Skip to content

wangychn/gpt-from-scratch

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

9 Commits
 
 
 
 
 
 
 
 
 
 

Repository files navigation

gpt-from-scratch

Latest project progress repo

Components

Encoder Model

positional_encoding.py: Implements sinusoidal positional encoding and learnt positional encoding (meeting 1)

attention.py: Implements scaled dot product attention and multi head attention (meetings 2 & 3)

transformer.py: Implements decoder-only transformer based on GPT2

Bigram Model

bigram.py: implments a basic bigram model

train_bigram.py: trains the basic bigram

About

Model and training code for a GPT style transformer made from scratching using PyTorch. Made as a project for Michigan Student Artificial Intelligence Lab (MSAIL)

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

 
 
 

Contributors

Languages