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

top-p: correct gt to gte

Browse source
This commit is contained in:
Alex Renda 2023-06-20 20:48:09 -04:00
parent 1e7755cfcb
commit 407b77cdb3
1 changed files with 3 additions and 2 deletions
Download patch file Download diff file Expand all files Collapse all files

5
llama.cpp
View file

@ -2005,8 +2005,9 @@ void llama_sample_top_p(struct llama_context * ctx, llama_token_data_array * can
for (size_t i = 0; i < candidates->size; ++i) {
cum_sum += candidates->data[i].p;
// Check if the running sum is greater than p or if we have kept at least min_keep tokens
if (cum_sum > p && i + 1 >= min_keep) {
// Check if the running sum is at least p or if we have kept at least min_keep tokens
// we set the last index to i+1 to indicate that the current iterate should be included in the set
if (cum_sum >= p && i + 1 >= min_keep) {
last_idx = i + 1;
break;
}