]> git.djapps.eu Git - pkg/ggml/sources/llama.cpp/commit
overview / pkg / ggml / sources / llama.cpp / commit
summary | shortlog | log | commit | commitdiff | tree
(parent: f14f4e4)
sampling: reuse token data buffer in llama_sampler_sample (#18365)
authorJay Zenith <redacted>
Tue, 30 Dec 2025 14:27:49 +0000 (06:27 -0800)
committerGitHub <redacted>
Tue, 30 Dec 2025 14:27:49 +0000 (16:27 +0200)
commitc32fa21db8a631e9127e55f69a3d2bdaa9f71824
tree06e00d781a09c0e42c34b1655e8657f3f7f9f575tree
parentf14f4e421b2177fadcf9d15ebccb0492e5464d86commit | diff
sampling: reuse token data buffer in llama_sampler_sample (#18365)

* sampling: reuse token data buffer in llama_sampler_sample

* move cur buffer before timing section, after samplers

* minor : fix build

---------

Co-authored-by: Georgi Gerganov <redacted>
src/llama-sampling.cpp diff | blob | history
src/llama-sampling.h diff | blob | history
Packaging of ggml-org/llama.cpp