CS348Project/llama.cpp
5
0
Fork 0
mirror of https://github.com/ggml-org/llama.cpp.git synced 2025-10-31 08:51:55 +00:00
Code Issues Packages Projects Releases Wiki Activity

fix server sampling: top k sampler first (#1977)

Browse Source 
Co-authored-by: anon <anon@example.org>
This commit is contained in:
anon998
2023-06-25 08:48:36 +00:00
committed by GitHub
parent 66a2555ba6
commit c2a08f87b8
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

@@ -325,10 +325,10 @@ struct llama_server_context {
id = llama_sample_token_mirostat_v2(ctx, &candidates_p, mirostat_tau, mirostat_eta, &mirostat_mu);
} else {
// Temperature sampling
llama_sample_top_k(ctx, &candidates_p, top_k, 1);
llama_sample_tail_free(ctx, &candidates_p, tfs_z, 1);
llama_sample_typical(ctx, &candidates_p, typical_p, 1);
llama_sample_top_p(ctx, &candidates_p, top_p, 1);
llama_sample_top_k(ctx, &candidates_p, top_k, 1);
llama_sample_temperature(ctx, &candidates_p, temp);
id = llama_sample_token(ctx, &candidates_p);
}