MCPcopy Create free account
hub / github.com/FastFlowLM/FastFlowLM / verbose

Method verbose

src/common/AutoModel/automodel.cpp:555–567  ·  view source on GitHub ↗

\brief Verbose \note The function will verbose \note The function will print the verbose

Source from the content-addressed store, hash-verified

553/// \note The function will verbose
554/// \note The function will print the verbose
555void AutoModel::verbose() {
556 std::cout << std::endl;
557 int total_tokens = this->get_current_context_length();
558 float prefill_speed = this->profiler_list[PREFILL_TIME].get_average_speed();
559 float decoding_speed = this->profiler_list[DECODING_TIME].get_average_speed();
560 time_utils::time_with_unit ttft_time = this->profiler_list[TTFT_TIME].get_total_time();
561 float context_percentage = (float)total_tokens / (float)this->MAX_L * 100;
562 std::cout << "Verbose: " << std::endl;
563 std::cout << " Total tokens: " << total_tokens << " (" << std::fixed << std::setprecision(2) << context_percentage << "%)" << std::endl;
564 std::cout << " TTFT: " << ttft_time.first << " " << ttft_time.second << std::endl;
565 std::cout << " Prefill speed: " << std::fixed << std::setprecision(2) << prefill_speed << " tokens/s" << std::endl;
566 std::cout << " Decoding speed: " << std::fixed << std::setprecision(2) << decoding_speed << " tokens/s" << std::endl << std::endl;
567}
568
569/// \brief Set the top-k
570/// \param topk the top-k

Callers 1

runMethod · 0.80

Calls 3

get_average_speedMethod · 0.80
get_total_timeMethod · 0.80

Tested by

no test coverage detected