tqcq/llama.cpp
0
0
Fork 0
mirror of https://github.com/ggml-org/llama.cpp.git synced 2025-07-28 03:55:06 -04:00
Code Issues Packages Projects Releases Wiki Activity

Adding logprobs to /v1/completions (#11344)

Browse Source 
Signed-off-by: Jiri Podivin <jpodivin@redhat.com>
This commit is contained in:
Jiří Podivín
2025-01-22 12:51:32 +01:00
committed by GitHub
parent a94f3b2727
commit 96f4053934
1 changed files with 5 additions and 0 deletions
Download Patch File Download Diff File Expand all files Collapse all files

5
examples/server/server.cpp
View File

@ -267,6 +267,11 @@ struct server_task {
params.speculative.n_min = std::max(params.speculative.n_min, 2); params.speculative.n_min = std::max(params.speculative.n_min, 2);
params.speculative.n_max = std::max(params.speculative.n_max, 0); params.speculative.n_max = std::max(params.speculative.n_max, 0);
// Use OpenAI API logprobs only if n_probs wasn't provided
if (data.contains("logprobs") && params.sampling.n_probs == defaults.sampling.n_probs){
params.sampling.n_probs = json_value(data, "logprobs", defaults.sampling.n_probs);
}
if (data.contains("lora")) { if (data.contains("lora")) {
if (data.at("lora").is_array()) { if (data.at("lora").is_array()) {
params.lora = parse_lora_request(params_base.lora_adapters, data.at("lora")); params.lora = parse_lora_request(params_base.lora_adapters, data.at("lora"));