Ecosyste.ms: Awesome
An open API service indexing awesome lists of open source software.
https://github.com/aazizisoufiane/llm-auto-labeler
An automated text labeling tool that employs multiple Language Model Mediators (LLMs) to assess, label, and justify text data, reducing the need for time-consuming human intervention
https://github.com/aazizisoufiane/llm-auto-labeler
Last synced: 3 months ago
JSON representation
An automated text labeling tool that employs multiple Language Model Mediators (LLMs) to assess, label, and justify text data, reducing the need for time-consuming human intervention
- Host: GitHub
- URL: https://github.com/aazizisoufiane/llm-auto-labeler
- Owner: aazizisoufiane
- Created: 2023-09-27T10:21:37.000Z (about 1 year ago)
- Default Branch: main
- Last Pushed: 2023-10-17T14:01:30.000Z (about 1 year ago)
- Last Synced: 2024-06-11T02:04:16.393Z (5 months ago)
- Language: Python
- Size: 21.5 KB
- Stars: 5
- Watchers: 1
- Forks: 0
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
Awesome Lists containing this project
README
# LLM-Auto-Labeler
[](https://medium.com/@aazizi.soufiane)
[](https://github.com/aazizisoufiane)
[](https://www.linkedin.com/in/soufiane-aazizi-phd-a502829/)
[](https://opensource.org/licenses/MIT)
## Table of Contents
- [Overview](#overview)
- [Features](#features)
- [Getting Started](#getting-started)
- [Prerequisites](#prerequisites)
- [Installation](#installation)
- [API Configuration and Environmental Variables](#api-configuration-and-environmental-variables)
- [Project Structure](#project-structure)
- [License](#license)
## Overview
LLM-Auto-Labeler is an automated text labeling tool designed to reduce the need for time-consuming human intervention.
It leverages multiple Language Model Mediators (LLMs) to assess, label, and provide justifications for text data.
## Features
- **Multi-Model Consensus**: Ensures robust labeling by employing multiple LLMs for the initial assessment.
- **Explanations**: Increases transparency by having each LLM provide justifications for its chosen label.
- **Mediation**: Utilizes a mediator LLM to make the final labeling decision, thereby reconciling any discrepancies
between different LLMs.
## Getting Started
### Prerequisites
To run this project, you'll need:
- Python 3.7 or higher
- Langchain library
- Pydantic library
- Your preferred LLM libraries (e.g., OpenAI's GPT, etc.)
### Installation
Clone the repository and install the necessary packages:
```bash
git clone https://github.com/yourusername/llm-auto-labeler.git
cd llm-auto-labeler
pip install -r requirements.txt
python src/main.py --description_file "input/descriptions"
```
### API Configuration and Environmental Variables
To run LLM-Auto-Labeler, you'll need access to specific Language Model Mediators (LLMs) APIs. For example, if you are
using OpenAI's GPT-based models and Replicate models, you'll need to set up the following API keys:
OPENAI_API_KEY for OpenAI's GPT-based models.
REPLICATE_API_TOKEN for Replicate models.
Place these API keys in a .env file located at the root directory of the project. The .env file should look like this:
```textmate
OPENAI_API_KEY=your_openai_api_key_here
REPLICATE_API_TOKEN=your_replicate_api_token_here
````
However, the modular architecture of LLM-Auto-Labeler allows you to use LLMs of your choice. If you want to integrate
other LLMs, make sure to read their respective documentation to obtain and configure additional API keys or access
tokens.
## Project Structure
```textmate
├── README.md
├── input
│ └── descriptions
├── logs
├── output
│ └── output.json
├── requirements.txt
└── src
├── __init__.py
├── autolabel.py
├── config.py
├── llm_handler.py
├── logger.py
├── main.py
├── schemas.py
└── utils.py
```
## License
Feel free to edit the placeholders like the documentation link, contributor's guide, or other specific details to fit
your project's needs.