Skip to content

Activity

Update nanogpt script

dewithpushed 1 commit to main • d7fe3d1…787007d • 
on Apr 20, 2024

Bugfix: Add to device on q k v layers

dewithpushed 1 commit to main • 74c476f…d7fe3d1 • 
on Apr 14, 2024

Add transformer notebook

dewithpushed 1 commit to main • 171f47c…74c476f • 
on Apr 13, 2024

Turn off viz

dewithpushed 1 commit to main • 4e17a38…171f47c • 
on Apr 13, 2024

Ignore all files in the "data" directory except for ".gitignore" and …

dewithpushed 1 commit to main • 945ea6a…4e17a38 • 
on Apr 13, 2024

Update requirements

dewithpushed 1 commit to main • fd92fc9…945ea6a • 
on Apr 13, 2024

add dropout and layer norm

dewithpushed 1 commit to main • 49f1840…fd92fc9 • 
on Apr 13, 2024

Minor changes in logging and data loading functions

dewithpushed 1 commit to main • 7e1772d…49f1840 • 
on Mar 31, 2024

Add version2 of bigram and do model visualization

dewithpushed 1 commit to main • 39c7fcc…7e1772d • 
on Mar 24, 2024

Add self attention notebook

dewithpushed 1 commit to main • 7890165…39c7fcc • 
on Mar 22, 2024

Add bigram class to a submodule

dewithpushed 2 commits to main • fcc025e…7890165 • 
on Mar 12, 2024

Re-organize the python project scripts

dewithpushed 1 commit to main • 82f4bdc…fcc025e • 
on Mar 12, 2024

Stop tracking a local conf file

dewithpushed 1 commit to main • e6fcb7c…82f4bdc • 
on Mar 12, 2024

Fix bug in the bigram model when reshaping for cross entropy loss

dewithpushed 1 commit to main • e3b98d5…e6fcb7c • 
on Mar 12, 2024

Add the training script for the baseline model

dewithpushed 1 commit to main • 1bc0175…e3b98d5 • 
on Mar 5, 2024

Add data loader notebook

dewithpushed 1 commit to main • 7ca6bb9…1bc0175 • 
on Mar 4, 2024

Add data tokenization

dewithpushed 1 commit to main • 6741c00…7ca6bb9 • 
on Mar 4, 2024

Curate raw data

dewithpushed 1 commit to main • 10860a6…6741c00 • 
on Mar 3, 2024

Add requirements.txt file and data ingestion notebook

dewithpushed 1 commit to main • 66a694a…10860a6 • 
on Mar 3, 2024

Initial commit

dewithcreated main • 66a694a • 
on Mar 2, 2024