prompt2model - Generate Deployable Models from Instructions

Prompt2Model is a system that takes a natural language task description (like the prompts used for LLMs such as ChatGPT) to train a small special-purpose model that is conducive for deployment.

Quick Start

pip install prompt2model

Our current prompt2model implementation uses the OpenAI API. Accordingly, you need to:

Sign up on the OpenAI website and obtain an OpenAI API key.
Set the environment variable OPENAI_API_KEY to your API key by running the following command in your terminal:

export OPENAI_API_KEY=<your key>

You can then run

python cli_demo.py

to create a small model from a prompt, as shown in the demo video below. This script must be run on a device with an internet connection to access the OpenAI API. For best results, run this script on a device with a GPU for training your model.

Demo

promptmodel_demo.mp4

Tips and Examples to Write a Good Prompt

You can see the tips and examples to write a good prompt in prompt_examples.

Components

The prompt2model package is composed of several components, each designed to fulfill a specific purpose. To gain a comprehensive understanding of how to utilize each component effectively, please consult the readme.md file situated in the directory of the respective component. These files can be found at ./prompt2model/<component>/readme.md. They provide detailed information and instructions on customizing and maximizing the functionality of each component within the package.

Contribution

If you're interested in contributing to the prompt2model project, please

refer to CONTRIBUTING.md
open an issue or submit a PR
join us on discord
or reach out to @vijaytarian and @Chenan3_Zhao on Twitter

Cite

We have written a paper describing Prompt2Model in detail.

If you use Prompt2Model in your research, please cite our paper:

@misc{prompt2model,
      title={Prompt2Model: Generating Deployable Models from Natural Language Instructions},
      author={Vijay Viswanathan and Chenyang Zhao and Amanda Bertsch and Tongshuang Wu and Graham Neubig},
      year={2023},
      eprint={2308.12261},
      archivePrefix={arXiv},
      primaryClass={cs.CL}
}