]> git.djapps.eu Git - pkg/ggml/sources/llama.cpp/commit
llama : add option to save memory in device buffers (#22679)
authorGeorgi Gerganov <redacted>
Tue, 5 May 2026 03:35:07 +0000 (06:35 +0300)
committerGitHub <redacted>
Tue, 5 May 2026 03:35:07 +0000 (06:35 +0300)
commitd6e7b033a42b2880496d1e9db1df5527e48a8869
tree13067ef9513b33c525cad6a3d8837b49ddae10c1
parentfa595462ca2d5ea0feca20c41bd1431b3d4285ef
llama : add option to save memory in device buffers (#22679)

* llama : add option to save memory in device buffers

* tests : extend llama-save-load-state
common/speculative.cpp
examples/save-load-state/save-load-state.cpp
ggml/src/ggml-metal/ggml-metal-device.h
ggml/src/ggml-metal/ggml-metal-device.m
ggml/src/ggml-metal/ggml-metal.cpp
include/llama.h
src/llama-context.cpp
src/llama-context.h
src/llama-io.h
src/llama-memory-recurrent.cpp
tools/server/server-context.cpp