llama.cpp
0dbfa66a
- return filter to save memory (#24125)
Go
Login via GitHub
Home
Pricing
FAQ
Install
Login
via GitHub
Commit
View On
GitHub
Commit
2 days ago
return filter to save memory (#24125) Co-authored-by: lvyichen <lvyichen@stepfun.com>
References
#24125 - fix: step35 MTP does not allocate KV cache for all layers
Author
forforever73
Parents
e8023568
Loading