mikhail/llama.cpp-mtp-turboquant
1
0
Fork 0
Code Issues Pull Requests Actions 28 Packages Projects Releases Wiki Activity

logs : reduce (#23021)

Browse Source 
* logs : reduce

* args : fix envs

* server : fix build

* common : print verbosity level at start

* server : clean-up logs

* server : print prompt processing timings + sampling params

* minor : whitespaces
This commit is contained in:
Georgi Gerganov
2026-05-14 13:05:52 +03:00
committed by GitHub
parent 81b0d882ae
commit 67b2b7f2f2
17 changed files with 207 additions and 147 deletions
Download Patch File Download Diff File Expand all files Collapse all files
common/common.cpp
+22 -7
View File
@@ -366,15 +366,29 @@ void common_init() {
SetConsoleCP(CP_UTF8);
#endif
llama_log_set(common_log_default_callback, NULL);
common_log_set_prefix(common_log_main(), true);
common_log_set_timestamps(common_log_main(), true);
llama_log_set(common_log_default_callback, NULL);
}
void common_params_print_info(const common_params & params) {
#ifdef NDEBUG
const char * build_type = "";
#else
const char * build_type = " (debug)";
#endif
LOG_TRC("%s: build %d (%s) with %s for %s%s\n", __func__, llama_build_number(), llama_commit(), llama_compiler(), llama_build_target(), build_type);
LOG_DBG("build: %d (%s) with %s for %s%s\n", llama_build_number(), llama_commit(), llama_compiler(), llama_build_target(), build_type);
LOG_INF("log_info: verbosity = %d (adjust with the `-lv N` CLI arg)\n", common_log_get_verbosity_thold());
LOG_INF("device_info:\n");
for (size_t i = 0; i < ggml_backend_dev_count(); ++i) {
auto * dev = ggml_backend_dev_get(i);
size_t free, total;
ggml_backend_dev_memory(dev, &free, &total);
LOG_INF(" - %-8s: %s (%zu MiB, %zu MiB free)\n", ggml_backend_dev_name(dev), ggml_backend_dev_description(dev), total / 1024 / 1024, free / 1024 / 1024);
}
LOG_INF("%s\n", common_params_get_system_info(params).c_str());
}
std::string common_params_get_system_info(const common_params & params) {
@@ -1147,7 +1161,8 @@ common_init_result::common_init_result(common_params & params) :
auto cparams = common_context_params_to_llama(params);
if (params.fit_params) {
LOG_INF("%s: fitting params to device memory, for bugs during this step try to reproduce them with -fit off, or provide --verbose logs if the bug only occurs with -fit on\n", __func__);
LOG_INF("%s: fitting params to device memory ...\n", __func__);
LOG_INF("%s: (for bugs during this step try to reproduce them with -fit off, or provide --verbose logs if the bug only occurs with -fit on)\n", __func__);
common_fit_params(params.model.path.c_str(), &mparams, &cparams,
params.tensor_split,
params.tensor_buft_overrides.data(),
@@ -1196,7 +1211,7 @@ common_init_result::common_init_result(common_params & params) :
// initialize once
for (llama_token i = 0; i < llama_vocab_n_tokens(vocab); i++) {
if (llama_vocab_is_eog(vocab, i)) {
LOG_INF("%s: added %s logit bias = %f\n", __func__, common_token_to_piece(vocab, i).c_str(), -INFINITY);
LOG_TRC("%s: added %s logit bias = %f\n", __func__, common_token_to_piece(vocab, i).c_str(), -INFINITY);
params.sampling.logit_bias_eog.push_back({i, -INFINITY});
}
}
@@ -1209,12 +1224,12 @@ common_init_result::common_init_result(common_params & params) :
}
//if (params.sampling.penalty_last_n == -1) {
// LOG_INF("%s: setting penalty_last_n to ctx_size = %d\n", __func__, llama_n_ctx(lctx));
// LOG_TRC("%s: setting penalty_last_n to ctx_size = %d\n", __func__, llama_n_ctx(lctx));
// params.sampling.penalty_last_n = llama_n_ctx(lctx);
//}
//if (params.sampling.dry_penalty_last_n == -1) {
// LOG_INF("%s: setting dry_penalty_last_n to ctx_size = %d\n", __func__, llama_n_ctx(lctx));
// LOG_TRC("%s: setting dry_penalty_last_n to ctx_size = %d\n", __func__, llama_n_ctx(lctx));
// params.sampling.dry_penalty_last_n = llama_n_ctx(lctx);
//}
@@ -1422,7 +1437,7 @@ common_context_seq_rm_type common_context_can_seq_rm(llama_context * ctx) {
// try to remove the last tokens
if (!llama_memory_seq_rm(mem, 0, 1, -1)) {
LOG_WRN("%s: the context does not support partial sequence removal\n", __func__);
LOG_TRC("%s: the context does not support partial sequence removal\n", __func__);
res = COMMON_CONTEXT_SEQ_RM_TYPE_FULL;
goto done;
}