GPT Controller

This python project is meant to integrate chatGPT in the process of controlling a robot in order to achieve zero-shot planning and control. The project is part of my bachelor's thesis.

Prerequisites

• Python 3.10
• poetry package manager (pip install poetry)

Installation

1. Clone the repository:

git clone https://github.com/andrei-calin-dragomir/gpt-controller.git

2. Navigate to the project directory:

cd gpt-controller

3. Install project dependencies:

poetry install

4. Configure project:

   Open the config.py file in the root directory and fill all the entries with the required information.

Usage

To run the project, first, run the poetry shell using the following command:

poetry shell

Then execute the following command:

poetry run python -m run.py

Upon running, you will be prompted to enter a message. After entering a message, the program will generate a response and print it to the console. You have access to utility functions in order to get a view on the knowledge of the system as well as other information about the system. To access them you can type --help in the console and you will be presented with a list of available commands.

Evaluation

The evaluation of the system's performance is done within the manual.ipynb notebook. The notebook contains a set of tests that can be run in order to evaluate the system's prompting accuracy, conversation length and other metrics.