Ecosyste.ms: Awesome

An open API service indexing awesome lists of open source software.

https://github.com/argonne-lcf/llm-workshop

Last synced: 27 days ago
JSON representation

Host: GitHub
URL: https://github.com/argonne-lcf/llm-workshop
Owner: argonne-lcf
Created: 2024-02-01T00:55:54.000Z (5 months ago)
Default Branch: main
Last Pushed: 2024-02-24T02:48:47.000Z (4 months ago)
Last Synced: 2024-04-16T03:44:52.178Z (2 months ago)
Language: Jupyter Notebook
Size: 43.9 MB
Stars: 14
Watchers: 7
Forks: 12
Open Issues: 0
Metadata Files:
- Readme: README.md

Lists

awesome-stars - argonne-lcf/llm-workshop - (Jupyter Notebook)

README

        # Large Language Models (LLMs): Tutorial Workshop

[Workshop Agenda](https://anl.app.box.com/file/1421615910690?s=woqtpw0o0tpnb6j9uljjme5wqmxsoz35)  

Argonne National Laboratory  

_February 12th and 13th, 2024_  

**Building 240**, **Room 1501** (in-person)

This repository contains the materials used in the LLM Tutorial Workshop, February 12th and 13th, 2024.

The  workshop material will rely on Jupyter Notebooks which are targeted for running on [Google's Colaboratory Platform](https://colab.research.google.com).  

    Tutorials

    


        LLMs 101

        Prompt Engineering

        Retrieval Augmented Generation (RAG)

        Fine-Tuning an Existing LLM

        LLMs from Scratch

    


## Before You Arrive ✅

> [!IMPORTANT]  

> Complete the following steps **BEFORE** you come to the tutorial

> 

> - [ ] [Google Colab](https://colab.research.google.com) **Setup Account**

> 

>     Google Colab Instructions

> 

>     The Colab platform gives the user a virtual machine in which to run Python codes including machine

>     learning codes.

> 

>     The VM comes with a preinstalled environment that includes most of what is needed

>     for these tutorials.

> 

>     * You need a Google Account to use Colaboratory

>     * Go to [Google's Colaboratory Platform](https://colab.research.google.com) and sign in with

>       your Google account

>     * You should see this page

>         ![start_page](./assets/colab_start_page_new.png)

>     * Click on the `New Notebook` at the bottom

>     * Now you will see a new notebook where you can type in python code.

>         ![clean_page](./assets/colab_start_page1.png)

>     * After you enter code, type ` + ` to execute the code cell.

>     * A full introduction to the notebook environment is out of scope for this tutorial, but many

>       can be found with a [simple Google

>       search](https://www.google.com/search?q=jupyter+notebook+tutorial)

>     * We will be using notebooks from this repository during the tutorial, so  you should be

>       familiar with how to import them into Colaboratory

>     * Now you can open the `File` menu at the top left and select `Open Notebook` which will open a

>       dialogue box.

>     * Select the `GitHub` tab in the dialogue box.

>     * From here you can enter the url for the github repo: `https://github.com/brettin/llm_tutorial`

>       and hit ``.

>         ![open_github](./assets/colab_open_github_1.png)

>     * This will show you a list of the Notebooks available in the repo.

>     * Select the `introduction.ipynb` file to open and work through it.

>     * As each session of the tutorial begins, you will simply select the corresponding notebook from

>       this list and it will create a copy for you in your Colaboratory account (all `*.ipynb` files in

>       the Colaboratory account will be stored in your Google Drive).

>     * To use a TPU, in the notbook the select `Runtime` -> `Change Runtime Type` and you have a

>       dropbox list of hardward settings to choose from where the notebook can run.

> 

>     

> 

> 

> - [ ] 🤗 [Hugging Face](https://huggingface.co): **Account and Access Token**

> 

>     Hugging Face Instructions

> 

>     - Sign up for a huggingface account and obtain an access token: https://huggingface.co

>     - Sign Up (top bar)

>       Log into huggingface and get an access token:

>         - Login -> Settings (left pane) -> Access Tokens (left pane) -> New token (center pane)

> 

>     

> 

> - [ ] 🦙 [Request access](https://ai.meta.com/resources/models-and-libraries/llama-downloads/) to Llama-2 model

> 

>     Llama-2 Access Instructions

> 

>     - Visit this https://huggingface.co/meta-llama/Llama-2-7b-hf and request access to the model

>     - vist meta website and accept the terms https://ai.meta.com/resources/models-and-libraries/llama-downloads/

>     - Note: Your Hugging Face account email address MUST match the email you provide on the Meta website, or your request will not be approved.

> 

>