DeepSpeed
26cefc57
- Removing need to grad_reduc attribute.
Go
Login via GitHub
Home
Pricing
FAQ
Install
Login
via GitHub
Commit
View On
GitHub
Commit
2 years ago
Removing need to grad_reduc attribute.
References
#2847 - ZeRO Gradient Accumulation Dtype.
Author
jomayeri
Parents
62a82368
Loading