tqcq/llama.cpp
0
0
Fork 0
mirror of https://github.com/ggml-org/llama.cpp.git synced 2025-07-01 13:05:52 +00:00
Code Issues Packages Projects Releases Wiki Activity

sampling: fix top_k <= 0 (#5388)

Browse Source 
* sampling: fix top_k <= 0

* Update llama.cpp

Co-authored-by: Georgi Gerganov <ggerganov@gmail.com>

---------

Co-authored-by: Georgi Gerganov <ggerganov@gmail.com>
This commit is contained in:
Johannes Gäßler
2024-02-08 09:46:30 +01:00
committed by GitHub
parent 8504d2d0da
commit 26d4efd11e
3 changed files with 7 additions and 1 deletions
Download Patch File Download Diff File Expand all files Collapse all files

4
llama.cpp
View File

@ -8585,6 +8585,10 @@ void llama_sample_top_k(struct llama_context * ctx, llama_token_data_array * can
// }
const int64_t t_start_sample_us = ggml_time_us();
if (k <= 0) {
k = candidates->size;
}
k = std::max(k, (int) min_keep);
k = std::min(k, (int) candidates->size);