happyz
/
llama.cpp
mirror of https://github.com/ggerganov/llama.cpp.git
1
0
Fork 0
Code Issues Packages Projects Releases Wiki Activity

more readable

This commit is contained in:
Xuan Son Nguyen 2025-11-20 17:45:37 +01:00
parent 3b195d301a
commit c0b9903a1a
1 changed files with 7 additions and 7 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

14
src/llama-grammar.cpp
View File

@ -349,9 +349,8 @@ const char * llama_grammar_parser::parse_sequence(
// use UINT64_MAX as the empty value because we aligned to the proper uint64_t type so -1 can't be used
// (though it's technically the same as -1 now)
// ref: https://github.com/ggml-org/llama.cpp/pull/17381
auto handle_repetitions = [&](uint64_t min_times, uint64_t max_times) {
bool no_max = max_times == UINT64_MAX;
if (last_sym_start == rule.size()) {
throw std::runtime_error(std::string("expecting preceding item to */+/?/{ at ") + pos);
}
@ -384,14 +383,14 @@ const char * llama_grammar_parser::parse_sequence(
}
uint32_t last_rec_rule_id = 0;
auto n_opt = max_times == UINT64_MAX ? 1 : max_times - min_times;
auto n_opt = no_max ? 1 : max_times - min_times;
llama_grammar_rule rec_rule(prev_rule);
for (uint64_t i = 0; i < n_opt; i++) {
rec_rule.resize(prev_rule.size());
uint32_t rec_rule_id = generate_symbol_id( rule_name);
if (i > 0 || max_times == UINT64_MAX) {
rec_rule.push_back({LLAMA_GRETYPE_RULE_REF, max_times == UINT64_MAX ? rec_rule_id : last_rec_rule_id});
if (i > 0 || no_max) {
rec_rule.push_back({LLAMA_GRETYPE_RULE_REF, no_max ? rec_rule_id : last_rec_rule_id});
}
rec_rule.push_back({LLAMA_GRETYPE_ALT, 0});
rec_rule.push_back({LLAMA_GRETYPE_END, 0});
@ -486,7 +485,7 @@ const char * llama_grammar_parser::parse_sequence(
uint64_t min_times = std::stoul(std::string(pos, int_end - pos));
pos = parse_space(int_end, is_nested);
uint64_t max_times = UINT64_MAX;
uint64_t max_times = UINT64_MAX; // default: no max limit
if (*pos == '}') {
max_times = min_times;
@ -507,7 +506,8 @@ const char * llama_grammar_parser::parse_sequence(
} else {
throw std::runtime_error(std::string("expecting ',' at ") + pos);
}
if (min_times > MAX_REPETITION_THRESHOLD || max_times > MAX_REPETITION_THRESHOLD) {
bool has_max = max_times != UINT64_MAX;
if (min_times > MAX_REPETITION_THRESHOLD || (has_max && max_times > MAX_REPETITION_THRESHOLD)) {
throw std::runtime_error(std::string("number of repetitions exceeds sane defaults, please reduce the number of repetitions"));
}
handle_repetitions(min_times, max_times);