Skip to content

Streamline sample_grammar #1606

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
LostRuins merged 1 commit into from
Jun 17, 2025
Merged

Streamline sample_grammar #1606

Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
17 changes: 10 additions & 7 deletions gpttype_adapter.cpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -1572,32 +1572,35 @@ void sample_grammar(FileFormat file_format, int32_t n_vocab, llama_token_data_ar

std::vector<std::pair<std::vector<uint32_t>, llama_partial_utf8>> candidates_decoded;
std::vector<llama_grammar_candidate> candidates_grammar;
std::vector<uint8_t> rejects;
candidates_decoded.reserve(candidates->size);
candidates_grammar.reserve(candidates->size);
rejects.assign(candidates->size, false);

for (size_t i = 0; i < candidates->size; ++i) {
const llama_token id = candidates->data[i].id;
const std::string piece = FileFormatTokenizeID(id,file_format);
bool found_eog = std::find(eog_tokens.begin(), eog_tokens.end(), id) != eog_tokens.end();
if (found_eog) {
if (!allow_eos) {
candidates->data[i].logit = -INFINITY;
rejects[i] = true;
}
} else if (piece.empty() || piece[0] == 0) {
candidates->data[i].logit = -INFINITY;
rejects[i] = true;
} else {
candidates_decoded.push_back(decode_utf8(piece.c_str(), grammar->partial_utf8));
candidates_grammar.push_back({ i, candidates_decoded.back().first.data(), candidates_decoded.back().second });
}
}

const auto rejects = llama_grammar_reject_candidates(grammar->rules, grammar->stacks, candidates_grammar);
for (const auto & reject : rejects) {
candidates->data[reject.index].logit = -INFINITY;
for (auto reject: llama_grammar_reject_candidates(grammar->rules, grammar->stacks, candidates_grammar)) {
rejects[reject.index] = true;
}

auto first = candidates->data;
auto last = first + candidates->size;
last = std::remove_if(first, last,
[&](const llama_token_data & tk){ return tk.logit == -INFINITY; });
[&](const llama_token_data & tk){ return rejects[&tk - first]; }); // tk.logit == -INFINITY; });
candidates->size = last - first;
}

Expand Down