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

server: fix llama_sample_top_k order

Browse source
This commit is contained in:
Wang Haoran(Robin) 2023-06-26 18:11:27 -07:00
parent b5c5c8e2b9
commit bc88fece87
1 changed files with 1 additions and 1 deletions
Download patch file Download diff file Expand all files Collapse all files

2
examples/server/server.cpp
View file

@ -378,10 +378,10 @@ struct llama_server_context {
} else { } else {
// Temperature sampling // Temperature sampling
size_t min_keep = std::max(1, n_probs); size_t min_keep = std::max(1, n_probs);
llama_sample_top_k(ctx, &candidates_p, top_k, min_keep);
llama_sample_tail_free(ctx, &candidates_p, tfs_z, min_keep); llama_sample_tail_free(ctx, &candidates_p, tfs_z, min_keep);
llama_sample_typical(ctx, &candidates_p, typical_p, min_keep); llama_sample_typical(ctx, &candidates_p, typical_p, min_keep);
llama_sample_top_p(ctx, &candidates_p, top_p, min_keep); llama_sample_top_p(ctx, &candidates_p, top_p, min_keep);
llama_sample_top_k(ctx, &candidates_p, top_k, min_keep);
llama_sample_temperature(ctx, &candidates_p, temp); llama_sample_temperature(ctx, &candidates_p, temp);
result.tok = llama_sample_token(ctx, &candidates_p); result.tok = llama_sample_token(ctx, &candidates_p);
} }