update debug statements

Author: kherud

src/main/cpp/jllama.cpp

@@ -358,6 +358,8 @@ JNIEXPORT void JNICALL Java_de_kherud_llama_LlamaModel_loadModel(JNIEnv *env, jo
 
     auto *ctx_server = new server_context();
 
+    std::cout << "New model: " << ctx_server << std::endl;
+
     std::string c_params = parse_jstring(env, jparams);
     json json_params = json::parse(c_params);
     server_params_parse(json_params, params);
@@ -476,15 +478,12 @@ JNIEXPORT jint JNICALL Java_de_kherud_llama_LlamaModel_requestCompletion(JNIEnv
 {
     jlong server_handle = env->GetLongField(obj, f_model_pointer);
     auto *ctx_server = reinterpret_cast<server_context *>(server_handle); // NOLINT(*-no-int-to-ptr)
-
-    std::cout << "DEBUG " << 1 << std::endl;
+    std::cout << "Request completion: " << ctx_server << std::endl;
 
     std::string c_params = parse_jstring(env, jparams);
     json json_params = json::parse(c_params);
     const bool infill = json_params.contains("input_prefix") || json_params.contains("input_suffix");
 
-    std::cout << "DEBUG " << 2 << std::endl;
-
     if (json_params.value("use_chat_template", false))
     {
         json chat;
@@ -493,12 +492,8 @@ JNIEXPORT jint JNICALL Java_de_kherud_llama_LlamaModel_requestCompletion(JNIEnv
         json_params["prompt"] = format_chat(ctx_server->model, ctx_server->params.chat_template, chat);
     }
 
-    std::cout << "DEBUG " << 3 << std::endl;
-
     const int id_task = ctx_server->queue_tasks.get_new_id();
-    std::cout << "DEBUG " << 4 << std::endl;
     ctx_server->queue_results.add_waiting_task_id(id_task);
-    std::cout << "DEBUG " << 5 << std::endl;
     ctx_server->request_completion(id_task, -1, json_params, infill, false);
 
     return id_task;
@@ -509,7 +504,6 @@ JNIEXPORT jobject JNICALL Java_de_kherud_llama_LlamaModel_receiveCompletion(JNIE
     jlong server_handle = env->GetLongField(obj, f_model_pointer);
     auto *ctx_server = reinterpret_cast<server_context *>(server_handle); // NOLINT(*-no-int-to-ptr)
 
-    std::cout << "DEBUG " << 8 << std::endl;
     server_task_result result = ctx_server->queue_results.recv(id_task);
 
     if (result.error)
@@ -519,14 +513,12 @@ JNIEXPORT jobject JNICALL Java_de_kherud_llama_LlamaModel_receiveCompletion(JNIE
         env->ThrowNew(c_llama_error, response.c_str());
         return nullptr;
     }
-    std::cout << "DEBUG " << 9 << std::endl;
 
     std::string response = result.data["content"].get<std::string>();
     if (result.stop)
     {
         ctx_server->queue_results.remove_waiting_task_id(id_task);
     }
-    std::cout << "DEBUG " << 10 << std::endl;
 
     jobject o_probabilities = env->NewObject(c_hash_map, cc_hash_map);
     if (result.data.contains("completion_probabilities"))
@@ -547,7 +539,6 @@ JNIEXPORT jobject JNICALL Java_de_kherud_llama_LlamaModel_receiveCompletion(JNIE
             }
         }
     }
-    std::cout << "DEBUG " << 11 << std::endl;
 
     jbyteArray jbytes = parse_jbytes(env, response);
     return env->NewObject(c_output, cc_output, jbytes, o_probabilities, result.stop);

src/main/cpp/server.hpp

@@ -1661,8 +1661,6 @@ struct server_context
         task.embedding = embedding;
         task.type = SERVER_TASK_TYPE_COMPLETION;
 
-        std::cout << "DEBUG " << 6 << std::endl;
-
         // when a completion task's prompt array is not a singleton, we split it into multiple requests
         // otherwise, it's a single-prompt task, we actually queue it
         // if there's numbers in the prompt array it will be treated as an array of tokens
@@ -1696,8 +1694,6 @@ struct server_context
         {
             queue_tasks.post(task);
         }
-
-        std::cout << "DEBUG " << 7 << std::endl;
     }
 
     void request_cancel(int id_task)

src/main/cpp/utils.hpp

@@ -89,7 +89,6 @@ static const char *log_level_to_string(ggml_log_level level)
 static inline void server_log(ggml_log_level level, const char *function, int line, const char *message,
                               const json &extra)
 {
-    std::cout << "DEBUG LOG " << 1 << std::endl;
     std::stringstream ss_tid;
     ss_tid << std::this_thread::get_id();
 
@@ -120,43 +119,32 @@ static inline void server_log(ggml_log_level level, const char *function, int li
     }
     else
     {
-        std::cout << "DEBUG LOG " << 2 << std::endl;
         std::stringstream ss;
         ss << message;
 
-        std::cout << "DEBUG LOG " << 3 << std::endl;
         if (!extra.empty())
         {
-            std::cout << "DEBUG LOG " << 4 << std::endl;
             for (const auto &el : extra.items())
             {
                 const std::string value = el.value().dump(-1, ' ', false, json::error_handler_t::replace);
                 ss << " " << el.key() << "=" << value;
             }
-            std::cout << "DEBUG LOG " << 5 << std::endl;
         }
 
-        std::cout << "DEBUG LOG " << 6 << std::endl;
 #if SERVER_VERBOSE
         ss << " | ts " << time(nullptr) << " | tid " << ss_tid.str() << " | " << function << " line " << line;
 #endif
 
-        std::cout << "DEBUG LOG " << 7 << std::endl;
         const std::string str = ss.str();
         if (log_callback == nullptr)
         {
-            std::cout << "DEBUG LOG " << 8 << std::endl;
             printf("[%4s] %.*s\n", log_level_to_string(level), (int)str.size(), str.data());
-            std::cout << "DEBUG LOG " << 9 << std::endl;
         }
         else
         {
-            std::cout << "DEBUG LOG " << 10 << std::endl;
             log_callback(level, str.c_str(), nullptr);
-            std::cout << "DEBUG LOG " << 11 << std::endl;
         }
     }
-    std::cout << "DEBUG LOG " << 12 << std::endl;
     fflush(stdout);
 }
 

src/main/java/de/kherud/llama/LlamaModel.java

@@ -42,6 +42,7 @@ public class LlamaModel implements AutoCloseable {
 	 */
 	public LlamaModel(ModelParameters parameters) {
 		loadModel(parameters.toString());
+		System.out.println(ctx);
 	}
 
 	/**