text-generation-inference
5f78ec32
- Do not convert weight scale to e4m3fnuz on CUDA (#2917)
Go
Login via GitHub
Home
Pricing
FAQ
Install
Login
via GitHub
Commit
View On
GitHub
Commit
337 days ago
Do not convert weight scale to e4m3fnuz on CUDA (#2917)
References
#2917 - Do not convert weight scale to e4m3fnuz on CUDA
Author
danieldk
Parents
922cc38f
Loading