Cerberus

Cerberus is a modular, containerized AI solution offering APIs for sentiment analysis, text summarization, and text generation. Built on top of fine-tuned and quantized Hugging Face models, Cerberus is designed for scalability, efficiency, and easy integration into workflows.

---

Features

• Sentiment Analysis: Quickly classify text sentiment as positive, negative, or neutral.
• Summarization: Generate concise and meaningful summaries for long text.
• Text Generation: Produce creative or structured text based on input prompts.
• Scalable Architecture: Deployed using containerized services for easy scaling.
• Optimized Models: Uses fine-tuned and quantized Hugging Face models for speed and accuracy.
• OpenAPI Documentation: Built-in API documentation for developers.
• Extensible: Easily add new AI models and services.

---

Technologies Used

• Programming Language: Python 3.10+
• Frameworks: FastAPI for API development
• AI Models: Hugging Face (Transformers, Optimum, Accelerate)
• Machine Learning Backend: PyTorch with MPS support for macOS (Apple Silicon)
• Data Handling: Datasets library by Hugging Face
• Deployment:
  • Docker for containerization
  • Docker Compose for service orchestration

---

Table of Contents

• Cerberus
  • Features
  • Technologies Used
  • Table of Contents
  • Getting Started
    • Prerequisites
    • Installation
  • Usage
    • Endpoints
    • Example Requests
  • API Documentation
  • Development
    • Run Locally Without Docker
    • Test the Models
  • Contributing
  • License

---

Getting Started

Prerequisites

• Python 3.10 or higher

• Poetry for dependency management:

  curl -sSL https://install.python-poetry.org | python3 -

• Docker and Docker Compose installed: Docker Installation Guide

Installation

1. Clone the Repository:

   git clone https://github.com/BonnardValentin/cerberus
   cd cerberus

2. Install Dependencies with Poetry:

   poetry install

3. Set Up Environment Variables: Create a .env file in the project root with the following:

   MODEL_PATH="./models/quantized"
   API_KEY="your-secret-key"
   

4. Start the Application: Using Docker Compose:

   docker-compose up --build

5. Test the Application: Access the API at http://localhost:8000.

---

Usage

Endpoints

• Sentiment Analysis: /analyze-sentiment/
• Summarization: /summarize/
• Text Generation: /generate-text/

Example Requests

Sentiment Analysis

curl -X POST http://localhost:8000/analyze-sentiment/ \
-H "Content-Type: application/json" \
-d '{"text": "I love Cerberus!"}'

Summarization

curl -X POST http://localhost:8000/summarize/ \
-H "Content-Type: application/json" \
-d '{"text": "Docker simplifies containerization by providing a platform to develop, ship, and run applications in isolated environments."}'

Text Generation

curl -X POST http://localhost:8000/generate-text/ \
-H "Content-Type: application/json" \
-d '{"text": "Once upon a time, in a faraway land,"}'

---

API Documentation

Cerberus provides an interactive OpenAPI UI:

• Visit http://localhost:8000/docs for Swagger UI.
• Visit http://localhost:8000/redoc for ReDoc.

---

Development

Run Locally Without Docker

1. Activate the Poetry environment:

   poetry env activate

2. Start the FastAPI server:

   uvicorn app.main:app --reload

Test the Models

Run a Python script to verify that models are loaded correctly:

from transformers import pipeline

model = pipeline("sentiment-analysis", model="./models/quantized")
print(model("I love open-source!"))

---

Contributing

We welcome contributions to Cerberus! To contribute:

1. Fork the repository.

2. Create a feature branch:

   git checkout -b feature-name

3. Commit your changes:

   git commit -m "Add new feature"

4. Push to your fork:

   git push origin feature-name

5. Open a pull request.

Please read our Contributing Guide for more details.

---

License

This project is licensed under the MIT License - see the LICENSE file for details.