torch.distributed.elastic

若API未标明支持情况,则代表该API的支持情况待验证。

API名称

是否支持

限制与说明

torch.distributed.elastic.agent.server.ElasticAgent

否

  

torch.distributed.elastic.agent.server.WorkerSpec

     

torch.distributed.elastic.agent.server.WorkerSpec.get_entrypoint_name

否

  

torch.distributed.elastic.agent.server.WorkerState

否

  

torch.distributed.elastic.agent.server.Worker

否

  

torch.distributed.elastic.agent.server.WorkerGroup

否

  

torch.distributed.elastic.agent.server.local_elastic_agent.LocalElasticAgent

否

  

torch.distributed.elastic.agent.server.SimpleElasticAgent

否

  

torch.distributed.elastic.agent.server.SimpleElasticAgent._assign_worker_ranks

否

  

torch.distributed.elastic.agent.server.SimpleElasticAgent._exit_barrier

否

  

torch.distributed.elastic.agent.server.SimpleElasticAgent._initialize_workers

否

  

torch.distributed.elastic.agent.server.SimpleElasticAgent._rendezvous

否

  

torch.distributed.elastic.agent.server.SimpleElasticAgent._restart_workers

否

  

torch.distributed.elastic.agent.server.api.RunResult

否

  

torch.distributed.elastic.multiprocessing.start_processes

否

  

torch.distributed.elastic.multiprocessing.api.PContext

否

  

torch.distributed.elastic.multiprocessing.api.MultiprocessContext

否

  

torch.distributed.elastic.multiprocessing.api.SubprocessContext

否

  

torch.distributed.elastic.multiprocessing.api.RunProcsResult

否

  

torch.distributed.elastic.multiprocessing.errors.record

否

  

torch.distributed.elastic.multiprocessing.errors.ChildFailedError

否

  

torch.distributed.elastic.multiprocessing.errors.ErrorHandler

否

  

torch.distributed.elastic.multiprocessing.errors.ProcessFailure

否

  

torch.distributed.elastic.rendezvous.RendezvousParameters

     

torch.distributed.elastic.rendezvous.RendezvousParameters.get

否

  

torch.distributed.elastic.rendezvous.RendezvousParameters.get_as_bool

否

  

torch.distributed.elastic.rendezvous.RendezvousParameters.get_as_int

否

  

torch.distributed.elastic.rendezvous.RendezvousHandlerRegistry

否

  

torch.distributed.elastic.rendezvous.RendezvousHandlerRegistry.create_handler

否

  

torch.distributed.elastic.rendezvous.RendezvousHandlerRegistry.register

否

  

torch.distributed.elastic.rendezvous.RendezvousHandler

否

  

torch.distributed.elastic.rendezvous.RendezvousHandler.shutdown

否

  

torch.distributed.elastic.rendezvous.RendezvousError

否

  

torch.distributed.elastic.rendezvous.RendezvousClosedError

否

  

torch.distributed.elastic.rendezvous.RendezvousTimeoutError

否

  

torch.distributed.elastic.rendezvous.RendezvousConnectionError

否

  

torch.distributed.elastic.rendezvous.RendezvousStateError

否

  

torch.distributed.elastic.rendezvous.dynamic_rendezvous.create_handler

否

  

torch.distributed.elastic.rendezvous.dynamic_rendezvous.DynamicRendezvousHandler

否

  

torch.distributed.elastic.rendezvous.dynamic_rendezvous.DynamicRendezvousHandler.from_backend

     

torch.distributed.elastic.rendezvous.dynamic_rendezvous.RendezvousBackend

否

  

torch.distributed.elastic.rendezvous.dynamic_rendezvous.RendezvousTimeout

否

  

torch.distributed.elastic.rendezvous.c10d_rendezvous_backend.create_backend

否

  

torch.distributed.elastic.rendezvous.c10d_rendezvous_backend.C10dRendezvousBackend

否

  

torch.distributed.elastic.rendezvous.c10d_rendezvous_backend.C10dRendezvousBackend.get_state

否

  

torch.distributed.elastic.rendezvous.c10d_rendezvous_backend.C10dRendezvousBackend.set_state

否

  

torch.distributed.elastic.rendezvous.etcd_rendezvous_backend.create_backend

否

  

torch.distributed.elastic.rendezvous.etcd_rendezvous_backend.EtcdRendezvousBackend

否

  

torch.distributed.elastic.rendezvous.etcd_rendezvous_backend.EtcdRendezvousBackend.get_state

否

  

torch.distributed.elastic.rendezvous.etcd_rendezvous_backend.EtcdRendezvousBackend.set_state

否

  

torch.distributed.elastic.rendezvous.etcd_rendezvous.EtcdRendezvousHandler

否

  

torch.distributed.elastic.rendezvous.etcd_store.EtcdStore

否

  

torch.distributed.elastic.rendezvous.etcd_store.EtcdStore.add

否

  

torch.distributed.elastic.rendezvous.etcd_store.EtcdStore.check

否

  

torch.distributed.elastic.rendezvous.etcd_store.EtcdStore.get

否

  

torch.distributed.elastic.rendezvous.etcd_store.EtcdStore.set

否

  

torch.distributed.elastic.rendezvous.etcd_store.EtcdStore.wait

否

  

torch.distributed.elastic.rendezvous.etcd_server.EtcdServer

否

  

torch.distributed.elastic.timer.configure

否

  

torch.distributed.elastic.timer.expires

否

  

torch.distributed.elastic.timer.LocalTimerServer

否

  

torch.distributed.elastic.timer.LocalTimerClient

否

  

torch.distributed.elastic.timer.FileTimerServer

否

  

torch.distributed.elastic.timer.FileTimerClient

否

  

torch.distributed.elastic.timer.TimerRequest

否

  

torch.distributed.elastic.timer.TimerServer

否

  

torch.distributed.elastic.timer.TimerClient

否

  

torch.distributed.elastic.metrics.api.MetricHandler

否

  

torch.distributed.elastic.metrics.api.ConsoleMetricHandler

否

  

torch.distributed.elastic.metrics.api.NullMetricHandler

否

  

torch.distributed.elastic.metrics.configure

否

  

torch.distributed.elastic.metrics.prof

否

  

torch.distributed.elastic.metrics.put_metric

否

  

torch.distributed.elastic.events.record

否

  

torch.distributed.elastic.events.get_logging_handler

否

  

torch.distributed.elastic.events.api.Event

否

  

torch.distributed.elastic.events.api.EventSource

否

  

torch.distributed.elastic.events.api.EventMetadataValue

否