model_name from llama_bench PR

Author: JohannesGaessler

examples/common.cpp

@@ -735,7 +735,7 @@ std::tuple<struct llama_model *, struct llama_context *> llama_init_from_gpt_par
 }
 
 void dump_non_result_info_yaml(FILE * stream, const gpt_params & params, const std::string & timestamp,
-                               const std::vector<int> & prompt_tokens) {
+                               const std::vector<int> & prompt_tokens, const char * model) {
     fprintf(stream, "build_commit: %s\n", BUILD_COMMIT);
     fprintf(stream, "build_number: %d\n", BUILD_NUMBER);
     fprintf(stream, "cpu_has_arm_fma: %s\n", ggml_cpu_has_arm_fma() ? "true" : "false");
@@ -763,6 +763,8 @@ void dump_non_result_info_yaml(FILE * stream, const gpt_params & params, const s
     fprintf(stream, "debug: true\n");
 #endif // NDEBUG
 
+    fprintf(stream, "model: %s\n", model);
+
 #ifdef __OPTIMIZE__
     fprintf(stream, "optimize: true\n");
 #else

examples/common.h

@@ -115,4 +115,6 @@ std::vector<llama_token> llama_tokenize(struct llama_context * ctx, const std::s
 std::tuple<struct llama_model *, struct llama_context *> llama_init_from_gpt_params(const gpt_params & params);
 struct llama_context_params llama_context_params_from_gpt_params(const gpt_params & params);
 
-void dump_non_result_info_yaml(FILE * stream, const gpt_params & params, const std::string & timestamp, const std::vector<int> & prompt_tokens);
+void dump_non_result_info_yaml(
+    FILE * stream, const gpt_params & params, const std::string & timestamp,
+    const std::vector<int> & prompt_tokens, const char * model);

examples/main/main.cpp

@@ -835,17 +835,16 @@ int main(int argc, char ** argv) {
 
             FILE * logfile = fopen((params.logdir + timestamp + ".yml").c_str(), "w");
             fprintf(logfile, "binary: main\n");
-            dump_non_result_info_yaml(logfile, params, timestamp, input_tokens);
+            char model_type[128];
+            llama_model_type(model, model_type, sizeof(model_type));
+            dump_non_result_info_yaml(logfile, params, timestamp, input_tokens, model_type);
 
             fprintf(logfile, "\n");
             fprintf(logfile, "######################\n");
             fprintf(logfile, "# Generation Results #\n");
             fprintf(logfile, "######################\n");
             fprintf(logfile, "\n");
 
-            // fprintf(logfile, "ftype: %u\n", ctx->model.hparams.ftype);
-            // fprintf(logfile, "ftype_str: %s\n", llama_ftype_name(ctx->model.hparams.ftype));
-            // fprintf(logfile, "model_type: %s\n", llama_model_type_name(ctx->model.type));
             dump_string_yaml_multiline(logfile, "output", output_ss.str().c_str(), false);
             dump_vector_int_yaml(logfile, "output_tokens", output_tokens);
 

examples/perplexity/perplexity.cpp

@@ -368,17 +368,16 @@ int main(int argc, char ** argv) {
 
             FILE * logfile = fopen((params.logdir + timestamp + ".yml").c_str(), "w");
             fprintf(logfile, "binary: perplexity\n");
-            dump_non_result_info_yaml(logfile, params, timestamp, tokens);
+            char model_type[128];
+            llama_model_type(model, model_type, sizeof(model_type));
+            dump_non_result_info_yaml(logfile, params, timestamp, tokens, model_type);
 
             fprintf(logfile, "\n");
             fprintf(logfile, "######################\n");
             fprintf(logfile, "# Perplexity Results #\n");
             fprintf(logfile, "######################\n");
             fprintf(logfile, "\n");
 
-            // fprintf(logfile, "ftype: %u\n", ctx->model.hparams.ftype);
-            // fprintf(logfile, "ftype_str: %s\n", llama_ftype_name(ctx->model.hparams.ftype));
-            // fprintf(logfile, "model_type: %s\n", llama_model_type_name(ctx->model.type));
             dump_vector_float_yaml(logfile, "probs", probs);
 
             llama_dump_timing_info_yaml(logfile, ctx);