root/llama.cpp
1
0
Fork 0
mirror of https://github.com/ggerganov/llama.cpp.git synced 2024-12-27 03:44:35 +00:00
Code Issues Actions 4 Packages Projects Releases Wiki Activity

speculative : print draft token pieces

Browse Source
This commit is contained in:
Georgi Gerganov 2023-09-04 12:54:38 +03:00
parent ebe41d49a6
commit 6c150d763e
No known key found for this signature in database
GPG Key ID: 449E073F9DC10735
1 changed files with 1 additions and 1 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

2
examples/speculative/speculative.cpp
View File

@ -241,7 +241,7 @@ int main(int argc, char ** argv) {
llama_sample_softmax(ctx_dft, &cur_p); llama_sample_softmax(ctx_dft, &cur_p);
for (int i = 0; i < 3; ++i) { for (int i = 0; i < 3; ++i) {
LOG(" - draft candidate %d: %d (%.3f)\n", i, cur_p.data[i].id, cur_p.data[i].p); LOG(" - draft candidate %3d: %6d (%8.3f) '%s'\n", i, cur_p.data[i].id, cur_p.data[i].p, llama_token_to_piece(ctx_dft, cur_p.data[i].id).c_str());
} }
// too low probability, stop drafting // too low probability, stop drafting