Geeks
jkraybill
/
gpt-2-b
New fork of nshepperd for colab finetuning work.
Primary Language
Jupyter Notebook
MIT License
MIT
gpt-2-b GitHub
Markdown
Readme
1
Issue
1
Stargazer
2
Watchers
Issues
Open issue
remove if layer % 12 == 0: and use AdafactorOptimizer can train all 432 vars
#1 opened
5 years ago
by
deepbluesea
3