Skip to content

LLM inference in C/C++

C++ ggml

completion : remove useless statics (#24226)

Adrien Gallouët6b80c74f2853
diff --git a/tools/completion/completion.cpp b/tools/completion/completion.cpp
index 2a957963f8..6747558fc5 100644
--- a/tools/completion/completion.cpp
+++ b/tools/completion/completion.cpp
@@ -33,12 +33,8 @@
#endif
static llama_context ** g_ctx;
-static llama_model ** g_model;
static common_sampler ** g_smpl;
static common_params * g_params;
-static std::vector<llama_token> * g_input_tokens;
-static std::ostringstream * g_output_ss;
-static std::vector<llama_token> * g_output_tokens;
static bool is_interacting = false;
static bool need_insert_eot = false;
@@ -136,7 +132,6 @@ int llama_completion(int argc, char ** argv) {
llama_context * ctx = nullptr;
common_sampler * smpl = nullptr;
- g_model = &model;
g_ctx = &ctx;
g_smpl = &smpl;
@@ -549,9 +544,9 @@ int llama_completion(int argc, char ** argv) {
int n_consumed = 0;
int n_session_consumed = 0;
- std::vector<int> input_tokens; g_input_tokens = &input_tokens;
- std::vector<int> output_tokens; g_output_tokens = &output_tokens;
- std::ostringstream output_ss; g_output_ss = &output_ss;
+ std::vector<int> input_tokens;
+ std::vector<int> output_tokens;
+ std::ostringstream output_ss;
std::ostringstream assistant_ss; // for storing current assistant message, used in conversation mode
// the first thing we will do is to output the prompt, so set color accordingly