]> git.djapps.eu Git - pkg/ggml/sources/llama.cpp/commit
overview / pkg / ggml / sources / llama.cpp / commit
summary | shortlog | log | commit | commitdiff | tree
(parent: aa6f918)
server : support max_completion_tokens request property (#19831)
authorRadoslav Gerganov <redacted>
Tue, 24 Feb 2026 08:30:00 +0000 (10:30 +0200)
committerGitHub <redacted>
Tue, 24 Feb 2026 08:30:00 +0000 (10:30 +0200)
commitc830f99cfa79d7e627e48de32280838f97b41115
tree1000d534d6c3a8244a9e3a7e72f70a1747dd3b02tree
parentaa6f918c1c786668db530c3a1c3ff8a93da928f7commit | diff
server : support max_completion_tokens request property (#19831)

"max_tokens" is deprectated in favor of "max_completion_tokens" which
sets the upper bound for reasoning+output token.

Closes: #13700
tools/server/server-task.cpp diff | blob | history
Packaging of ggml-org/llama.cpp