]> git.djapps.eu Git - pkg/ggml/sources/llama.cpp/commit
kv-cache : support V-less cache (#19067)
authorGeorgi Gerganov <redacted>
Sun, 25 Jan 2026 13:48:56 +0000 (15:48 +0200)
committerGitHub <redacted>
Sun, 25 Jan 2026 13:48:56 +0000 (15:48 +0200)
commitd9c6ce46f747189cd6238ca7699253613f77c016
tree753ada122bfe8bacac8fa696ccabcb150d8303e9
parent70d860824a72c3daaea678b74d5051fdd46d9c38
kv-cache : support V-less cache (#19067)

* kv-cache : support V-less cache

* cuda : better check for V_is_K_view

* cuda : improve V_is_K_view check

* graph : add comments

* hparams : refactor
ggml/src/ggml-cuda/fattn-common.cuh
ggml/src/ggml-cuda/fattn.cu
src/llama-context.cpp
src/llama-graph.cpp
src/llama-graph.h
src/llama-hparams.cpp
src/llama-hparams.h
src/llama-kv-cache.cpp
src/llama-model-saver.cpp
src/llama-model.cpp
src/models/deepseek2.cpp