llama.cpp
5b1a8c23
- cuda: top-k MoE should always fire
Go
Login via GitHub
Home
Pricing
FAQ
Install
Login
via GitHub
Commit
View On
GitHub
Commit
7 days ago
cuda: top-k MoE should always fire
References
aman/top-k-alloc-dep
#28432 - cuda: top-k MoE should always fire
Author
am17an
Parents
4d917609
Loading