transformers
47d77657
- fix gemma4 gradient accumulation loss and last token incorrect labels (#45354)
Go
Login via GitHub
Home
Pricing
FAQ
Install
Login
via GitHub
Commit
View On
GitHub
Commit
53 days ago
fix gemma4 gradient accumulation loss and last token incorrect labels (#45354) * fix gemma4 gradient accumulation loss and last token incorrect labels * modular + also gemma3n --------- Co-authored-by: Cyril Vallez <cyril.vallez@gmail.com>
References
#45354 - fix gemma4 gradient accumulation loss and last token incorrect labels
Author
winglian
Parents
d13b44b6
Loading