Mirrors/llama.cpp
1
0
Fork 0
mirror of https://github.com/ggerganov/llama.cpp.git synced 2024-12-26 14:20:31 +01:00
Code Issues Packages Projects Releases Wiki Activity

log : add CONT level for continuing previous log entry (#9610)

Browse Source
This commit is contained in:
Georgi Gerganov 2024-09-24 10:15:35 +03:00 committed by GitHub
parent 0aa15011e3
commit cea1486ecf
No known key found for this signature in database
GPG Key ID: B5690EEEBB952194
7 changed files with 29 additions and 24 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

2
common/log.cpp
View File

@ -82,7 +82,7 @@ struct gpt_log_entry {
}
}
if (level != GGML_LOG_LEVEL_NONE && prefix) {
if (level != GGML_LOG_LEVEL_NONE && level != GGML_LOG_LEVEL_CONT && prefix) {
if (timestamp) {
// [M.s.ms.us]
fprintf(fcur, "%s%d.%02d.%03d.%03d%s ",

2
common/log.h
View File

@ -83,8 +83,10 @@ void gpt_log_set_timestamps(struct gpt_log * log, bool timestamps); // w
#define LOG_WRN(...) LOG_TMPL(GGML_LOG_LEVEL_WARN, 0, __VA_ARGS__)
#define LOG_ERR(...) LOG_TMPL(GGML_LOG_LEVEL_ERROR, 0, __VA_ARGS__)
#define LOG_DBG(...) LOG_TMPL(GGML_LOG_LEVEL_DEBUG, LOG_DEFAULT_DEBUG, __VA_ARGS__)
#define LOG_CNT(...) LOG_TMPL(GGML_LOG_LEVEL_CONT, 0, __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_ERRV(verbosity, ...) LOG_TMPL(GGML_LOG_LEVEL_ERROR, verbosity, __VA_ARGS__)
#define LOG_DBGV(verbosity, ...) LOG_TMPL(GGML_LOG_LEVEL_DEBUG, verbosity, __VA_ARGS__)
#define LOG_CNTV(verbosity, ...) LOG_TMPL(GGML_LOG_LEVEL_CONT, verbosity, __VA_ARGS__)

14
examples/infill/infill.cpp
View File

@ -263,9 +263,9 @@ int main(int argc, char ** argv) {
if (params.n_keep > 0) {
LOG_INF("%s: static prompt based on n_keep: '", __func__);
for (int i = 0; i < params.n_keep; i++) {
LOG("%s", llama_token_to_piece(ctx, embd_inp[i]).c_str());
LOG_CNT("%s", llama_token_to_piece(ctx, embd_inp[i]).c_str());
}
LOG("'\n");
LOG_CNT("'\n");
}
LOG_INF("\n");
}
@ -306,8 +306,8 @@ int main(int argc, char ** argv) {
LOG_INF("generate: n_ctx = %d, n_batch = %d, n_predict = %d, n_keep = %d\n", n_ctx, params.n_batch, params.n_predict, params.n_keep);
LOG("\n");
LOG("\n##### Infill mode #####\n\n");
LOG_INF("\n");
LOG_INF("\n##### Infill mode #####\n\n");
if (params.interactive) {
const char *control_message;
if (params.multiline_input) {
@ -318,11 +318,11 @@ int main(int argc, char ** argv) {
" - To return control without starting a new line, end your input with '/'.\n"
" - If you want to submit another line, end your input with '\\'.\n";
}
LOG("== Running in interactive mode. ==\n");
LOG_INF("== Running in interactive mode. ==\n");
#if defined (__unix__) || (defined (__APPLE__) && defined (__MACH__)) || defined (_WIN32)
LOG( " - Press Ctrl+C to interject at any time.\n");
LOG_INF( " - Press Ctrl+C to interject at any time.\n");
#endif
LOG( "%s\n", control_message);
LOG_INF( "%s\n", control_message);
is_interacting = params.interactive_first;
}

28
examples/main/main.cpp
View File

@ -385,9 +385,9 @@ int main(int argc, char ** argv) {
if (params.n_keep > add_bos) {
LOG_INF("%s: static prompt based on n_keep: '", __func__);
for (int i = 0; i < params.n_keep; i++) {
LOG("%s", llama_token_to_piece(ctx, embd_inp[i]).c_str());
LOG_CNT("%s", llama_token_to_piece(ctx, embd_inp[i]).c_str());
}
LOG("'\n");
LOG_CNT("'\n");
}
LOG_INF("\n");
}
@ -409,40 +409,40 @@ int main(int argc, char ** argv) {
}
if (params.interactive) {
LOG("%s: interactive mode on.\n", __func__);
LOG_INF("%s: interactive mode on.\n", __func__);
if (!params.antiprompt.empty()) {
for (const auto & antiprompt : params.antiprompt) {
LOG("Reverse prompt: '%s'\n", antiprompt.c_str());
LOG_INF("Reverse prompt: '%s'\n", antiprompt.c_str());
if (params.verbose_prompt) {
auto tmp = ::llama_tokenize(ctx, antiprompt, false, true);
for (int i = 0; i < (int) tmp.size(); i++) {
LOG("%6d -> '%s'\n", tmp[i], llama_token_to_piece(ctx, tmp[i]).c_str());
LOG_INF("%6d -> '%s'\n", tmp[i], llama_token_to_piece(ctx, tmp[i]).c_str());
}
}
}
}
if (params.input_prefix_bos) {
LOG("Input prefix with BOS\n");
LOG_INF("Input prefix with BOS\n");
}
if (!params.input_prefix.empty()) {
LOG("Input prefix: '%s'\n", params.input_prefix.c_str());
LOG_INF("Input prefix: '%s'\n", params.input_prefix.c_str());
if (params.verbose_prompt) {
auto tmp = ::llama_tokenize(ctx, params.input_prefix, true, true);
for (int i = 0; i < (int) tmp.size(); i++) {
LOG("%6d -> '%s'\n", tmp[i], llama_token_to_piece(ctx, tmp[i]).c_str());
LOG_INF("%6d -> '%s'\n", tmp[i], llama_token_to_piece(ctx, tmp[i]).c_str());
}
}
}
if (!params.input_suffix.empty()) {
LOG("Input suffix: '%s'\n", params.input_suffix.c_str());
LOG_INF("Input suffix: '%s'\n", params.input_suffix.c_str());
if (params.verbose_prompt) {
auto tmp = ::llama_tokenize(ctx, params.input_suffix, false, true);
for (int i = 0; i < (int) tmp.size(); i++) {
LOG("%6d -> '%s'\n", tmp[i], llama_token_to_piece(ctx, tmp[i]).c_str());
LOG_INF("%6d -> '%s'\n", tmp[i], llama_token_to_piece(ctx, tmp[i]).c_str());
}
}
}
@ -474,7 +474,7 @@ int main(int argc, char ** argv) {
//GGML_ASSERT(n_ctx >= n_ctx_train * ga_n && "n_ctx must be at least n_ctx_train * grp_attn_n"); // NOLINT
LOG_INF("self-extend: n_ctx_train = %d, grp_attn_n = %d, grp_attn_w = %d\n", n_ctx_train, ga_n, ga_w);
}
LOG("\n");
LOG_INF("\n");
if (params.interactive) {
const char * control_message;
@ -486,11 +486,11 @@ int main(int argc, char ** argv) {
" - To return control without starting a new line, end your input with '/'.\n"
" - If you want to submit another line, end your input with '\\'.\n";
}
LOG("== Running in interactive mode. ==\n");
LOG_INF("== Running in interactive mode. ==\n");
#if defined (__unix__) || (defined (__APPLE__) && defined (__MACH__)) || defined (_WIN32)
LOG( " - Press Ctrl+C to interject at any time.\n");
LOG_INF( " - Press Ctrl+C to interject at any time.\n");
#endif
LOG( "%s\n", control_message);
LOG_INF( "%s\n", control_message);
is_interacting = params.interactive_first;
}

1
ggml/include/ggml.h
View File

@ -570,6 +570,7 @@ extern "C" {
GGML_LOG_LEVEL_WARN = 2,
GGML_LOG_LEVEL_ERROR = 3,
GGML_LOG_LEVEL_DEBUG = 4,
GGML_LOG_LEVEL_CONT = 5, // continue previous log
};
// this tensor...

2
src/llama-impl.h
View File

@ -28,6 +28,8 @@ void llama_log_callback_default(ggml_log_level level, const char * text, void *
#define LLAMA_LOG_INFO(...) llama_log_internal(GGML_LOG_LEVEL_INFO , __VA_ARGS__)
#define LLAMA_LOG_WARN(...) llama_log_internal(GGML_LOG_LEVEL_WARN , __VA_ARGS__)
#define LLAMA_LOG_ERROR(...) llama_log_internal(GGML_LOG_LEVEL_ERROR, __VA_ARGS__)
#define LLAMA_LOG_DEBUG(...) llama_log_internal(GGML_LOG_LEVEL_DEBUG, __VA_ARGS__)
#define LLAMA_LOG_CONT(...) llama_log_internal(GGML_LOG_LEVEL_CONT , __VA_ARGS__)
//
// helpers

4
src/llama.cpp
View File

@ -18671,9 +18671,9 @@ struct llama_model * llama_load_model_from_file(
unsigned percentage = (unsigned) (100 * progress);
while (percentage > *cur_percentage_p) {
*cur_percentage_p = percentage;
LLAMA_LOG(".");
LLAMA_LOG_CONT(".");
if (percentage >= 100) {
LLAMA_LOG("\n");
LLAMA_LOG_CONT("\n");
}
}
return true;