Learn how to build a custom deep learning machine in this comprehensive 49-minute video. Discover the benefits of building your own workstation, compare costs with pre-built options and cloud services, and gain insights into selecting the right components. Explore essential resources for researching parts, understand the importance of GPU memory, and receive guidance on choosing CPUs, motherboards, memory, storage, cases, and power supplies. Get practical advice on budgeting, component purchasing, and avoiding common pitfalls in the build process. Whether you're a beginner or an experienced builder, gain valuable knowledge to create a powerful machine tailored to your deep learning needs.
How to Build a Deep Learning Machine - Everything You Need to Know
Aleksa Gordić - The AI Epiphany via YouTube
Overview
Syllabus
Intro - announcing the new series
Why should you build your deep learning workstation?
Quick skim of the components
pcpartpicker is your friend
Cost comparisons - Lambda Labs workstation & laptop, cloud
Researching the components - useful resources
Overview Tim Dettmers' blogs
Do you need 11+ GBs?
Don't follow someone's advice blindly!
Upgrade to NVIDIA 40x series?
Other resources
How to pick the right components?
CPU - Ryzen 9 7950x or Threadripper?
CPU Cooler
MoBo
Memory and storage
Case
PSU
Quick mention of my peripherals
Advice for targeting your particular budget
Tips for buying components - site aggregators, Amazon, legitness
Outro
Taught by
Aleksa Gordić - The AI Epiphany
