]> git.djapps.eu Git - pkg/ggml/sources/llama.cpp/commit
overview / pkg / ggml / sources / llama.cpp / commit
summary | shortlog | log | commit | commitdiff | tree
(parent: 79c137f)
server: fix regression on streamed non-chat completion w/ stops (#13785)
authorOlivier Chafik <redacted>
Mon, 26 May 2025 13:16:37 +0000 (06:16 -0700)
committerGitHub <redacted>
Mon, 26 May 2025 13:16:37 +0000 (14:16 +0100)
commitf13847cfb560d92492b480cca2c5d6aa9473cde3
treede87a0deddb7ebfed7c000c846f835c831733b22tree
parent79c137f77677b3c8ee3c60a7da033721b938399acommit | diff
server: fix regression on streamed non-chat completion w/ stops (#13785)

* more forgiving message diffs: partial stop words aren't erased, full stops are

* Add (slow) server test for completion + stream + stop
common/chat.cpp diff | blob | history
tools/server/tests/unit/test_completion.py diff | blob | history
Packaging of ggml-org/llama.cpp