GPT-3 Swarm

This project provides a Python implementation for running multiple OpenAI's GPT-3 models in parallel, which is particularly useful for high-throughput applications. The implementation uses asyncio and ZeroMQ to communicate between the workers and the collector, and handles the rate-limiting restrictions imposed by OpenAI. The code provides a clean separation of concerns, with the swarm logic and the GPT-3 client being in different classes. The project also includes a sample client that demonstrates how to use the implementation to generate responses to multiple messages concurrently.

Features

• Supports running multiple GPT-3 models concurrently.
• Uses asyncio and ZeroMQ for efficient communication between the workers and the collector.
• Handles the rate-limiting restrictions imposed by OpenAI.
• Provides a clean separation of concerns between the swarm logic and the GPT-3 client.
• Includes a sample client that demonstrates how to use the implementation.

Technologies used

• Python 3.8
• OpenAI API
• asyncio
• ZeroMQ

Installation

• Clone the repository.
• Install the dependencies using pip.
• Set the OpenAI API key in the configuration file.
• Run the sample client to generate responses to multiple messages concurrently.

python -m venv env 
source env/bin/activate 
pip install -r requirements.txt
export OPENAI_API_KEY=sk-XXXXXXXXXXXXXXXXXXX; python main.py start-swarming

Contributing

Contributions are welcome! Please open an issue or pull request if you have any suggestions or improvements.

License

This project is licensed under the MIT License.