Skip to content

Train GPT-3 at home by combining the latest research with custom innovations.

License

Notifications You must be signed in to change notification settings

FOULAB/HomebrewNLP

 
 

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

HomebrewNLP

Overview

"...Our goal is to open up the space by combining every form of efficient training we have. If we throw enough tradeoffs against it, a model of this size (GPT-3) should be trainable on commodity hardware (<1k if purchased as upgrades) ... Compute-memory tradeoffs (like MOE) aren't enough ... we want more efficient training using extragradient methods and better optimizers (Shampoo)" - Lucas Nestler

Example Command

python3 main.py configs/small.yaml

DeepSource | Dataset | Discord

About

Train GPT-3 at home by combining the latest research with custom innovations.

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages

  • Python 100.0%