metadata

title: Story
emoji: 📈
colorFrom: yellow
colorTo: red
sdk: gradio
sdk_version: 5.18.0
app_file: app.py
pinned: false

Check out the configuration reference at https://huggingface.co/docs/hub/spaces-config-reference Here's a README template for your Gradio Space project that you can modify as needed:

AI Kids' Story Generator - English & Arabic 📖

Project Objectives

The objective of this project is to create a bilingual AI-powered story generator that produces engaging short stories for children in both English and Arabic. Users can input a character and a setting, and the AI will generate a fun, age-appropriate story.

Description of Implemented Pipelines

The project utilizes two main pipelines:

Text Generation Pipeline: This is based on the HF中国镜像站 Transformers library, utilizing models suitable for English and Arabic text generation:
- English: microsoft/Phi-3-mini-4k-instruct
- Arabic: ALLaM-AI/ALLaM-7B-Instruct-preview
Arabic Tashkeel Pipeline: The project employs the MohamedRashad/arabic-auto-tashkeel API to add diacritics (tashkeel) to Arabic text, enhancing readability for children.

Instructions for Using the Interface

Input Fields:
- Character Name: Enter a name for the character (e.g., "Benny the Bunny").
- Setting: Specify the setting for the adventure (e.g., "a magical forest").
- Choose Language: Select either "English" or "Arabic" from the dropdown menu.
Generate Story: Click the "Submit" button to generate the story.
Output: The generated story will appear in the output textbox.

Justifications for Model and Pipeline Choices

Model Selection: The chosen models are specifically fine-tuned for text generation tasks, making them suitable for crafting creative narratives. The English model is lightweight yet effective, while the Arabic model is designed to handle intricate language features.
Pipeline Selection: The use of the Arabic tashkeel API addresses the specific needs of Arabic-speaking children, ensuring that the generated stories are accessible and easy to read.

Bilingual Implementation

The project supports both English and Arabic, allowing for a diverse user base.