This tool leverages the capabilities of OpenAI's GPT model to automate hardware design based on user prompts. It's an easy-to-use tool that requires minimal inputs from the user.
The tool works by directing the AI to through a waterfall design procedure based on a set of goals and requirements. Follow this link for an updated version of the accompanying paper: https://arxiv.org/abs/2405.16072
Ensure you have Python 3.10 installed and the pip package manager.
Before running the project, you'll need to have,
- OpenAI API key (purchage/get one from here: https://platform.openai.com/). Make sure to set this up, as it is a crucial part of the project.
- Tavily search API (get a free one or purchase here: https://app.tavily.com/home)
-
Clone the repository:
-
Navigate to the project directory:
cd FPGA_AGI -
Install the necessary requirements using pip:
pip install -r requirements.txt
Disclaimer: ** The Agents will be able to autonomously generate and run python code on your machine. You are responsible for sandboxing this tool. Run this at your own risk.**
- For a quick and dirty intro follow the sample jupyter notebook, Experiments.ipynb, given in the repo.
This software is licensed under the GNU General Public License (GPL), version 3 (GPL-3.0). For the full text of the license, see the LICENSE file in this repository.
For more information on GPL-3.0, visit: https://www.gnu.org/licenses/gpl-3.0.en.html
If you use this for any evaluation or wrote any article that levrages this tool, please cite me: https://arxiv.org/abs/2405.16072 Here is the bibtext version for your convenience:
@misc{sheikholeslam2024synthai,
title={SynthAI: A Multi Agent Generative AI Framework for Automated Modular HLS Design Generation},
author={Seyed Arash Sheikholeslam and Andre Ivanov},
year={2024},
eprint={2405.16072},
archivePrefix={arXiv},
primaryClass={cs.AI}
}
Please note that this project is currently under active development. Feedback and contributions are always welcome!