llama2-finetune

LLAMA-FIT CO.™ - Your LLaMA Will Like The Way It Looks

Cracking the LLaMA2 Interview: A Weekend Guide to Impressing AI Startups

Introduction

Goal is to finetune a LLaMA2 so that it can impress during a Tech Interview
business model
- I got suited up llama interns waiting to work for you
operation model
- read the doc (TODO: make a doc); the joke here is that it's basically the code and I have no moat
street cred maxxing
impress employers/vc/people/ai
benchmark llama2 fine-tune against base model
- split training set against validation set (90/10?)
- benchmark fine-tune against base
  - goal: achieve higher "cosine similarity" (between llama output and validation set) using fine-tuned version than base version
  - (exclaim - oh shit! - I'm a regmonkey)

Quick Start Guide

Haphazard Benchmark

run all cells in notebook
last cell can be ran to view simple haphazard benchmark

Basic Research

LLaMA2 Paper
- definitive primary source
LLaMA2 Github
- actual code
- not to be confused with llamacode library
  - (aside) spent way to much time trying to figure out why the download code wasn't working ...
- 7b model
Deepgram Video Analysis
- good take aimed for layman
Karparthy is a Beast
- look it's clear that I'm not legit until I make it onto the README.md page
- but as a weekend project ... nah ... (at least right now)
Fiverr
- there's academic integrity which I adhere to
- but I'm also technically an entrepreneur ... had the look
  - these rates seems decent
  - side hustle oppurtunity?
Random Paper
- literally a random paper I pulled from arxiv talking about training and finetunning
- insight...
  - AlpacaFarm
  - is this useful?
Yannic Kilcher segment
- Original LLaMA
- LLaMA-Accessory (potential fine-tune tool)
  - together.ai, openchat, lmsys.org (tools that leverage llama)

Design Experiment/Hack Minimum Viable Product/Engineering Draft

Dataset Creation

Kaggle

kaggle always have some interesting datasets (interview quetions related)
able to manually compile a list of interesting interview questions
- potential OCR (future... TODO)
use GPT4 API to synthetically generate responses
- potential human reinforcement here
  - scale.ai/mechanical turk stuff? (TODO)

Implementation

Joke

Discuss B4 Implementation

Data Processing

acquire datasets from kaggle
random code in data_processing
- Use GTP4 to do data formatting work
create .env with OPENAI_API_KEY=...
alternatively, upload/pull datasets into/from huggingface

Fine-Tune Framework for Pipeline

Code to be executed are found in Notebooks, which contains a bunch of colab notebooks that should be one-click solutions

FineTune Script

good no fluff script
copy of code to use in pipeline in Notebooks

Promising FineTune Framework

this looks like an out of the box solution - one click colab solution
cons is that it uses some strange sharded model of llama 7b

Fine-tune Code Examples

Benchmarks

LlamaIndex Semantic Similarity Evaluator

compare text
basis for more rigorous benchmarking
- eval-2-base_llama vs eval-2-llama_finetune

MMLU (Massive Multitask Language Understanding)

left as an excersize to better understand benchmarking

Initial Proposal

choose/create fine-tuning dataset
prepare dataset
choose fine-tuning framework
configure fine-tuning process
train the model
evaluate the model

Code

Replit (MarkDown Renderer, Jupyter Env)

Github (Acutal Codebase - I hear it's on chain)

https://github.com/joexu22/llama2-finetune

Executable Code in Notebook that runs on Collab

Collab Notesbooks in Repo

Lambda Labs/Cloud Compute

Yo, all instances are reserved
- WTF
- need to containerize application so that it can run on any compute cloud/otherwise/etc.
  - this is the actual product
  - buried Alpha if you read this far

Hugging Face

https://huggingface.co/UrbanJoe

joexu22 / llama2-finetune Goto Github PK

llama2-finetune's Introduction