tqcq/llama.cpp
0
0
Fork 0
mirror of https://github.com/ggml-org/llama.cpp.git synced 2025-06-28 12:25:03 +00:00
Code Issues Packages Projects Releases 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
26 changed files with 1314 additions and 408 deletions
Download Patch File Download Diff File Expand all files Collapse all files

6
examples/server/utils.hpp
View File

@ -607,6 +607,7 @@ static json oaicompat_completion_params_parse(
inputs.use_jinja = use_jinja;
inputs.parallel_tool_calls = json_value(body, "parallel_tool_calls", false);
inputs.extract_reasoning = reasoning_format != COMMON_REASONING_FORMAT_NONE;
inputs.add_generation_prompt = json_value(body, "add_generation_prompt", true);
if (!inputs.tools.empty() && inputs.tool_choice != COMMON_CHAT_TOOL_CHOICE_NONE && body.contains("grammar")) {
throw std::runtime_error("Cannot use custom grammar constraints with tools.");
}
@ -620,10 +621,7 @@ static json oaicompat_completion_params_parse(
llama_params["grammar_lazy"] = chat_params.grammar_lazy;
auto grammar_triggers = json::array();
for (const auto & trigger : chat_params.grammar_triggers) {
grammar_triggers.push_back({
{"word", trigger.word},
{"at_start", trigger.at_start},
});
grammar_triggers.push_back(trigger.to_json<json>());
}
llama_params["grammar_triggers"] = grammar_triggers;
llama_params["preserved_tokens"] = chat_params.preserved_tokens;