NOTE: This article is a sub-project of Open-Source-AI-Research, an experimental non-profit project focusing on AI research in an open source mode.
A few days ago, my supervisor asked me to write a document for some fellow grad students on an introduction of LLMs (Large Language Models) and how to simply apply them into researches. At first, I decided to delay without any interest and motivation. However, one day I suddenly thought that, since I believe actually LLMs could be considered as a process pipeline for almost any task in any field**,** maybe it is a better idea to write such an article for anyone who is major in another field without much knowledge on LLMs. It is indeed the real motivation for me. Thus, the title of the passage is LLM4PP (Large Language Models for a Process Pipeline ). So, let's start...
P.S. : Since I am a complete noob instead of a master, because of potential mistakes, this article is for reference only. I would appreciate it if you are willing to help by providing feedbacks/suggestions to bebetterest+at+outlook.com . In addition, you are welcome to contribute.
First of all, please read the contribution terms carefully in Open-Source-AI-Research.
Second, fork the corresponding repository, make some improvements, add a record in Update Logs and just pull a request, which also means that you have already accepted the terms by default.
Thanks for your contribution!!!
DONE: initialized the first version of LLM4PP.
TODO: polish the article; more examples on language/feature interface; more details on the accesses...
Please cite the article if your work is motivated from it.
@misc{LLM4PP-QuickStart,
title = {LLM4PP-QuickStart},
author = {O.S.R.},
howpublished = {\url{https://llm4pp-quickstart.gitbook.io/llm4pp-quickstart/}},
}