thek0tyara/llama-cpp-turboquant
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity Actions 4

common/parser: add proper reasoning tag prefill reading (#20424)

Browse source 
* Implement proper prefill extraction

* Refactor cli parameters, update docs, move reasoning budget sampler part to common/reasoning-budget.cpp

* Update tools/server/server-task.cpp

* refactor: move grammars to variant, remove grammar_external, handle exception internally

* Make code less C++y

Co-authored-by: Georgi Gerganov <ggerganov@gmail.com>
This commit is contained in:
Piotr Wilkin (ilintar) 2026-03-19 16:58:21 +01:00 • committed by GitHub
parent c1258830b2
commit 5e54d51b19
No known key found for this signature in database
GPG key ID: B5690EEEBB952194
33 changed files with 651 additions and 454 deletions
Download patch file Download diff file Expand all files Collapse all files

10
common/arg.cpp
View file

@ -1830,23 +1830,23 @@ common_params_context common_params_parser_init(common_params & params, llama_ex
).set_sparam());
add_opt(common_arg(
{"--grammar"}, "GRAMMAR",
string_format("BNF-like grammar to constrain generations (see samples in grammars/ dir) (default: '%s')", params.sampling.grammar.c_str()),
"BNF-like grammar to constrain generations (see samples in grammars/ dir)",
[](common_params & params, const std::string & value) {
params.sampling.grammar = value;
params.sampling.grammar = {COMMON_GRAMMAR_TYPE_USER, value};
}
).set_sparam());
add_opt(common_arg(
{"--grammar-file"}, "FNAME",
"file to read grammar from",
[](common_params & params, const std::string & value) {
params.sampling.grammar = read_file(value);
params.sampling.grammar = {COMMON_GRAMMAR_TYPE_USER, read_file(value)};
}
).set_sparam());
add_opt(common_arg(
{"-j", "--json-schema"}, "SCHEMA",
"JSON schema to constrain generations (https://json-schema.org/), e.g. `{}` for any JSON object\nFor schemas w/ external $refs, use --grammar + example/json_schema_to_grammar.py instead",
[](common_params & params, const std::string & value) {
params.sampling.grammar = json_schema_to_grammar(json::parse(value));
params.sampling.grammar = {COMMON_GRAMMAR_TYPE_OUTPUT_FORMAT, json_schema_to_grammar(json::parse(value))};
}
).set_sparam());
add_opt(common_arg(
@ -1863,7 +1863,7 @@ common_params_context common_params_parser_init(common_params & params, llama_ex
std::istreambuf_iterator<char>(),
std::back_inserter(schema)
);
params.sampling.grammar = json_schema_to_grammar(json::parse(schema));
params.sampling.grammar = {COMMON_GRAMMAR_TYPE_OUTPUT_FORMAT, json_schema_to_grammar(json::parse(schema))};
}
).set_sparam());
add_opt(common_arg(