cturan/llama.cpp @ a71a4075cdb8e81eaaa834e48ffda88b038286bc

spogulis no https://github.com/cturan/llama.cpp

David Huang 7f323a589f Add `--no-op-offload` to improve `-ot` pp perf in MoE models like llama4 400B (#13386)		8 mēneši atpakaļ
..
llama-cpp.h	afa8a9ec9b llama : add `llama_vocab`, functions -> methods, naming (#11110)	1 gadu atpakaļ
llama.h	7f323a589f Add `--no-op-offload` to improve `-ot` pp perf in MoE models like llama4 400B (#13386)	8 mēneši atpakaļ