ver4a/llama.cpp
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity

tool-call: fix Qwen 2.5 Coder support, add micro benchmarks, support trigger patterns for lazy grammars (#12034)

Browse source 
* sampler: turn lazy grammar trigger words to regexes

* add scripts/tool_bench.sh & .py

* constrain llama json output regardless of function name if matches at beginning

* update relaxed newline space rule in grammar tests

* support add_generation_prompt query parameter (useful for /apply_template)

* Update src/llama-grammar.cpp

Co-authored-by: Georgi Gerganov <ggerganov@gmail.com>

---------

Co-authored-by: Georgi Gerganov <ggerganov@gmail.com>
This commit is contained in:
Olivier Chafik 2025-03-05 13:05:13 +00:00 • committed by GitHub
parent fa31c438e0
commit 669912d9a5
No known key found for this signature in database
GPG key ID: B5690EEEBB952194
26 changed files with 1314 additions and 408 deletions
Download patch file Download diff file Expand all files Collapse all files

2
examples/server/public_legacy/json-schema-to-grammar.mjs
View file

@ -1,5 +1,5 @@
// WARNING: This file was ported from json_schema_to_grammar.py, please fix bugs / add features there first.
const SPACE_RULE = '| " " | "\\n" [ \\t]{0,20}';
const SPACE_RULE = '| " " | "\\n"{1,2} [ \\t]{0,20}';
function _buildRepetition(itemRule, minItems, maxItems, opts={}) {
if (minItems === 0 && maxItems === 1) {