first_token
Method
This is a method of the StreamingTracker class in the profiler module.Mark time to first token.
Guide
Reference
- execute
- execute_once
- get_stats
- start
- stop
- generate_crew_and_kickoff
- is_function_or_decorated
- load_tools_from_module
- load_tools_from_module_class
- load_tools_from_package
- load_tools_from_tools_py
- convert_and_save
- discover_tools_for_topic
- generate
- get_user_content
- merge_with_existing_agents
- recommend_pattern
- analyze_complexity
- get_available_tools
- create_api_file
- create_dockerfile
- run_commands
- set_environment_variables
- deploy
- deploy
- deploy_once
- set_deployer
- start
- stop
- get_imports
- get_slowest
- execute
- api_call
- block
- clear
- cprofile
- disable
- enable
- export_flamegraph
- export_html
- export_json
- export_to_file
- get_api_calls
- get_cprofile_stats
- get_files_accessed
- get_flamegraph_data
- get_flow
- get_imports
- get_line_profile_data
- get_memory_records
- get_statistics
- get_streaming_records
- get_summary
- get_timings
- is_enabled
- memory
- memory_snapshot
- record_api_call
- record_flow
- record_import
- record_memory
- record_streaming
- record_timing
- report
- set_line_profile_data
- streaming
- streaming_async
- parse
- chunk
- elapsed_ms
- end
- first_token
- start
- ttft_ms
- check_gpu
- check_ram
- create_and_push_ollama_model
- inference
- load_config
- load_datasets
- load_model
- prepare_model
- prepare_modelfile_content
- print_system_info
- process_dataset
- push_model_gguf
- run
- save_model_gguf
- save_model_merged
- tokenize_dataset
- train_model
- check_gpu
- check_ram
- convert_sample
- create_and_push_ollama_model
- load_config
- load_datasets
- prepare_model
- prepare_modelfile_content
- print_system_info
- push_model_gguf
- run
- save_model_gguf
- save_model_merged
- train_model
- vision_inference
- create_and_push_ollama_model
- load_config
- prepare_model
- prepare_modelfile_content
- push_model_gguf
- save_model_merged
- upload
- generate
- merge_with_existing_workflow
- recommend_pattern
- recommend_pattern_llm
- agent
- auth_callback
- check_module_available
- create_agent_scheduler
- create_scheduler
- disable_crewai_telemetry
- embedding
- formatting_prompts_func
- get_all_available_tools
- get_registered_agents
- get_tools_for_task
- main
- noop
- on_chat_resume
- on_modify
- on_run
- on_settings_update
- output
- profile
- profile_api
- profile_api_async
- profile_async
- profile_detailed
- profile_imports
- profile_lines
- recommend_agent_count
- run_agents
- safe_format
- sanitize_agent_name_for_autogen_v4
- set_profiles
- start_chat
- start_chat_server
- task
- time_import
- tokenize_function