Graduation-project

Automated Content Generation System

Team Members:

1. Mohammed Saieed Abdelnaby
2. Gehad Abdallah Mohammed
3. Abdelrahman Mamduh Eldwiny
4. Ayat Abdelnaby Mohammed Hussain
5. Mohammed Saeed Fath-elbab
6. Mahmoud Taha Abdelmaboud Behiery

Project Overview

Our graduation project focuses on simplifying and enhancing content creation using AI technology. We have developed two core applications:

1. Content Flow – A sophisticated AI Writing Assistant built on Llama 3.2 3B, enabling users to generate custom content through prompt engineering. The application supports customizable text generation, catering to a variety of writing tasks such as research papers, essays, cover letters, and more.

2. Inspire AI – A prompt optimizer designed to improve AI image model outputs. By refining user inputs, it enhances the quality of AI-generated images for models like Stable Diffusion and Flux.

Why It Matters

We aim to make AI technologies more accessible and customizable for users, bridging the gap between technical advancements and creative processes. Our tools help users generate high-quality written content and optimized prompts for AI-based image generation, reducing the complexity of these tasks.

Methodology

• Content Flow: Utilizes Llama 3.2 3B to assist with natural language generation through prompt engineering. Users input desired tone, style, and content type, while the system generates tailored text.

• Inspire AI: Leverages NLP techniques to optimize prompts for image models. It offers improved image quality and accuracy through better prompt formulation.

Key Features

• Custom AI Writing Assistant: Adjust tone, style, and structure for personalized content.
• AI Prompt Optimizer: Enhances AI-generated images by refining user inputs and prompts.

Technologies Used

• AI Writing Assistant: Python, Hugging Face, Llama models.
• Prompt Optimizer: NLP frameworks, Llama model, Llava API integrations.

Challenges & Results

• Challenges: Ensuring model accuracy and managing computational resources for real-time optimizations.
• Results: Faster content creation and improved AI-generated image quality, enhancing user experience in both applications.

Acknowledgments

We would like to thank our supervisor, Engineer Ehab Ibrahim, and the Ministry of Communication for their support throughout this project.