Megatron-DeepSpeed
fce1a98e
- Clarify loss on targets & remove kwarg
Go
Login via GitHub
Home
Pricing
FAQ
Install
Login
via GitHub
Commit
View On
GitHub
Commit
4 years ago
Clarify loss on targets & remove kwarg
References
#326 - Add option to normalize loss per target
#336 - Add multiple evaluation compat
Author
Muennighoff
Parents
7bc1dd20
Loading