]> git.djapps.eu Git - pkg/ggml/sources/llama.cpp/commit
overview / pkg / ggml / sources / llama.cpp / commit
summary | shortlog | log | commit | commitdiff | tree
(parent: 89d5d90)
We could use std::unordered_map over std::map (#305)
authorFabio R. Sluzala <redacted>
Tue, 21 Mar 2023 17:21:50 +0000 (14:21 -0300)
committerGitHub <redacted>
Tue, 21 Mar 2023 17:21:50 +0000 (19:21 +0200)
commit353ec251a42491f5192c48561da4b444ef67f23c
tree95783d81ab8be8a6875ec708e2e2a7882222250atree
parent89d5d90f3b6d25f134da7a8e252c3432bffcf674commit | diff
We could use std::unordered_map over std::map (#305)

* Improve performance by changing std::map to std::unordered_map and std::map<id, token> id_to_token; to std::vector<token> id_to_token;

* fix last commit on gpt_vocab_init add vocab.id_to_token.resize(vocab.token_to_id.size());

* Removed include <map>

* Nest struct token score inside gpt_vocab

* renamed token to tok
main.cpp diff | blob | history
quantize.cpp diff | blob | history
utils.cpp diff | blob | history
utils.h diff | blob | history
Packaging of ggml-org/llama.cpp