists of conceptual captions translated into Arabic using the Google Translate API. It serves as a resource for researchers and developers interested in exploring the vision-language tasks and biases introduced during the translation process.

Dataset Information

Source Dataset: Conceptual Captions
Translation Tool: Google Translate API
Translation Language: English to Arabic

Important Notes

Translation Quality: The translations are machine-generated and may contain errors, inaccuracies, or cultural nuances that were not appropriately captured. Researchers are encouraged to verify translations for accuracy.
Biases: The dataset is prone to various types of biases, including but not limited to gender bias. Google Translate API, like any other machine translation tool, may inadvertently introduce biases present in its training data.
Usage Guidelines: Please refer to the original Conceptual Captions dataset usage guidelines, as they apply to this translated version. Respect copyright and licensing agreements associated with the source dataset.

Description

the dataset contains two file train and validation and each has:

arabic_caption: This column includes the machine-translated captions in Arabic, generated using the Google Translate API.
caption: This column contains the original captions in English sourced from the Conceptual Captions dataset.
link: This column contains links to images corresponding to the captions in the dataset.