Skip to content

LLM inference in C/C++

C++ ggml

llama-bench : init params.offline (#25476)

Adrien Gallouët259f2e2a531a
diff --git a/tools/llama-bench/llama-bench.cpp b/tools/llama-bench/llama-bench.cpp
index 2695f58785..0756893881 100644
--- a/tools/llama-bench/llama-bench.cpp
+++ b/tools/llama-bench/llama-bench.cpp
@@ -520,6 +520,7 @@ static cmd_params parse_cmd_params(int argc, char ** argv) {
params.delay = cmd_params_defaults.delay;
params.progress = cmd_params_defaults.progress;
params.no_warmup = cmd_params_defaults.no_warmup;
+ params.offline = cmd_params_defaults.offline;
if (const char * env = getenv("HF_TOKEN")) {
params.hf_token = env;