为了满足大家对大语言模型Agent（代理/智能体），RAG（检索增强生成），Tools box(工具包）以及模型Fine-tuning微调技术的需求， Research Triangle AI Society (RTAI) 特邀北卡州立大学 Professor Paul Liu 给大家两天从零基础入门到精通的强化训练。

内容：
Day-1: Using LLMs via APIs and Locally

1: Introduction to Large Language Models
 Overview of popular LLMs and platforms (Openai, Llama, Gemini, Claude, Mistral):
 Building a real-time data backed ChatBot using LLM function call and APIs.
2: Running LLMs on Your Local Laptop/PC
 Setting up and running Ollama.
 Using the local Ollama API for advanced open source LLM projects.
3: Implementing Your LLM App on the Web
 Case study using Streamlit (TXR, CSV, PDF+LLM).
 Case study using Flask, Gunicorn, and Nginx.
4: Creating and Using Retrieval-Augmented Generation (RAG)
 Introduce Tokenization, Embedding, Vector DB and Query
 How to use Llama-idex and Langchain for multiple document searching

Day2:  Fine-Tuning and Creating LLMs Using Your Own Data

5. Fine-tuning GPT-3.5 and 4o with your own data.
 Data preparation and cleaning
 Steps and techniques for fine-tuning models.
6. Using QLoRA and PEFT to fine-tune Llama2/3，Phi3,  and Mixtral.
 Introduction to QLoRA and PEFT.
 Practical project: Fine-tune Llama3.2 text and vision models.
7: Building an Auto-Coding Agent,
 Auto Python and SQL for querying and displaying large data frames.
 Introduction to writing and scheduling automated scripts.
 Practical project: Build an auto-coding agent to query and display your data.
8: Creating and Building a Complete New Small Language Model GPT2 from Scratch

时间：
1) Sat. Dec 7th,  9:00-12:00 (半天）
2) Sat. Dec 14th, 9:00-12:00,  1:00-4:00pm （全天）
3) Sat. Dec 21th, 1:00pm-4:00pm （半天）

学费: $300/person
上课方式： 线下面对面为主 （特殊情况可以申请线上）
上课语言：中文为主，英文为辅
地点： 待定

要求： 1）有基本的编程经验；2）乐意花时间，花代价进入大语言模型应用领域； 3）完成每周的作业及Projects.

课程注册地址： https://forms.gle/zRGvohN4uRY8wNhF8