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

impl::load change map bpe_ranks to onordered map for reduce time of impl::load on 30%

Browse source
This commit is contained in:
lexasub 2025-01-27 15:09:34 +04:00
parent acd38efee3
commit 5144a18e67
1 changed files with 7 additions and 2 deletions
Download patch file Download diff file Expand all files Collapse all files

9
src/llama-vocab.cpp
View file

@ -1245,8 +1245,13 @@ struct llama_vocab::impl {
std::vector<llama_token> cache_special_tokens; std::vector<llama_token> cache_special_tokens;
std::vector<std::string> cache_token_to_piece; // llama_token_to_piece(special = true); std::vector<std::string> cache_token_to_piece; // llama_token_to_piece(special = true);
struct PairHash {
std::map<std::pair<std::string, std::string>, int> bpe_ranks; size_t operator()(const std::pair<std::string, std::string>& p) const {
return std::hash<std::string>{}(p.first) ^ //create some hash for pair
(std::hash<std::string>{}(p.second) << 1);
}
};
std::unordered_map<std::pair<std::string, std::string>, int, PairHash> bpe_ranks;
// set of all tokens that cause "end of generation" // set of all tokens that cause "end of generation"
std::set<llama_token> special_eog_ids; std::set<llama_token> special_eog_ids;