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

💡 Get 30 (free) AI project ideas: https://30aiprojects.com/

In this video, I discuss fine-tuning an LLM using QLoRA (i.e. Quantized Low-rank Adaptation). Example code is provided for training a custom YouTube comment responder using Mistral-7b-Instruct.

More Resources:
▶️ Series Playlist:    • Large Language Models (LLMs)  
🎥 Fine-tuning with OpenAI:    • 3 Ways to Make a Custom AI Assistant | RAG...  

📰 Read more: https://medium.com/towards-data-scien...
💻 Colab: https://colab.research.google.com/dri...
💻 GitHub: https://github.com/ShawhinT/YouTube-B...
🤗 Model: https://huggingface.co/shawhin/shawgp...
🤗 Dataset: https://huggingface.co/datasets/shawh...

[1] Fine-tuning LLMs:    • Fine-tuning Large Language Models (LLMs) |...  
[2] ZeRO paper: https://arxiv.org/abs/1910.02054
[3] QLoRA paper: https://arxiv.org/abs/2305.14314
[4] Phi-1 paper: https://arxiv.org/abs/2306.11644
[5] LoRA paper: https://arxiv.org/abs/2106.09685

--
Homepage: https://www.shawhintalebi.com/

Intro - 0:00
Fine-tuning (recap) - 0:45
LLMs are (computationally) expensive - 1:22
What is Quantization? - 4:49
4 Ingredients of QLoRA - 7:10
Ingredient 1: 4-bit NormalFloat - 7:28
Ingredient 2: Double Quantization - 9:54
Ingredient 3: Paged Optimizer - 13:45
Ingredient 4: LoRA - 15:40
Bringing it all together - 18:24
Example code: Fine-tuning Mistral-7b-Instruct for YT Comments - 20:35
What's Next? - 35:22