vbatts/llama.cpp
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity Actions

prefer const reference parameters (clang-tidy)

Browse source 
This is recommended by the 'performance-unnecessary-value-param' check.
This commit is contained in:
Cebtenzzre 2023-09-05 18:42:19 -04:00
parent aae2be0f08
commit 0b0fe663f8
4 changed files with 8 additions and 7 deletions
Download patch file Download diff file Expand all files Collapse all files

5
examples/main/main.cpp
View file

@ -48,8 +48,9 @@ static bool is_interacting = false;
void write_logfile( void write_logfile(
const llama_context * ctx, const gpt_params & params, const llama_model * model, const llama_context * ctx, const gpt_params & params, const llama_model * model,
const std::vector<llama_token> input_tokens, const std::string output, const std::vector<llama_token> output_tokens) { const std::vector<llama_token> & input_tokens, const std::string & output,
const std::vector<llama_token> & output_tokens
) {
if (params.logdir.empty()) { if (params.logdir.empty()) {
return; return;
} }

2
examples/quantize-stats/quantize-stats.cpp
View file

@ -71,7 +71,7 @@ void quantize_stats_print_usage(int /*argc*/, char ** argv) {
} }
// Check if a layer is included/excluded by command line // Check if a layer is included/excluded by command line
bool layer_included(const quantize_stats_params params, const std::string & layer) { bool layer_included(const quantize_stats_params & params, const std::string & layer) {
for (const auto& excluded : params.exclude_layers) { for (const auto& excluded : params.exclude_layers) {
if (std::regex_search(layer, std::regex(excluded))) { if (std::regex_search(layer, std::regex(excluded))) {
return false; return false;

6
examples/server/server.cpp
View file

@ -139,7 +139,7 @@ static std::string tokens_to_output_formatted_string(const llama_context *ctx, c
} }
// convert a vector of completion_token_output to json // convert a vector of completion_token_output to json
static json probs_vector_to_json(const llama_context *ctx, const std::vector<completion_token_output> probs) static json probs_vector_to_json(const llama_context *ctx, const std::vector<completion_token_output> & probs)
{ {
json out = json::array(); json out = json::array();
for (const auto &prob : probs) for (const auto &prob : probs)
@ -271,7 +271,7 @@ struct llama_server_context
return true; return true;
} }
std::vector<llama_token> tokenize(json json_prompt, bool add_bos) std::vector<llama_token> tokenize(const json & json_prompt, bool add_bos) const
{ {
// If `add_bos` is true, we only add BOS, when json_prompt is a string, // If `add_bos` is true, we only add BOS, when json_prompt is a string,
// or the first element of the json_prompt array is a string. // or the first element of the json_prompt array is a string.
@ -1255,7 +1255,7 @@ void beam_search_callback(void * callback_data, llama_beams_state beams_state) {
struct token_translator { struct token_translator {
llama_context * ctx; llama_context * ctx;
std::string operator()(llama_token tok) const { return llama_token_to_piece(ctx, tok); } std::string operator()(llama_token tok) const { return llama_token_to_piece(ctx, tok); }
std::string operator()(completion_token_output cto) const { return (*this)(cto.tok); } std::string operator()(const completion_token_output & cto) const { return (*this)(cto.tok); }
}; };
void append_to_generated_text_from_generated_token_probs(llama_server_context & llama) { void append_to_generated_text_from_generated_token_probs(llama_server_context & llama) {

2
tests/test-quantize-perf.cpp
View file

@ -76,7 +76,7 @@ void * align_with_offset(void * ptr, int offset) {
return (char *) std::align(MAX_ALIGNMENT, MAX_ALIGNMENT, ptr, dummy_size) + offset; return (char *) std::align(MAX_ALIGNMENT, MAX_ALIGNMENT, ptr, dummy_size) + offset;
} }
void benchmark_function(size_t size, size_t q_size, int64_t iterations, std::function<size_t(void)> function) { void benchmark_function(size_t size, size_t q_size, int64_t iterations, const std::function<size_t(void)> & function) {
int64_t min_time_us = INT64_MAX; int64_t min_time_us = INT64_MAX;
int64_t total_time_us = 0; int64_t total_time_us = 0;
int64_t min_time_cycles = INT64_MAX; int64_t min_time_cycles = INT64_MAX;