Multi-Author Story Generation

This project generates a collaborative story using multiple language models (LLMs) in sequence. Each LLM generates a part of the story, building upon the previous parts.

Table of Contents

• Introduction
• Features
• Installation
• Usage
• Contributing
• Contact

Introduction

This project aims to create collaborative stories by leveraging various LLMs in a sequential manner. Each LLM contributes to the story by adding a new part, based on the previous parts generated by other models. The story-generation process includes summarization and continuation prompts to ensure coherence and creativity.

Features

• Multi-author story generation using different LLMs
• Dynamic prompt generation for story continuation
• Summarization of the story so far for better context
• Batch processing of data for efficiency
• Customizable parameters for different use cases

Installation

1. Clone the repository:

   git clone https://github.com/.git  ## Not linked to github yet for anonymity

2. Navigate to the project directory:

   cd repo-folder

3. Install the required dependencies:

   pip install -r requirements.txt

4. Set your HuggingFace token as an environment variable:

   export HF_TOKEN="ENTER YOUR HUGGINGFACE TOKEN HERE"

Usage

To run the story generation script, use the following example command:

python write_middle_parts.py --author_num 1 --start 0 --end 100 --n 10 --total_authors 5 --llm llama

• author_num: The order of this LLM or part of the story being written.
• start: Prompt start index (default is 0).
• end: Prompt end index (default is -1, which means till the end).
• n: Batch size.
• total_authors: Total number of authors contributing to the story.
• llm: The LLM being used to generate the story (e.g., llama, mistral,gemma, olmo, orca).

Contributing

Contributions are welcome! Please follow these steps to contribute:

1. Fork the repository.
2. Create a new branch (git checkout -b feature-branch).
3. Commit your changes (git commit -am 'Add new feature').
4. Push to the branch (git push origin feature-branch).
5. Create a new Pull Request.

Contact

For any questions or issues, please contact:

• Email: removed for anonymity