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

common : remove json.hpp from common.cpp (#12697)

Browse source 
* common : remove json.hpp from common.cpp

* fix comment
This commit is contained in:
Xuan-Son Nguyen 2025-04-02 09:58:34 +02:00 • committed by GitHub
parent 9bacd6b374
commit 42eb248f46
No known key found for this signature in database
GPG key ID: B5690EEEBB952194
4 changed files with 34 additions and 38 deletions
Download patch file Download diff file Expand all files Collapse all files

11
examples/server/server.cpp
View file

@ -133,7 +133,8 @@ struct slot_params {
auto grammar_triggers = json::array();
for (const auto & trigger : sampling.grammar_triggers) {
grammar_triggers.push_back(trigger.to_json<json>());
server_grammar_trigger ct(std::move(trigger));
grammar_triggers.push_back(ct.to_json());
}
return json {
@ -372,9 +373,9 @@ struct server_task {
const auto grammar_triggers = data.find("grammar_triggers");
if (grammar_triggers != data.end()) {
for (const auto & t : *grammar_triggers) {
auto ct = common_grammar_trigger::from_json(t);
if (ct.type == COMMON_GRAMMAR_TRIGGER_TYPE_WORD) {
const auto & word = ct.value;
server_grammar_trigger ct(t);
if (ct.value.type == COMMON_GRAMMAR_TRIGGER_TYPE_WORD) {
const auto & word = ct.value.value;
auto ids = common_tokenize(vocab, word, /* add_special= */ false, /* parse_special= */ true);
if (ids.size() == 1) {
auto token = ids[0];
@ -392,7 +393,7 @@ struct server_task {
params.sampling.grammar_triggers.push_back({COMMON_GRAMMAR_TRIGGER_TYPE_WORD, word});
}
} else {
params.sampling.grammar_triggers.push_back(ct);
params.sampling.grammar_triggers.push_back(std::move(ct.value));
}
}
}

29
examples/server/utils.hpp
View file

@ -58,6 +58,32 @@ static T json_value(const json & body, const std::string & key, const T & defaul
const static std::string build_info("b" + std::to_string(LLAMA_BUILD_NUMBER) + "-" + LLAMA_COMMIT);
// thin wrapper around common_grammar_trigger with (de)serialization functions
struct server_grammar_trigger {
common_grammar_trigger value;
server_grammar_trigger() = default;
server_grammar_trigger(const common_grammar_trigger & value) : value(value) {}
server_grammar_trigger(const json & in) {
value.type = (common_grammar_trigger_type) in.at("type").get<int>();
value.value = in.at("value").get<std::string>();
if (value.type == COMMON_GRAMMAR_TRIGGER_TYPE_TOKEN) {
value.token = (llama_token) in.at("token").get<int>();
}
}
json to_json() const {
json out {
{"type", (int) value.type},
{"value", value.value},
};
if (value.type == COMMON_GRAMMAR_TRIGGER_TYPE_TOKEN) {
out["token"] = (int) value.token;
}
return out;
}
};
//
// tokenizer and input processing utils
//
@ -627,7 +653,8 @@ 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(trigger.to_json<json>());
server_grammar_trigger ct(trigger);
grammar_triggers.push_back(ct.to_json());
}
llama_params["grammar_triggers"] = grammar_triggers;
llama_params["preserved_tokens"] = chat_params.preserved_tokens;