ArchiCraft: Solution Architecture Insights for AI Engineering cover image

#002 - How long to train a 70B LLM on 15T tokens using 1024 H100s?

ArchiCraft: Solution Architecture Insights for AI Engineering

00:00

Estimating Compute Power for Training Large Language Models

This chapter explores the mathematical requirements for training large language models on GPU clusters, detailing the necessary FLOPs and training duration. It also discusses the challenges faced by organizations in building these models and offers insights into the future of AI development as technology advances.

Transcript
Play full episode

The AI-powered Podcast Player

Save insights by tapping your headphones, chat with episodes, discover the best highlights - and more!
App store bannerPlay store banner
Get the app