model_analyzer
Triton Model Analyzer is a CLI tool to help with better understanding of the compute and memory requirements of the Triton Inference Server models.
파일 탐색기
최종 버전 다운로드 (.zip)- codeql.yml
- pre-commit.yml
- python-package.yaml
- bls_quick_start.md
- checkpoints.md
- cli.md
- config.md
- config_search.md
- ensemble_quick_start.md
- install.md
- kubernetes_deploy.md
- launch_modes.md
- metrics.md
- mm_quick_start.md
- model_types.md
- quick_start.md
- README.md
- report.md
- model.py
- config.pbtxt
- model.pt
- config.pbtxt
- output0_labels.txt
- model.py
- bls_input_data.json
- config.pbtxt
- .gitkeep
- config.pbtxt
- model.py
- config.pbtxt
- model.py
- config.pbtxt
- bls_input_data.json
- bls_result_summary_table.jpg
- bls_result_summary_top.jpg
- ensemble_result_summary_table.jpg
- ensemble_result_summary_top.jpg
- mm_result_summary_constraint_table.jpg
- mm_result_summary_constraint_top.jpg
- mm_result_summary_table.jpg
- mm_result_summary_top.jpg
- offline_detailed_report.pdf
- offline_summary.pdf
- online_detailed_report.pdf
- online_summary.pdf
- 0.ckpt
- config.pbtxt
- 0.ckpt
- config.pbtxt
- minimize_latency.yml
- sweep_checkpoints.py
- test_pa.py
- checkpoint_experiment_data.py
- config_command_experiment.py
- evaluate_config_generator.py
- experiment_config_command_creator.py
- experiment_data.py
- experiment_evaluator.py
- experiment_file_writer.py
- experiment_plot.py
- generator_experiment_factory.py
- main.py
- start_profiles.sh
- _helpers.tpl
- config-map.yaml
- job.yaml
- .helmignore
- Chart.yaml
- values.yaml
- __init__.py
- cli.py
- __init__.py
- automatic_model_config_generator.py
- base_model_config_generator.py
- brute_plus_binary_parameter_search_run_config_generator.py
- brute_run_config_generator.py
- concurrency_sweeper.py
- config_generator_interface.py
- coordinate.py
- coordinate_data.py
- generator_utils.py
- manual_model_config_generator.py
- model_config_generator_factory.py
- model_profile_spec.py
- model_run_config_generator.py
- model_variant_name_manager.py
- neighborhood.py
- optuna_plus_concurrency_sweep_run_config_generator.py
- optuna_run_config_generator.py
- perf_analyzer_config_generator.py
- quick_plus_concurrency_sweep_run_config_generator.py
- quick_run_config_generator.py
- run_config_generator_factory.py
- search_config.py
- search_dimension.py
- search_dimensions.py
- search_parameter.py
- search_parameters.py
- __init__.py
- config_model_profile_spec.py
- config_model_report_spec.py
- config_plot.py
- config_protobuf_utils.py
- __init__.py
- config_command.py
- config_command_profile.py
- config_command_report.py
- config_defaults.py
- config_enum.py
- config_field.py
- config_list_generic.py
- config_list_numeric.py
- config_list_string.py
- config_none.py
- config_object.py
- config_primitive.py
- config_status.py
- config_sweep.py
- config_union.py
- config_utils.py
- config_value.py
- yaml_config_validator.py
- __init__.py
- model_run_config.py
- run_config.py
- __init__.py
- __init__.py
- device.py
- gpu_device.py
- gpu_device_factory.py
- __init__.py
- dcgm_client_cli_parser.py
- dcgm_client_main.py
- __init__.py
- dcgm_agent.py
- dcgm_collectd_plugin.py
- dcgm_errors.py
- dcgm_field_helpers.py
- dcgm_fields.py
- dcgm_fields_collectd.py
- dcgm_fields_internal.py
- dcgm_fluentd.py
- dcgm_monitor.py
- dcgm_prometheus.py
- dcgm_structs.py
- dcgm_telegraf.py
- dcgm_value.py
- DcgmDiag.py
- DcgmFieldGroup.py
- DcgmGroup.py
- DcgmHandle.py
- DcgmJsonReader.py
- DcgmReader.py
- DcgmStatus.py
- DcgmSystem.py
- dcgmvalue.py
- denylist_recommendations.py
- pydcgm.py
- __init__.py
- cpu_monitor.py
- monitor.py
- remote_monitor.py
- __init__.py
- file_writer.py
- output_writer.py
- __init__.py
- genai_perf_config.py
- perf_analyzer.py
- perf_config.py
- __init__.py
- detailed_plot.py
- plot_manager.py
- simple_plot.py
- __init__.py
- cpu_available_ram.py
- cpu_used_ram.py
- gpu_free_memory.py
- gpu_power_usage.py
- gpu_total_memory.py
- gpu_used_memory.py
- gpu_utilization.py
- inter_token_latency_avg.py
- inter_token_latency_base.py
- inter_token_latency_max.py
- inter_token_latency_min.py
- inter_token_latency_p25.py
- inter_token_latency_p50.py
- inter_token_latency_p75.py
- inter_token_latency_p90.py
- inter_token_latency_p95.py
- inter_token_latency_p99.py
- output_token_throughput.py
- perf_client_response_wait.py
- perf_client_send_recv.py
- perf_latency.py
- perf_latency_avg.py
- perf_latency_base.py
- perf_latency_p90.py
- perf_latency_p95.py
- perf_latency_p99.py
- perf_server_compute_infer.py
- perf_server_compute_input.py
- perf_server_compute_output.py
- perf_server_queue.py
- perf_throughput.py
- time_to_first_token_avg.py
- time_to_first_token_base.py
- time_to_first_token_max.py
- time_to_first_token_min.py
- time_to_first_token_p25.py
- time_to_first_token_p50.py
- time_to_first_token_p75.py
- time_to_first_token_p90.py
- time_to_first_token_p95.py
- time_to_first_token_p99.py
- __init__.py
- gpu_record.py
- metrics_manager.py
- record.py
- record_aggregator.py
- __init__.py
- html_report.py
- pdf_report.py
- report.py
- report_factory.py
- report_manager.py
- report_utils.py
- __init__.py
- constraint_manager.py
- model_config_measurement.py
- model_constraints.py
- parameter_search.py
- result_manager.py
- result_statistics.py
- result_table.py
- result_table_manager.py
- result_utils.py
- results.py
- run_config_measurement.py
- run_config_result.py
- run_config_result_comparator.py
- sorted_results.py
- __init__.py
- analyzer_state.py
- analyzer_state_manager.py
- __init__.py
- client.py
- client_factory.py
- grpc_client.py
- http_client.py
- __init__.py
- model_config.py
- model_config_variant.py
- __init__.py
- server.py
- server_config.py
- server_docker.py
- server_factory.py
- server_local.py
- __init__.py
- __init__.py
- analyzer.py
- constants.py
- entrypoint.py
- log_formatter.py
- model_analyzer_exceptions.py
- model_manager.py
- check_analyzer_results.sh
- check_copyright.py
- util.sh
- check_results.py
- input_data.json
- test.sh
- test_config_generator.py
- config_generator.py
- test.sh
- config_generator.py
- test.sh
- test.sh
- test_config_generator.py
- test.sh
- test.sh
- test_config_generator.py
- check_results.py
- test.sh
- test_config_generator.py
- test.sh
- test_config_generator.py
- test.sh
- mkdocs.yml
- test.sh
- check_results.py
- config_composing_ranges.yml
- test.sh
- test_config_generator.py
- test.sh
- test_config_generator.py
- check_results.py
- test.sh
- test_config_generator.py
- test.sh
- test_config_generator.py
- check_results.py
- input_data.json
- test.sh
- test_config_generator.py
- check_results.py
- test.sh
- test_config_generator.py
- check_results.py
- test.sh
- test_config_generator.py
- check_results.py
- test.sh
- test_config_generator.py
- config_generator.py
- test.sh
- check_results.py
- test.sh
- test_config_generator.py
- check_results.py
- test.sh
- test_config_generator.py
- test.sh
- check_results.py
- test.sh
- test_config_generator.py
- check_results.py
- test.sh
- test_config_generator.py
- check_results.py
- test.sh
- test_config_generator.py
- test.sh
- check_results.py
- test.sh
- test_config_generator.py
- check_gpus.py
- test.sh
- test_config_generator.py
- broken_config.yml
- test.sh
- working_config.yml
- broken_config.yml
- nginx.conf
- test.sh
- working_config.yml
- check_results.py
- test.sh
- test_config_generator.py
- check_results.py
- test.sh
- test_config_generator.py
- check_results.py
- test.sh
- test_config_generator.py
- check_results.py
- test.sh
- test_config_generator.py
- check_results.py
- test.sh
- test_config_generator.py
- count_tests.py
- test.sh
- 0.ckpt
- golden-metrics-model-gpu.csv
- golden-metrics-model-inference.csv
- golden-metrics-server-only.csv
- README
- 0.ckpt
- golden-metrics-model-gpu.csv
- golden-metrics-model-inference.csv
- golden-metrics-server-only.csv
- README
- 0.ckpt
- golden-metrics-model-gpu.csv
- golden-metrics-model-inference.csv
- golden-metrics-server-only.csv
- plot_manager.json
- README
- 0.ckpt
- golden-metrics-model-gpu.csv
- golden-metrics-model-inference.csv
- golden-metrics-server-only.csv
- README
- 0.ckpt
- golden-metrics-model-gpu.csv
- golden-metrics-model-inference.csv
- golden-metrics-server-only.csv
- plot_manager.json
- README
- test_result_collector.py
- test_utils.py
- __init__.py
- mock_api_error.py
- mock_base.py
- mock_client.py
- mock_config.py
- mock_dcgm.py
- mock_dcgm_agent.py
- mock_dcgm_field_group_watcher.py
- mock_glob.py
- mock_gpu_device_factory.py
- mock_io.py
- mock_json.py
- mock_matplotlib.py
- mock_model_config.py
- mock_numba.py
- mock_os.py
- mock_perf_analyzer.py
- mock_psutil.py
- mock_requests.py
- mock_run_configs.py
- mock_server.py
- mock_server_docker.py
- mock_server_local.py
- __init__.py
- test_analyzer.py
- test_bls_report_manager.py
- test_cli.py
- test_config.py
- test_config_composing_model_validation.py
- test_constraint_manager.py
- test_coordinate.py
- test_coordinate_data.py
- test_cpu_monitor.py
- test_dcgm_monitor.py
- test_ensemble_composing_model_integration.py
- test_ensemble_report_manager.py
- test_extract_instance_group_kind.py
- test_file_writer.py
- test_html_report.py
- test_logger.py
- test_model_config.py
- test_model_config_generator.py
- test_model_config_measurement.py
- test_model_constraints.py
- test_model_manager.py
- test_model_variant_name_manager.py
- test_neighborhood.py
- test_optuna_run_config_generator.py
- test_parameter_search.py
- test_pdf_report.py
- test_perf_analyzer.py
- test_perf_analyzer_config_generator.py
- test_plot.py
- test_plot_manager.py
- test_quick_run_config_generator.py
- test_record_aggregator.py
- test_record_types.py
- test_remote_monitor.py
- test_report_manager.py
- test_report_utils.py
- test_result_manager.py
- test_result_table.py
- test_result_table_manager.py
- test_results.py
- test_run_config.py
- test_run_config_generator.py
- test_run_config_measurement.py
- test_run_config_result.py
- test_run_config_result_comparator.py
- test_search_config.py
- test_search_dimension.py
- test_search_dimensions.py
- test_search_dimensions_factory.py
- test_search_parameters.py
- test_sorted_results.py
- test_state_manager.py
- test_triton_client.py
- test_triton_server.py
- test_triton_server_factory.py
- test_yaml_options.py
- add_copyright.py
- COPYRIGHT
- .clang-format
- .gitattributes
- .gitignore
- .pre-commit-config.yaml
- CONTRIBUTING.md
- Dockerfile
- LICENSE
- nvidia_entrypoint.sh
- pyproject.toml
- README.md
- VERSION
# 설치 가이드
1. 코드 내려받기
git clone https://github.com/triton-inference-server/model_analyzer
깃허브에서 프로젝트 코드 전체를 내 컴퓨터로 내려받습니다.
cd model_analyzer
방금 내려받은 프로젝트 폴더 안으로 이동합니다.
2. Docker
쉬움 추천사전 준비물
- Git GitHub에서 프로젝트 코드를 내려받으려면 필요합니다.
- Docker Desktop 컨테이너를 빌드하고 실행하려면 필요합니다. 설치 후 실행해서 백그라운드에 켜두세요.
docker build -t model-analyzer .
Dockerfile을 기반으로 실행 가능한 이미지를 빌드합니다.
docker run -p 8080:80 model-analyzer
빌드된 이미지를 실제 컨테이너로 실행합니다.
터미널에 docker compose ps 를 입력해 컨테이너들이 Up 상태인지 확인하세요. README에 포트 번호가 적혀있다면 브라우저에서 http://localhost:포트번호 로 접속해보세요.
// repository documentation
Was this content helpful?
(0 ratings)
