Mirrors/llama.cpp
1
0
Fork 0
mirror of https://github.com/ggerganov/llama.cpp.git synced 2025-01-30 22:03:03 +01:00
Code Issues Packages Projects Releases Wiki Activity

format

Browse Source
This commit is contained in:
VJHack 2025-01-13 19:08:30 -06:00
parent 0f7501c913
commit b29deb83cd
2 changed files with 2 additions and 2 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

2
examples/main/README.md
View File

@ -273,7 +273,7 @@ Top-nσ sampling is a text generation method that selects tokens based on a stat
Example usage: `--top-nsigma 1` Example usage: `--top-nsigma 1`
### Logit Bias ### Logit Bias
- `-l TOKEN_ID(+/-)BIAS, --logit-bias TOKEN_ID(+/-)BIAS`: Modify the likelihood of a token appearing in the generated text completion. - `-l TOKEN_ID(+/-)BIAS, --logit-bias TOKEN_ID(+/-)BIAS`: Modify the likelihood of a token appearing in the generated text completion.

2
src/llama-sampling.cpp
View File

@ -1672,7 +1672,7 @@ static void llama_sampler_top_n_sigma_apply(struct llama_sampler * smpl, llama_t
logits_sum += cur_p->data[i].logit; logits_sum += cur_p->data[i].logit;
} }
float mean = (float)logits_sum/cur_p->size; float mean = (float)logits_sum/cur_p->size;
// calculate standard deviation // calculate standard deviation
float acc = 0; float acc = 0;
for(size_t i = 0; i < cur_p->size; ++i){ for(size_t i = 0; i < cur_p->size; ++i){