execute_training_operation async ¶

async def execute_training_operation(
    self,
    training_session_name: "capo_sagemakertrainingsessionruntime.types.training_session_name.TrainingSessionName",
    sequence_number: "capo_sagemakertrainingsessionruntime.types.sequence_number.SequenceNumber",
    operation_config: "capo_sagemakertrainingsessionruntime.types.operation_config.OperationConfig",
    *,
    config_overrides: Optional[
        AsyncSageMakerTrainingSessionRuntimeClientConfig
    ] = None,
    client_token: Optional[
        "capo_sagemakertrainingsessionruntime.types.client_token.ClientToken"
    ] = None,
) -> "capo_sagemakertrainingsessionruntime.types.execute_training_operation_output.ExecuteTrainingOperationOutput":
    """<p>Submits an asynchronous training operation to the specified training session. The service executes operations in sequence number order. The session must be in <code>Running</code> or <code>Blocked</code> state. Use <code>GetTrainingOperationResult</code> to poll for the result. Supported operation types: <code>forward</code>, <code>forwardBackward</code>, <code>optimizerStep</code>, <code>saveWeights</code>, <code>saveState</code>, and <code>loadState</code>.</p>

    Args:
        training_session_name: <p>The name of the training session to submit the operation to.</p>
        sequence_number: <p>The monotonically increasing sequence number that determines execution order. The training session processes operations strictly in sequence number order. If a gap exists, the session blocks until the missing operation arrives or times out. Must be at least 1.</p>
        client_token: <p>A unique, case-sensitive identifier that you provide to ensure that the operation completes no more than one time. If you retry a request with the same client token and the same parameters, the service ignores the request but does not return an error.</p>
        operation_config: <p>The configuration that specifies the type and parameters of the training operation. Exactly one member of the union must be set.</p>

    Raises:
        capo_sagemakertrainingsessionruntime.errors.conflict_exception.ConflictException: <p>The training session is in a state that cannot accept new operations. The session must be in <code>Running</code> or <code>Blocked</code> state to accept requests.</p>
        capo_sagemakertrainingsessionruntime.errors.internal_server_exception.InternalServerException: <p>An internal service error occurred. Retry the request using exponential backoff.</p>
        capo_sagemakertrainingsessionruntime.errors.resource_not_found_exception.ResourceNotFoundException: <p>The specified training session or operation does not exist, or the operation does not belong to the specified training session.</p>
        capo_sagemakertrainingsessionruntime.errors.throttling_exception.ThrottlingException: <p>The request was throttled. Retry the request after a brief wait using exponential backoff.</p>
        capo_sagemakertrainingsessionruntime.errors.validation_exception.ValidationException: <p>The request is not valid. Check the request syntax and parameters. Examples include invalid sequence numbers, malformed client tokens, missing required fields, or invalid operation configuration.</p>
        capo_sagemakertrainingsessionruntime.errors.UnknownServiceError: The service returned an error code this client does not model.

    Examples:
        Run a forward and backward pass
        Computes gradients for a batch of tokenized prompts and accumulates them into the session. Operations are applied in `sequenceNumber` order, so each call must supply the next number in the sequence.

        >>> await client.execute_training_operation(training_session_name='my-training-session', sequence_number=1, operation_config={'forwardBackward': {'lossFnName': 'cross_entropy', 'datumList': [{'modelInput': {'chunks': [{'tokens': [791, 6864, 315, 9822, 374, 12366], 'type': 'encoded_text'}]}}]}})
    """

    async def _handler(
        req: "AsyncOperationRequest[capo_sagemakertrainingsessionruntime.types.execute_training_operation_input.ExecuteTrainingOperationInput]",
    ) -> AsyncOperationResponse[
        "capo_sagemakertrainingsessionruntime.types.execute_training_operation_output.ExecuteTrainingOperationOutput"
    ]:
        import capo_sagemakertrainingsessionruntime._operations.lego_router_service.execute_training_operation

        (
            output,
            http_response,
        ) = await capo_sagemakertrainingsessionruntime._operations.lego_router_service.execute_training_operation.async_execute_training_operation(
            req.options, req.input
        )
        return AsyncOperationResponse(output=output, response=http_response)

    def _redact_input(
        input_: "capo_sagemakertrainingsessionruntime.types.execute_training_operation_input.ExecuteTrainingOperationInput",
    ) -> dict[str, Any]:
        import capo_sagemakertrainingsessionruntime.types.execute_training_operation_input

        return capo_sagemakertrainingsessionruntime.types.execute_training_operation_input.redact_sensitive(
            input_
        )

    interceptors_, options_ = self.operation_options(
        config_overrides,
        OperationLogging("ExecuteTrainingOperation", redact_input=_redact_input),
    )
    input_: capo_sagemakertrainingsessionruntime.types.execute_training_operation_input.ExecuteTrainingOperationInput = {
        "training_session_name": training_session_name,
        "sequence_number": sequence_number,
        "operation_config": operation_config,
    }
    if client_token is not None:
        input_["client_token"] = client_token

    response = await aexecute_pipeline(
        AsyncOperationRequest(input=input_, options=options_),
        handler=_handler,
        interceptors=list(interceptors_),
    )
    await response.response.aclose()
    return response.output