happyz
/
llama.cpp
mirror of https://github.com/ggerganov/llama.cpp.git
1
0
Fork 0
Code Issues Packages Projects Releases Wiki Activity

common: improve verbosity level definitions (#17630) 

* common: improve verbosity level definitions

* string_format

* update autogen docs
This commit is contained in:
Xuan-Son Nguyen 2025-12-01 14:38:13 +01:00 committed by GitHub
parent cd3c118908
commit 7733409734
No known key found for this signature in database
GPG Key ID: B5690EEEBB952194
6 changed files with 46 additions and 19 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

8
common/arg.cpp
View File

@ -2674,7 +2674,13 @@ common_params_context common_params_parser_init(common_params & params, llama_ex
).set_env("LLAMA_OFFLINE")); ).set_env("LLAMA_OFFLINE"));
add_opt(common_arg( add_opt(common_arg(
{"-lv", "--verbosity", "--log-verbosity"}, "N", {"-lv", "--verbosity", "--log-verbosity"}, "N",
"Set the verbosity threshold. Messages with a higher verbosity will be ignored.", string_format("Set the verbosity threshold. Messages with a higher verbosity will be ignored. Values:\n"
" - 0: generic output\n"
" - 1: error\n"
" - 2: warning\n"
" - 3: info\n"
" - 4: debug\n"
"(default: %d)\n", params.verbosity),
[](common_params & params, int value) { [](common_params & params, int value) {
params.verbosity = value; params.verbosity = value;
common_log_set_verbosity_thold(value); common_log_set_verbosity_thold(value);

2
common/common.h
View File

@ -369,7 +369,7 @@ struct common_params {
std::vector<common_control_vector_load_info> control_vectors; // control vector with user defined scale std::vector<common_control_vector_load_info> control_vectors; // control vector with user defined scale
int32_t verbosity = 0; int32_t verbosity = 3; // LOG_LEVEL_INFO
int32_t control_vector_layer_start = -1; // layer range for control vector int32_t control_vector_layer_start = -1; // layer range for control vector
int32_t control_vector_layer_end = -1; // layer range for control vector int32_t control_vector_layer_end = -1; // layer range for control vector
bool offline = false; bool offline = false;

2
common/download.cpp
View File

@ -430,7 +430,7 @@ std::pair<long, std::vector<char>> common_remote_get_content(const std::string &
curl_easy_setopt(curl.get(), CURLOPT_URL, url.c_str()); curl_easy_setopt(curl.get(), CURLOPT_URL, url.c_str());
curl_easy_setopt(curl.get(), CURLOPT_NOPROGRESS, 1L); curl_easy_setopt(curl.get(), CURLOPT_NOPROGRESS, 1L);
curl_easy_setopt(curl.get(), CURLOPT_FOLLOWLOCATION, 1L); curl_easy_setopt(curl.get(), CURLOPT_FOLLOWLOCATION, 1L);
curl_easy_setopt(curl.get(), CURLOPT_VERBOSE, 1L); curl_easy_setopt(curl.get(), CURLOPT_VERBOSE, 0L);
typedef size_t(*CURLOPT_WRITEFUNCTION_PTR)(void * ptr, size_t size, size_t nmemb, void * data); typedef size_t(*CURLOPT_WRITEFUNCTION_PTR)(void * ptr, size_t size, size_t nmemb, void * data);
auto write_callback = [](void * ptr, size_t size, size_t nmemb, void * data) -> size_t { auto write_callback = [](void * ptr, size_t size, size_t nmemb, void * data) -> size_t {
auto data_vec = static_cast<std::vector<char> *>(data); auto data_vec = static_cast<std::vector<char> *>(data);

16
common/log.cpp
View File

@ -443,8 +443,22 @@ void common_log_set_timestamps(struct common_log * log, bool timestamps) {
log->set_timestamps(timestamps); log->set_timestamps(timestamps);
} }
static int common_get_verbosity(enum ggml_log_level level) {
switch (level) {
case GGML_LOG_LEVEL_DEBUG: return LOG_LEVEL_DEBUG;
case GGML_LOG_LEVEL_INFO: return LOG_LEVEL_INFO;
case GGML_LOG_LEVEL_WARN: return LOG_LEVEL_WARN;
case GGML_LOG_LEVEL_ERROR: return LOG_LEVEL_ERROR;
case GGML_LOG_LEVEL_CONT: return LOG_LEVEL_INFO; // same as INFO
case GGML_LOG_LEVEL_NONE:
default:
return LOG_LEVEL_OUTPUT;
}
}
void common_log_default_callback(enum ggml_log_level level, const char * text, void * /*user_data*/) { void common_log_default_callback(enum ggml_log_level level, const char * text, void * /*user_data*/) {
if (LOG_DEFAULT_LLAMA <= common_log_verbosity_thold) { auto verbosity = common_get_verbosity(level);
if (verbosity <= common_log_verbosity_thold) {
common_log_add(common_log_main(), level, "%s", text); common_log_add(common_log_main(), level, "%s", text);
} }
} }

29
common/log.h
View File

@ -21,8 +21,14 @@
# define LOG_ATTRIBUTE_FORMAT(...) __attribute__((format(printf, __VA_ARGS__))) # define LOG_ATTRIBUTE_FORMAT(...) __attribute__((format(printf, __VA_ARGS__)))
#endif #endif
#define LOG_DEFAULT_DEBUG 1 #define LOG_LEVEL_DEBUG 4
#define LOG_DEFAULT_LLAMA 0 #define LOG_LEVEL_INFO 3
#define LOG_LEVEL_WARN 2
#define LOG_LEVEL_ERROR 1
#define LOG_LEVEL_OUTPUT 0 // output data from tools
#define LOG_DEFAULT_DEBUG LOG_LEVEL_DEBUG
#define LOG_DEFAULT_LLAMA LOG_LEVEL_INFO
enum log_colors { enum log_colors {
LOG_COLORS_AUTO = -1, LOG_COLORS_AUTO = -1,
@ -67,10 +73,11 @@ void common_log_add(struct common_log * log, enum ggml_log_level level, const ch
// 0.00.090.578 I llm_load_tensors: offloading 32 repeating layers to GPU // 0.00.090.578 I llm_load_tensors: offloading 32 repeating layers to GPU
// 0.00.090.579 I llm_load_tensors: offloading non-repeating layers to GPU // 0.00.090.579 I llm_load_tensors: offloading non-repeating layers to GPU
// //
// I - info (stdout, V = 0)
// W - warning (stderr, V = 0)
// E - error (stderr, V = 0)
// D - debug (stderr, V = LOG_DEFAULT_DEBUG) // D - debug (stderr, V = LOG_DEFAULT_DEBUG)
// I - info (stdout, V = LOG_DEFAULT_INFO)
// W - warning (stderr, V = LOG_DEFAULT_WARN)
// E - error (stderr, V = LOG_DEFAULT_ERROR)
// O - output (stdout, V = LOG_DEFAULT_OUTPUT)
// //
void common_log_set_file (struct common_log * log, const char * file); // not thread-safe void common_log_set_file (struct common_log * log, const char * file); // not thread-safe
@ -95,14 +102,14 @@ void common_log_set_timestamps(struct common_log * log, bool timestamps); // w
} \ } \
} while (0) } while (0)
#define LOG(...) LOG_TMPL(GGML_LOG_LEVEL_NONE, 0, __VA_ARGS__) #define LOG(...) LOG_TMPL(GGML_LOG_LEVEL_NONE, LOG_LEVEL_OUTPUT, __VA_ARGS__)
#define LOGV(verbosity, ...) LOG_TMPL(GGML_LOG_LEVEL_NONE, verbosity, __VA_ARGS__) #define LOGV(verbosity, ...) LOG_TMPL(GGML_LOG_LEVEL_NONE, verbosity, __VA_ARGS__)
#define LOG_INF(...) LOG_TMPL(GGML_LOG_LEVEL_INFO, 0, __VA_ARGS__) #define LOG_DBG(...) LOG_TMPL(GGML_LOG_LEVEL_DEBUG, LOG_LEVEL_DEBUG, __VA_ARGS__)
#define LOG_WRN(...) LOG_TMPL(GGML_LOG_LEVEL_WARN, 0, __VA_ARGS__) #define LOG_INF(...) LOG_TMPL(GGML_LOG_LEVEL_INFO, LOG_LEVEL_INFO, __VA_ARGS__)
#define LOG_ERR(...) LOG_TMPL(GGML_LOG_LEVEL_ERROR, 0, __VA_ARGS__) #define LOG_WRN(...) LOG_TMPL(GGML_LOG_LEVEL_WARN, LOG_LEVEL_WARN, __VA_ARGS__)
#define LOG_DBG(...) LOG_TMPL(GGML_LOG_LEVEL_DEBUG, LOG_DEFAULT_DEBUG, __VA_ARGS__) #define LOG_ERR(...) LOG_TMPL(GGML_LOG_LEVEL_ERROR, LOG_LEVEL_ERROR, __VA_ARGS__)
#define LOG_CNT(...) LOG_TMPL(GGML_LOG_LEVEL_CONT, 0, __VA_ARGS__) #define LOG_CNT(...) LOG_TMPL(GGML_LOG_LEVEL_CONT, LOG_LEVEL_INFO, __VA_ARGS__) // same as INFO
#define LOG_INFV(verbosity, ...) LOG_TMPL(GGML_LOG_LEVEL_INFO, verbosity, __VA_ARGS__) #define LOG_INFV(verbosity, ...) LOG_TMPL(GGML_LOG_LEVEL_INFO, verbosity, __VA_ARGS__)
#define LOG_WRNV(verbosity, ...) LOG_TMPL(GGML_LOG_LEVEL_WARN, verbosity, __VA_ARGS__) #define LOG_WRNV(verbosity, ...) LOG_TMPL(GGML_LOG_LEVEL_WARN, verbosity, __VA_ARGS__)

6
tools/server/README.md
View File

@ -52,7 +52,7 @@ The project is under active development, and we are [looking for feedback and co
| `-ub, --ubatch-size N` | physical maximum batch size (default: 512)<br/>(env: LLAMA_ARG_UBATCH) | | `-ub, --ubatch-size N` | physical maximum batch size (default: 512)<br/>(env: LLAMA_ARG_UBATCH) |
| `--keep N` | number of tokens to keep from the initial prompt (default: 0, -1 = all) | | `--keep N` | number of tokens to keep from the initial prompt (default: 0, -1 = all) |
| `--swa-full` | use full-size SWA cache (default: false)<br/>[(more info)](https://github.com/ggml-org/llama.cpp/pull/13194#issuecomment-2868343055)<br/>(env: LLAMA_ARG_SWA_FULL) | | `--swa-full` | use full-size SWA cache (default: false)<br/>[(more info)](https://github.com/ggml-org/llama.cpp/pull/13194#issuecomment-2868343055)<br/>(env: LLAMA_ARG_SWA_FULL) |
| `--kv-unified, -kvu` | use single unified KV buffer for the KV cache of all sequences (default: false)<br/>[(more info)](https://github.com/ggml-org/llama.cpp/pull/14363)<br/>(env: LLAMA_ARG_KV_SPLIT) | | `--kv-unified, -kvu` | use single unified KV buffer for the KV cache of all sequences (default: false)<br/>[(more info)](https://github.com/ggml-org/llama.cpp/pull/14363)<br/>(env: LLAMA_ARG_KV_UNIFIED) |
| `-fa, --flash-attn [on\|off\|auto]` | set Flash Attention use ('on', 'off', or 'auto', default: 'auto')<br/>(env: LLAMA_ARG_FLASH_ATTN) | | `-fa, --flash-attn [on\|off\|auto]` | set Flash Attention use ('on', 'off', or 'auto', default: 'auto')<br/>(env: LLAMA_ARG_FLASH_ATTN) |
| `--no-perf` | disable internal libllama performance timings (default: false)<br/>(env: LLAMA_ARG_NO_PERF) | | `--no-perf` | disable internal libllama performance timings (default: false)<br/>(env: LLAMA_ARG_NO_PERF) |
| `-e, --escape` | process escapes sequences (\n, \r, \t, \', \", \\) (default: true) | | `-e, --escape` | process escapes sequences (\n, \r, \t, \', \", \\) (default: true) |
@ -103,11 +103,11 @@ The project is under active development, and we are [looking for feedback and co
| `-hffv, --hf-file-v FILE` | Hugging Face model file for the vocoder model (default: unused)<br/>(env: LLAMA_ARG_HF_FILE_V) | | `-hffv, --hf-file-v FILE` | Hugging Face model file for the vocoder model (default: unused)<br/>(env: LLAMA_ARG_HF_FILE_V) |
| `-hft, --hf-token TOKEN` | Hugging Face access token (default: value from HF_TOKEN environment variable)<br/>(env: HF_TOKEN) | | `-hft, --hf-token TOKEN` | Hugging Face access token (default: value from HF_TOKEN environment variable)<br/>(env: HF_TOKEN) |
| `--log-disable` | Log disable | | `--log-disable` | Log disable |
| `--log-file FNAME` | Log to file | | `--log-file FNAME` | Log to file<br/>(env: LLAMA_LOG_FILE) |
| `--log-colors [on\|off\|auto]` | Set colored logging ('on', 'off', or 'auto', default: 'auto')<br/>'auto' enables colors when output is to a terminal<br/>(env: LLAMA_LOG_COLORS) | | `--log-colors [on\|off\|auto]` | Set colored logging ('on', 'off', or 'auto', default: 'auto')<br/>'auto' enables colors when output is to a terminal<br/>(env: LLAMA_LOG_COLORS) |
| `-v, --verbose, --log-verbose` | Set verbosity level to infinity (i.e. log all messages, useful for debugging) | | `-v, --verbose, --log-verbose` | Set verbosity level to infinity (i.e. log all messages, useful for debugging) |
| `--offline` | Offline mode: forces use of cache, prevents network access<br/>(env: LLAMA_OFFLINE) | | `--offline` | Offline mode: forces use of cache, prevents network access<br/>(env: LLAMA_OFFLINE) |
| `-lv, --verbosity, --log-verbosity N` | Set the verbosity threshold. Messages with a higher verbosity will be ignored.<br/>(env: LLAMA_LOG_VERBOSITY) | | `-lv, --verbosity, --log-verbosity N` | Set the verbosity threshold. Messages with a higher verbosity will be ignored. Values:<br/> - 0: generic output<br/> - 1: error<br/> - 2: warning<br/> - 3: info<br/> - 4: debug<br/>(default: 3)<br/><br/>(env: LLAMA_LOG_VERBOSITY) |
| `--log-prefix` | Enable prefix in log messages<br/>(env: LLAMA_LOG_PREFIX) | | `--log-prefix` | Enable prefix in log messages<br/>(env: LLAMA_LOG_PREFIX) |
| `--log-timestamps` | Enable timestamps in log messages<br/>(env: LLAMA_LOG_TIMESTAMPS) | | `--log-timestamps` | Enable timestamps in log messages<br/>(env: LLAMA_LOG_TIMESTAMPS) |
| `-ctkd, --cache-type-k-draft TYPE` | KV cache data type for K for the draft model<br/>allowed values: f32, f16, bf16, q8_0, q4_0, q4_1, iq4_nl, q5_0, q5_1<br/>(default: f16)<br/>(env: LLAMA_ARG_CACHE_TYPE_K_DRAFT) | | `-ctkd, --cache-type-k-draft TYPE` | KV cache data type for K for the draft model<br/>allowed values: f32, f16, bf16, q8_0, q4_0, q4_1, iq4_nl, q5_0, q5_1<br/>(default: f16)<br/>(env: LLAMA_ARG_CACHE_TYPE_K_DRAFT) |