server : add --flash-attn usage output (#3152)

This commit adds the `--flash-attn` option to the usage output of the
server example.

The motivation for this change is that while it is possible to set this
option it is not printed in the usage output.

diff --git a/examples/server/server.cpp b/examples/server/server.cpp
index 14462707ef181d3677bc732f72f9f0d592f2e16c..bf81f792e64eb292940ce6034a34822b6b8bde44 100644 (file)
--- a/examples/server/server.cpp
+++ b/examples/server/server.cpp
@@ -139,6 +139,7 @@ void whisper_print_usage(int /*argc*/, char ** argv, const whisper_params & para
     fprintf(stderr, "  -nth N,    --no-speech-thold N [%-7.2f] no speech threshold\n",   params.no_speech_thold);
     fprintf(stderr, "  -nc,       --no-context        [%-7s] do not use previous audio context\n", params.no_context ? "true" : "false");
     fprintf(stderr, "  -ng,       --no-gpu            [%-7s] do not use gpu\n", params.use_gpu ? "false" : "true");
+    fprintf(stderr, "  -fa,       --flash-attn        [%-7s] flash attention\n", params.flash_attn ? "true" : "false");
     fprintf(stderr, "\n");
 }