Fine-Tune Language Models with LoRA! OobaBooga Walkthrough and Explanation.

Показать описание

In this video, we dive into the world of LoRA (Low-Rank Approximation) to fine-tune large language models. We'll explore how LoRA works, its significance in reducing memory usage, and how to implement it using oobabooga's text generation web UI. Whether you're a beginner or a pro, this step-by-step tutorial will help you harness the power of LoRA to improve your language model's performance. Don't miss out on our explanation of the underlying linear algebra concepts, as well as a detailed breakdown of the hyperparameters involved in LoRA training. Join us in our quest for efficient language model fine-tuning!

#LoRA #LanguageModel #FineTuning #NLP #AI #machinelearning

0:00 Intro
0:30 What are LoRA's
4:48 How to use LoRA's in OobaBooga

Рекомендации по теме

Комментарии

That is exactly what I was waiting for!! Thank you so much for the insight!

swannschilling

Just keep these coming. Tons of helpful information.

kaymcneely

Thank you for this. I am sick of seeing click bait or low quality content when Im trying to self educate about llms and get some good knowledge .

MichaelSalaA

Didn't know I wanted Kurt Cobain to teach me ML but I'm here for it!

victarion

Hey thanks for sharing your time and knowledge with us!

nartrab

I am very hopeful for local OpenSource AI this year in particular. Having LoRA's fine tuned to certain characters would be incredible rather than just using the old prompt method.

infini_ryu

This was AWSOME!!!! Thank you, Thank you, Thank you, this was a total mystery to me you managed to clear it up completely. I truly thank you for this video, have a blessed year!

timothymaggenti

I'd love to see a video on building these data sets. I am currently working on an AI safety and morality project, and this is super helpful.

manslaughterinc.

Nice. Will check it out over the weekend.

ArjunKrishnaUserProfile

Thanks for also explaining math "basics" in a nutshell. Unfortunately I've forgotten about everything in the regards what I learned some 20 years ago when I studied because I never had to use it. Who would have thought that one day I actually coule make use of it? But it's as one of my teachers once said: Used it or lose it.

testales

Really helpful and informative. Thank you! Could you talk about using these for query, key, and value matrices in attention? Also, explanations of where in particular it can be used in models that aren’t transformers would be interesting 😊

amelieschreiber

I think this video was very informative! It just boils down to math which is kind of cool

blondiedawn

I love your content, very tangible, but still ambitious with theory, thanks a lot!

cz

Great video! Do expand on the parameters section. We are all out of ram.

wpyx

Could you make a video going through an example from start to finish? For example, expanding on a base model by creating a Lora with specific knowledge that it can only know with the Lora trained (e.g. the contents of a certain book that you feed it) and have it answer questions about the knowledge that you added. A lot of people are searching for this information, especially now that models have come on the market that can be used for commercial use (e.g. MPT-7B). Easy 10.000+ subscribers if you were to make such a video where you for example take MPT-7B or other commercially usable model and show how to make the data set and instructions and feed it, I don't know, scripts on every Simpsons episode and then ask the model to make a list of all the characters featured in episode "Marge Vs The Monorail".

netwar

Just found your channel. Excellent Content - Another sub for you sir!

andre-le-bone-aparte

Can you give examples of training data and results?
that would help a lot

cryptoape

Great work, you are doing a great job at helping the community with your videos. I had one question or confusion. At the start in the architecture diagram, the talk was about fine tuning using Lora on historical medical appeals however during the hands on part the example was to generate question and answer on some other book dataset. I wasn’t able to relate to it.

karrtikiyer

This was great. I would love to see you do this in python code at a simple level to set these parameters instead of the GUI

logan

Just subscribed. Would love a video on training and fine-tuning to clone characters or yourself.

trackerprince

Fine-Tune Language Models with LoRA! OobaBooga Walkthrough and Explanation.

Fine-tuning LLMs with PEFT and LoRA

Fine-Tune Language Models with LoRA! OobaBooga Walkthrough and Explanation.

Fine-tuning Large Language Models (LLMs) | w/ Example Code

LoRA - Low-rank Adaption of AI Large Language Models: LoRA and QLoRA Explained Simply

Steps By Step Tutorial To Fine Tune LLAMA 2 With Custom Dataset Using LoRA And QLoRA Techniques

LoRA & QLoRA Fine-tuning Explained In-Depth

Low-rank Adaption of Large Language Models: Explaining the Key Concepts Behind LoRA

What is LoRA? Low-Rank Adaptation for finetuning LLMs EXPLAINED

Generative AI - Techniques for Fine Tuning LLMs

Fine-Tune Large LLMs with QLoRA (Free Colab Tutorial)

How to Fine-tune Large Language Models Like ChatGPT with Low-Rank Adaptation (LoRA)

How to fine-tune a model using LoRA (step by step)

Fine-tune Gemma models With Custom Data in Keras using LoRA

LoRA explained (and a bit about precision and quantization)

Fine Tune LLaMA 2 In FIVE MINUTES! - 'Perform 10x Better For My Use Case'

Fine Tuning LLM Models – Generative AI Course

LoRA: Low-Rank Adaptation of Large Language Models - Explained visually + PyTorch code from scratch

'okay, but I want Llama 3 for my specific use case' - Here's how

Fine-tuning a CRAZY Local Mistral 7B Model - Step by Step - together.ai

Low-rank Adaption of Large Language Models Part 2: Simple Fine-tuning with LoRA

LoRA fine-tuning for custom dataset codes explained

PEFT LoRA Finetuning With Oobabooga! How To Configure Other Models Than Alpaca/LLaMA Step-By-Step.

QLoRA—How to Fine-tune an LLM on a Single GPU (w/ Python Code)

LLAMA-3.1 🦙: EASIET WAY To FINE-TUNE ON YOUR DATA 🙌