Welcome to the LLM Design Patterns repository! This is your go-to resource for a collection of design patterns aimed at helping you harness the full potential of large language models (LLMs).
Our goal is to provide design patterns that are platform and language-agnostic. However, as Python is the predominant language in the machine learning ecosystem, you'll find that most of our patterns can be readily implemented using Python. All sample codes are written in Python, utilizing its rich ecosystem of libraries.
We hope that this repository serves as a valuable resource for both beginners and experienced practitioners alike, empowering you to make the most of LLMs in your applications. We invite you to explore, experiment, and contribute to this ever-growing collection of design patterns. Happy coding!
Please refer below for ease of reading: GitHub pages link for this repo
The patterns aren't classified into logical groups yet. We are working on it.
- Simple LLM Generation
- Retrieval Augmented Generation
- [TODO] Example Augmented Task Generation
- [TODO] Autonomous Agent
- [TODO] LLM Tool
- [TODO] Guard rails
- [TODO] Simple LLM Embeddings
- [TODO] LLM Memory
- [TODO] Reflection and Planning pattern
For adding a new pattern, please use pattern_template.md as a template, and raise an issue and
followed by a PR.
To request for a new pattern, please raise an issue.
To contribute to an existing pattern, please raise a PR.
Copyright (c) 2023 Krishna Gogineni
Licensed under the MIT License.