Index

_

__init__() (controllers.benchmark_controller.BenchmarkController method) (controllers.direct_benchmark_controller.DirectBenchmarkController method) (controllers.docker_controller.DockerController method) (controllers.local_controller.LocalController method) (controllers.ome_controller.OMEController method) (orchestrator.AutotunerOrchestrator method) (utils.gpu_discovery.NodeGPUSummary method) (utils.gpu_monitor.GPUMonitor method) (utils.gpu_monitor.GPUSnapshot method) (utils.gpu_pool.GPUAllocation method) (utils.gpu_pool.GPUResourcePool method) (utils.optimizer.BayesianStrategy method) (utils.optimizer.GridSearchStrategy method) (utils.optimizer.OptimizationStrategy method) (utils.optimizer.RandomSearchStrategy method) (web.db.models.AgentEventSubscription method) (web.db.models.ChatMessage method) (web.db.models.ChatSession method) (web.db.models.Experiment method) (web.db.models.ParameterPreset method) (web.db.models.Task method)

__new__() (web.db.models.ExperimentStatus method) (web.db.models.MessageRole method) (web.db.models.TaskStatus method)

A

acquire() (utils.gpu_pool.GPUResourcePool method) AgentEventSubscription (class in web.db.models) allocatable_gpus (utils.gpu_discovery.NodeGPUSummary attribute) allocate_gpus() (utils.gpu_monitor.GPUMonitor method) allocated_at (utils.gpu_pool.GPUAllocation attribute)

ASSISTANT (web.db.models.MessageRole attribute) AutotunerOrchestrator (class in orchestrator) available_gpus (utils.gpu_monitor.GPUSnapshot attribute) avg_memory_usage (utils.gpu_discovery.NodeGPUSummary attribute) avg_utilization (utils.gpu_discovery.NodeGPUSummary attribute)

B

base_runtime (web.db.models.Task attribute) BaseModelController (class in controllers.base_controller) BayesianStrategy (class in utils.optimizer) benchmark_config (web.db.models.Task attribute)

BenchmarkController (class in controllers.benchmark_controller) BENCHMARKING (web.db.models.ExperimentStatus attribute) best_experiment (web.db.models.Task attribute) best_experiment_id (web.db.models.Task attribute)

C

calculate_objective_score() (in module utils.optimizer) calculate_slo_penalty() (in module utils.optimizer) CANCELLED (web.db.models.TaskStatus attribute) category (web.db.models.ParameterPreset attribute) ChatMessage (class in web.db.models) ChatSession (class in web.db.models) check_batch_slo_compliance() (in module utils.optimizer) cleanup_experiment() (orchestrator.AutotunerOrchestrator method) cleanup_port_forward() (controllers.direct_benchmark_controller.DirectBenchmarkController method) cleanup_results() (controllers.direct_benchmark_controller.DirectBenchmarkController method) clusterbasemodel_config (web.db.models.Task attribute) clusterservingruntime_config (web.db.models.Task attribute) COMPLETED (web.db.models.TaskStatus attribute) completed_at (web.db.models.Experiment attribute) (web.db.models.Task attribute) content (web.db.models.ChatMessage attribute) context_summary (web.db.models.ChatSession attribute) controllers.base_controller module controllers.benchmark_controller module

controllers.direct_benchmark_controller module controllers.docker_controller module controllers.local_controller module controllers.ome_controller module create_benchmark_job() (controllers.benchmark_controller.BenchmarkController method) create_namespace() (controllers.ome_controller.OMEController method) create_optimization_strategy() (in module utils.optimizer) created_at (web.db.models.AgentEventSubscription attribute) (web.db.models.ChatMessage attribute) (web.db.models.ChatSession attribute) (web.db.models.Experiment attribute) (web.db.models.ParameterPreset attribute) (web.db.models.Task attribute) created_clusterbasemodel (web.db.models.Task attribute) created_clusterservingruntime (web.db.models.Task attribute)

D

delete_benchmark_job() (controllers.benchmark_controller.BenchmarkController method) delete_inference_service() (controllers.base_controller.BaseModelController method) (controllers.docker_controller.DockerController method) (controllers.local_controller.LocalController method) (controllers.ome_controller.OMEController method) deploy_inference_service() (controllers.base_controller.BaseModelController method) (controllers.docker_controller.DockerController method) (controllers.local_controller.LocalController method) (controllers.ome_controller.OMEController method)

DEPLOYING (web.db.models.ExperimentStatus attribute) deployment_mode (web.db.models.Task attribute) description (web.db.models.ParameterPreset attribute) (web.db.models.Task attribute) DirectBenchmarkController (class in controllers.direct_benchmark_controller) DockerController (class in controllers.docker_controller)

E

elapsed_time (web.db.models.Experiment attribute) (web.db.models.Task attribute) ensure_clusterbasemodel() (controllers.ome_controller.OMEController method) ensure_clusterservingruntime() (controllers.ome_controller.OMEController method) ensure_model_downloaded() (controllers.local_controller.LocalController method) error_message (web.db.models.Experiment attribute) estimate_and_acquire() (in module utils.gpu_pool)

event_types (web.db.models.AgentEventSubscription attribute) Experiment (class in web.db.models) experiment_id (utils.gpu_pool.GPUAllocation attribute) (web.db.models.Experiment attribute) experiments (web.db.models.Task attribute) ExperimentStatus (class in web.db.models) expires_at (web.db.models.AgentEventSubscription attribute)

F

FAILED (web.db.models.ExperimentStatus attribute) (web.db.models.TaskStatus attribute) find_best_node_for_deployment() (in module utils.gpu_discovery)

from_state() (utils.optimizer.BayesianStrategy class method) (utils.optimizer.GridSearchStrategy class method) (utils.optimizer.OptimizationStrategy class method) (utils.optimizer.RandomSearchStrategy class method)

G

generate_parameter_grid() (in module utils.optimizer) get_available_gpus() (utils.gpu_monitor.GPUMonitor method) get_benchmark_results() (controllers.benchmark_controller.BenchmarkController method) get_best_params() (utils.optimizer.BayesianStrategy method) get_best_score() (utils.optimizer.BayesianStrategy method) get_cluster_gpu_status() (in module utils.gpu_discovery) get_container_logs() (controllers.docker_controller.DockerController method) (controllers.local_controller.LocalController method) get_gpu_count() (utils.gpu_monitor.GPUMonitor method) get_gpu_info() (controllers.docker_controller.DockerController method) (controllers.local_controller.LocalController method) (utils.gpu_monitor.GPUMonitor method) get_gpu_monitor() (in module utils.gpu_monitor) get_node_gpu_summaries() (in module utils.gpu_discovery) get_service_url() (controllers.base_controller.BaseModelController method) (controllers.docker_controller.DockerController method) (controllers.local_controller.LocalController method) (controllers.ome_controller.OMEController method)

get_state() (utils.optimizer.BayesianStrategy method) (utils.optimizer.GridSearchStrategy method) (utils.optimizer.OptimizationStrategy method) (utils.optimizer.RandomSearchStrategy method) get_status() (utils.gpu_pool.GPUResourcePool method) get_summary_stats() (utils.gpu_monitor.GPUMonitor method) gpu_indices (utils.gpu_pool.GPUAllocation attribute) gpu_info (web.db.models.Experiment attribute) GPUAllocation (class in utils.gpu_pool) GPUMonitor (class in utils.gpu_monitor) GPUResourcePool (class in utils.gpu_pool) gpus (utils.gpu_monitor.GPUSnapshot attribute) gpus_with_metrics (utils.gpu_discovery.NodeGPUSummary attribute) GPUSnapshot (class in utils.gpu_monitor) GridSearchStrategy (class in utils.optimizer)

I

id (web.db.models.AgentEventSubscription attribute) (web.db.models.ChatMessage attribute) (web.db.models.ChatSession attribute) (web.db.models.Experiment attribute) (web.db.models.ParameterPreset attribute) (web.db.models.Task attribute)

idle_gpu_count (utils.gpu_discovery.NodeGPUSummary attribute) is_active (web.db.models.AgentEventSubscription attribute) (web.db.models.ChatSession attribute) is_available() (utils.gpu_monitor.GPUMonitor method) is_system (web.db.models.ParameterPreset attribute)

L

list_clusterbasemodels() (controllers.ome_controller.OMEController method)

list_clusterservingruntimes() (controllers.ome_controller.OMEController method) LocalController (class in controllers.local_controller)

M

message_metadata (web.db.models.ChatMessage attribute) MessageRole (class in web.db.models) messages (web.db.models.ChatSession attribute) metrics (web.db.models.Experiment attribute) model_config (web.db.models.Task attribute) module controllers.base_controller controllers.benchmark_controller controllers.direct_benchmark_controller controllers.docker_controller controllers.local_controller controllers.ome_controller orchestrator utils.gpu_discovery utils.gpu_monitor utils.gpu_pool utils.optimizer web.db.models

monitor_gpus() (utils.gpu_monitor.GPUMonitor method)

N

name (web.db.models.ParameterPreset attribute)

node_name (utils.gpu_discovery.NodeGPUSummary attribute) NodeGPUSummary (class in utils.gpu_discovery)

O

objective_score (web.db.models.Experiment attribute) OMEController (class in controllers.ome_controller) optimization_config (web.db.models.Task attribute)

OptimizationStrategy (class in utils.optimizer) orchestrator module

P

parallel_config (web.db.models.Task attribute) ParameterPreset (class in web.db.models) parameters (web.db.models.Experiment attribute) (web.db.models.ParameterPreset attribute) (web.db.models.Task attribute)

params (utils.gpu_pool.GPUAllocation attribute) PENDING (web.db.models.ExperimentStatus attribute) (web.db.models.TaskStatus attribute) preset_metadata (web.db.models.ParameterPreset attribute)

Q

quant_config (web.db.models.Task attribute)

query_gpus() (utils.gpu_monitor.GPUMonitor method)

R

RandomSearchStrategy (class in utils.optimizer) release() (utils.gpu_pool.GPUResourcePool method) restore_optimization_strategy() (in module utils.optimizer) role (web.db.models.ChatMessage attribute) run_benchmark() (controllers.direct_benchmark_controller.DirectBenchmarkController method)

run_experiment() (orchestrator.AutotunerOrchestrator method) run_task() (orchestrator.AutotunerOrchestrator method) RUNNING (web.db.models.TaskStatus attribute) runtime (web.db.models.ParameterPreset attribute) runtime_image_tag (web.db.models.Task attribute)

S

service_name (web.db.models.Experiment attribute) service_url (web.db.models.Experiment attribute) session (web.db.models.AgentEventSubscription attribute) (web.db.models.ChatMessage attribute) session_id (web.db.models.AgentEventSubscription attribute) (web.db.models.ChatMessage attribute) (web.db.models.ChatSession attribute) session_metadata (web.db.models.ChatSession attribute) setup_port_forward() (controllers.direct_benchmark_controller.DirectBenchmarkController method) should_stop() (utils.optimizer.BayesianStrategy method) (utils.optimizer.GridSearchStrategy method) (utils.optimizer.OptimizationStrategy method) (utils.optimizer.RandomSearchStrategy method)

slo_config (web.db.models.Task attribute) started_at (web.db.models.Experiment attribute) (web.db.models.Task attribute) status (web.db.models.Experiment attribute) (web.db.models.Task attribute) subscriptions (web.db.models.ChatSession attribute) SUCCESS (web.db.models.ExperimentStatus attribute) successful_experiments (web.db.models.Task attribute) suggest_parameters() (utils.optimizer.BayesianStrategy method) (utils.optimizer.GridSearchStrategy method) (utils.optimizer.OptimizationStrategy method) (utils.optimizer.RandomSearchStrategy method) SYSTEM (web.db.models.MessageRole attribute)

T

Task (class in web.db.models) task (web.db.models.AgentEventSubscription attribute) (web.db.models.Experiment attribute) task_id (web.db.models.AgentEventSubscription attribute) (web.db.models.Experiment attribute) task_metadata (web.db.models.Task attribute) task_name (web.db.models.Task attribute) TaskStatus (class in web.db.models) tell_result() (utils.optimizer.BayesianStrategy method) (utils.optimizer.GridSearchStrategy method) (utils.optimizer.OptimizationStrategy method) (utils.optimizer.RandomSearchStrategy method)

timestamp (utils.gpu_monitor.GPUSnapshot attribute) title (web.db.models.ChatSession attribute) to_dict() (utils.gpu_monitor.GPUSnapshot method) (web.db.models.Experiment method) (web.db.models.ParameterPreset method) (web.db.models.Task method) token_count (web.db.models.ChatMessage attribute) tool_calls (web.db.models.ChatMessage attribute) total_experiments (web.db.models.Task attribute) total_gpus (utils.gpu_discovery.NodeGPUSummary attribute) (utils.gpu_monitor.GPUSnapshot attribute)

U

updated_at (web.db.models.ChatSession attribute) (web.db.models.ParameterPreset attribute) USER (web.db.models.MessageRole attribute) user_id (web.db.models.ChatSession attribute) utils.gpu_discovery module

utils.gpu_monitor module utils.gpu_pool module utils.optimizer module

W

wait_for_completion() (controllers.benchmark_controller.BenchmarkController method) wait_for_ready() (controllers.base_controller.BaseModelController method) (controllers.docker_controller.DockerController method) (controllers.local_controller.LocalController method) (controllers.ome_controller.OMEController method)

web.db.models module