Скачать или смотреть Muon vs AdamW - Why Muon Is Better Optimizer (for LLMs)

Muon vs AdamW - Why Muon Is Better Optimizer (for LLMs)

GitHub - https://github.com/vukrosic/muon-opti...

PDF Research - https://github.com/vukrosic/muon-opti...

Jupyther - https://github.com/vukrosic/muon-opti...

NovitaAI GPUs - https://novita.ai/?ref=mjqyndm&utm_so...

--

My open course to become AI researcher / engineer - https://github.com/vukrosic/ultimate-...

Code DeepSeek V3 From Scratch Full Course -    • Understand & Code DeepSeek V3 From Scratch...  

contact: [email protected]

00:00 – Intro: Muon vs AdamW
00:12 – Experiment 1: Learning Rate Robustness
01:32 – Experiment 2: Scaling to Larger Models
03:22 – Research Repository Overview
04:01 – Running on Google Colab & Novita GPU
05:08 – Novita Setup & Spot Instance Tips
06:32 – Uploading & Running the Notebook
07:40 – Installing Dependencies & Setup
08:29 – Running the Experiments
09:06 – Getting the Results & Tips
09:43 – Final Thoughts & Outro -
-
Support my Free & Open AI research - https://ko-fi.com/vukai
Help me do this full time