]> git.djapps.eu Git - pkg/ggml/sources/llama.cpp/commit
overview / pkg / ggml / sources / llama.cpp / commit
summary | shortlog | log | commit | commitdiff | tree
(parent: 1e22599)
llama : enforce the same K and V cache types for DeepSeek V4; enable FA if V cache...
authorfairydreaming <redacted>
Fri, 31 Jul 2026 07:03:30 +0000 (09:03 +0200)
committerGitHub <redacted>
Fri, 31 Jul 2026 07:03:30 +0000 (10:03 +0300)
commit69e62fc77c911da169cc8726b490028d53bb90fe
treebf6b475e219fc425a72f2fe109aeeb0b2bb55c09tree
parent1e22599522269c9d31f4ec3c4b0b75d01f07f9a7commit | diff
llama : enforce the same K and V cache types for DeepSeek V4; enable FA if V cache is quantized (#25871)

* llama : enforce the same K and V cache types for DeepSeek V4; enable FA if V cache is quantized

* llama : enforce the same K and V cache types for MLA models

---------

Co-authored-by: Stanisław Szymczyk <redacted>
src/llama-context.cpp diff | blob | history
Packaging of ggml-org/llama.cpp