CS348Project/llama.cpp
5
0
Fork 0
mirror of https://github.com/ggml-org/llama.cpp.git synced 2025-10-28 08:31:25 +00:00
Code Issues Packages Projects Releases Wiki Activity

llama : minor fixes for up llama load model speed (#11448)

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

* llama_model_loader::init_mapping - replace new llama_mmap to std::make_unique<llama_mmap> for clean code & reduce (/2) time of running init_mappings

* Update src/llama-vocab.cpp

---------

Co-authored-by: lexasub <empty@empty.ru>
Co-authored-by: Diego Devesa <slarengh@gmail.com>
This commit is contained in:
lexasub
2025-01-27 17:42:09 +04:00
committed by GitHub
parent df984e0147
commit a5203b4465
2 changed files with 8 additions and 3 deletions
Download Patch File Download Diff File Expand all files Collapse all files

2
src/llama-model-loader.cpp
View File

@@ -819,7 +819,7 @@ void llama_model_loader::init_mappings(bool prefetch, llama_mlocks * mlock_mmaps
for (const auto & file : files) { for (const auto & file : files) {
auto * reg = ggml_backend_dev_backend_reg(ggml_backend_dev_by_type(GGML_BACKEND_DEVICE_TYPE_CPU)); auto * reg = ggml_backend_dev_backend_reg(ggml_backend_dev_by_type(GGML_BACKEND_DEVICE_TYPE_CPU));
auto * is_numa_fn = (decltype(ggml_is_numa) *) ggml_backend_reg_get_proc_address(reg, "ggml_backend_cpu_is_numa"); auto * is_numa_fn = (decltype(ggml_is_numa) *) ggml_backend_reg_get_proc_address(reg, "ggml_backend_cpu_is_numa");
std::unique_ptr<llama_mmap> mapping(new llama_mmap(file.get(), prefetch ? -1 : 0, is_numa_fn())); std::unique_ptr<llama_mmap> mapping = std::make_unique<llama_mmap>(file.get(), prefetch ? -1 : 0, is_numa_fn());
mmaps_used.emplace_back(mapping->size(), 0); mmaps_used.emplace_back(mapping->size(), 0);
if (mlock_mmaps) { if (mlock_mmaps) {
std::unique_ptr<llama_mlock> mlock_mmap(new llama_mlock()); std::unique_ptr<llama_mlock> mlock_mmap(new llama_mlock());

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 pair_hash {
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, pair_hash> 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;