Skip to content

LLM inference in C/C++

C++ ggml

server : add missing task parameters(adaptive_target, adaptive_decay) in generation_settings (#25830)

Tekin Ertekin2cfc7670ed10
diff --git a/tools/server/server-task.cpp b/tools/server/server-task.cpp
index f01780100c..1fd7cce27b 100644
--- a/tools/server/server-task.cpp
+++ b/tools/server/server-task.cpp
@@ -63,6 +63,8 @@ json task_params::to_json(bool only_metrics) const {
{"mirostat", sampling.mirostat},
{"mirostat_tau", sampling.mirostat_tau},
{"mirostat_eta", sampling.mirostat_eta},
+ {"adaptive_target", sampling.adaptive_target},
+ {"adaptive_decay", sampling.adaptive_decay},
{"max_tokens", n_predict},
{"n_predict", n_predict}, // TODO: deduplicate?
{"n_keep", n_keep},
@@ -114,6 +116,8 @@ json task_params::to_json(bool only_metrics) const {
{"mirostat", sampling.mirostat},
{"mirostat_tau", sampling.mirostat_tau},
{"mirostat_eta", sampling.mirostat_eta},
+ {"adaptive_target", sampling.adaptive_target},
+ {"adaptive_decay", sampling.adaptive_decay},
{"stop", antiprompt},
{"max_tokens", n_predict},
{"n_predict", n_predict}, // TODO: deduplicate?
diff --git a/tools/server/tests/unit/test_completion.py b/tools/server/tests/unit/test_completion.py
index 1e0891987a..9375e0110e 100644
--- a/tools/server/tests/unit/test_completion.py
+++ b/tools/server/tests/unit/test_completion.py
@@ -66,6 +66,8 @@ def test_completion_stream(prompt: str, n_predict: int, re_content: str, n_promp
assert server.n_predict is not None
assert data["generation_settings"]["n_predict"] == min(n_predict, server.n_predict)
assert data["generation_settings"]["seed"] == server.seed
+ assert "adaptive_target" in data["generation_settings"]
+ assert "adaptive_decay" in data["generation_settings"]
assert match_regex(re_content, content)
else:
assert len(data["tokens"]) > 0