furiosa.server.api.grpc package

Subpackages

• furiosa.server.api.grpc.generated package
  • Submodules
  • furiosa.server.api.grpc.generated.model_repository_pb2 module
  • furiosa.server.api.grpc.generated.model_repository_pb2_grpc module
  • furiosa.server.api.grpc.generated.predict_pb2 module
  • furiosa.server.api.grpc.generated.predict_pb2_grpc module
  • Module contents

Submodules

furiosa.server.api.grpc.converters module

class furiosa.server.api.grpc.converters.InferInputTensorConverter

Bases: object

classmethod from_types(type_object: furiosa.server.types.predict.RequestInput) → predict_pb2.InferInputTensor

classmethod to_types(pb_object: predict_pb2.InferInputTensor) → furiosa.server.types.predict.RequestInput

class furiosa.server.api.grpc.converters.InferOutputTensorConverter

Bases: object

classmethod from_types(type_object: furiosa.server.types.predict.ResponseOutput) → predict_pb2.InferOutputTensor

classmethod to_types(pb_object: predict_pb2.InferOutputTensor) → furiosa.server.types.predict.ResponseOutput

class furiosa.server.api.grpc.converters.InferRequestedOutputTensorConverter

Bases: object

classmethod from_types(type_object: furiosa.server.types.predict.RequestOutput) → predict_pb2.InferRequestedOutputTensor

classmethod to_types(pb_object: predict_pb2.InferRequestedOutputTensor) → furiosa.server.types.predict.RequestOutput

class furiosa.server.api.grpc.converters.InferTensorContentsConverter

Bases: object

classmethod from_types(type_object: furiosa.server.types.predict.TensorData, datatype: str) → predict_pb2.InferTensorContents

classmethod to_types(pb_object: predict_pb2.InferTensorContents) → furiosa.server.types.predict.TensorData

class furiosa.server.api.grpc.converters.ModelInferRequestConverter

Bases: object

classmethod from_types(type_object: furiosa.server.types.predict.InferenceRequest, model_name: str, model_version: str) → predict_pb2.ModelInferRequest

classmethod to_types(pb_object: predict_pb2.ModelInferRequest) → furiosa.server.types.predict.InferenceRequest

class furiosa.server.api.grpc.converters.ModelInferResponseConverter

Bases: object

classmethod from_types(type_object: furiosa.server.types.predict.InferenceResponse) → predict_pb2.ModelInferResponse

classmethod to_types(pb_object: predict_pb2.ModelInferResponse) → furiosa.server.types.predict.InferenceResponse

class furiosa.server.api.grpc.converters.ModelMetadataResponseConverter

Bases: object

classmethod from_types(type_object: furiosa.server.types.predict.MetadataModelResponse) → predict_pb2.ModelMetadataResponse

classmethod to_types(pb_object: predict_pb2.ModelMetadataResponse) → furiosa.server.types.predict.MetadataModelResponse

class furiosa.server.api.grpc.converters.ParametersConverter

Bases: object

classmethod from_types(type_object: furiosa.server.types.predict.Parameters) → Mapping[str, predict_pb2.InferParameter]

classmethod to_types(pb_object: Mapping[str, predict_pb2.InferParameter]) → Optional[furiosa.server.types.predict.Parameters]

class furiosa.server.api.grpc.converters.RepositoryIndexRequestConverter

Bases: object

classmethod from_types(type_object: furiosa.server.types.model_repository.RepositoryIndexRequest) → model_repository_pb2.RepositoryIndexRequest

classmethod to_types(pb_object: model_repository_pb2.RepositoryIndexRequest) → furiosa.server.types.model_repository.RepositoryIndexRequest

class furiosa.server.api.grpc.converters.RepositoryIndexResponseConverter

Bases: object

classmethod from_types(type_object: furiosa.server.types.model_repository.RepositoryIndexResponse) → model_repository_pb2.RepositoryIndexResponse

classmethod to_types(pb_object: model_repository_pb2.RepositoryIndexResponse) → furiosa.server.types.model_repository.RepositoryIndexResponse

class furiosa.server.api.grpc.converters.RepositoryIndexResponseItemConverter

Bases: object

classmethod from_types(type_object: furiosa.server.types.model_repository.RepositoryIndexResponseItem) → model_repository_pb2.ModelIndex

classmethod to_types(pb_object: model_repository_pb2.ModelIndex) → furiosa.server.types.model_repository.RepositoryIndexResponseItem

class furiosa.server.api.grpc.converters.ServerMetadataResponseConverter

Bases: object

classmethod from_types(type_object: furiosa.server.types.predict.MetadataServerResponse) → predict_pb2.ServerMetadataResponse

classmethod to_types(pb_object: predict_pb2.ServerMetadataResponse) → furiosa.server.types.predict.MetadataServerResponse

class furiosa.server.api.grpc.converters.TensorMetadataConverter

Bases: object

classmethod from_types(type_object: furiosa.server.types.predict.MetadataTensor) → predict_pb2.TensorMetadata

classmethod to_types(pb_object: predict_pb2.TensorMetadata) → furiosa.server.types.predict.MetadataTensor

furiosa.server.api.grpc.server module

class furiosa.server.api.grpc.server.GRPCServer(config: furiosa.server.settings.GRPCServerConfig, predict_handler: furiosa.server.handlers.predict.PredictHandler, repository_handler: furiosa.server.handlers.repository.RepositoryHandler)

Bases: object

async start()

async stop()

furiosa.server.api.grpc.servicers module

class furiosa.server.api.grpc.servicers.InferenceServicer(predict_handler: furiosa.server.handlers.predict.PredictHandler)

Bases: furiosa.server.api.grpc.generated.predict_pb2_grpc.GRPCInferenceServiceServicer

async ModelInfer(request, context)

The ModelInfer API performs inference using the specified model. Errors are indicated by the google.rpc.Status returned for the request. The OK code indicates success and other codes indicate failure.

async ModelMetadata(request, context)

The per-model metadata API provides information about a model. Errors are indicated by the google.rpc.Status returned for the request. The OK code indicates success and other codes indicate failure.

async ModelReady(request: predict_pb2.ModelReadyRequest, context) → predict_pb2.ModelReadyResponse

The ModelReady API indicates if a specific model is ready for inferencing.

async ServerLive(request: predict_pb2.ServerLiveRequest, context) → predict_pb2.ServerLiveResponse

The ServerLive API indicates if the inference server is able to receive and respond to metadata and inference requests.

async ServerMetadata(request: predict_pb2.ServerMetadataRequest, context) → predict_pb2.ServerMetadataResponse

The ServerMetadata API provides information about the server. Errors are indicated by the google.rpc.Status returned for the request. The OK code indicates success and other codes indicate failure.

async ServerReady(request: predict_pb2.ServerReadyRequest, context) → predict_pb2.ServerReadyResponse

The ServerReady API indicates if the server is ready for inferencing.

class furiosa.server.api.grpc.servicers.ModelRepositoryServicer(handler: furiosa.server.handlers.repository.RepositoryHandler)

Bases: furiosa.server.api.grpc.generated.model_repository_pb2_grpc.ModelRepositoryServiceServicer

async RepositoryIndex(request: model_repository_pb2.RepositoryIndexRequest, context) → model_repository_pb2.RepositoryIndexResponse

Get the index of model repository contents.

async RepositoryModelLoad(request, context)

Load or reload a model from a repository.

async RepositoryModelUnload(request, context)

Unload a model.

Module contents