vrr/koboldcpp
2
0
Fork 0
mirror of https://github.com/LostRuins/koboldcpp.git synced 2025-09-10 17:14:36 +00:00
Code Issues Projects Releases Packages Wiki Activity Actions

improved rep pen speed

Browse source
This commit is contained in:
Concedo 2024-04-24 21:29:21 +08:00
parent a681cdd9ef
commit cb2dbe9e9a
1 changed files with 8 additions and 2 deletions
Download patch file Download diff file Expand all files Collapse all files

10
gpttype_adapter.cpp
View file

@ -429,9 +429,15 @@ void sample_rep_pen(int n_ctx, int rep_pen_range, float rep_pen, float presence_
const int64_t t_start_sample_us = ggml_time_us(); const int64_t t_start_sample_us = ggml_time_us();
// Create a frequency map to count occurrences of each token in last_tokens
std::unordered_map<llama_token, int> token_count;
for (size_t i = 0; i < last_n_repeat; ++i) {
token_count[last_tokens[i]]++;
}
for (size_t i = 0; i < candidates->size; ++i) { for (size_t i = 0; i < candidates->size; ++i) {
const auto * token_iter = std::find(last_tokens, last_tokens + last_tokens_size, candidates->data[i].id); const auto token_iter = token_count.find(candidates->data[i].id);
if (token_iter == last_tokens + last_tokens_size) { if (token_iter == token_count.end()) {
continue; continue;
} }