This repository provides learning notes and practical examples for ChatGPT prompt engineering, targeting developers who want to leverage large language models (LLMs). It aims to demystify LLM capabilities and offer best practices for crafting effective prompts to build powerful applications.

How It Works

The project is structured around the DeepLearning.AI "ChatGPT Prompt Engineering for Developers" course. It translates course content, summarizes key concepts using ChatGPT, and provides runnable Jupyter Notebooks with OpenAI API integration. The notes emphasize systematic prompt design, iterative refinement, and application across various tasks like summarization, inference, transformation, and expansion.

Quick Start & Requirements

• Install/Run: Clone the repository and run Jupyter Notebooks locally. CLI scripts are also provided: python source/cli/cli_py.py --prompt "hello chatgpt" --model "gpt-3.5"
• Prerequisites: OpenAI API Key, Python.
• Resources: Access to OpenAI API.
• Docs: https://islinxu.github.io/prompt-engineering-note/

Highlighted Details

• Comprehensive notes covering prompt engineering principles and applications.
• Runnable Jupyter Notebooks for hands-on practice with OpenAI API.
• CLI scripts for direct interaction with LLMs.
• Curated list of "awesome" prompt engineering projects.
• Includes a "tiny-gpt" toy project for building a simple GPT model from scratch.

Maintenance & Community

The project is a personal learning initiative, with ongoing updates noted for CLI scripts and project lists. Links to external resources like OpenAI's cookbook and DeepLearning.AI's course are provided.

Licensing & Compatibility

The repository does not explicitly state a license. The content is derived from a course, and usage of OpenAI API requires adherence to OpenAI's terms of service.

Limitations & Caveats

This is a personal collection of learning notes and may not be exhaustive. The project relies on external OpenAI API access, and the quality of generated content is dependent on the prompts used and the underlying LLM.