tensorflow API reference

400 public APIs from tensorflow (tensorflow/tensorflow) — 243 classes, 78 functions, 79 methods. Signatures extracted by static analysis of the actual source.

Repository: tensorflow/tensorflow

Kind	Count
Classes	243
Functions	78
Methods	79

API list

funcci.official.utilities.extract_resultstore_links.parse_args() -> argparse.Namespace

Parses the commandline args.

funcconfigure.configure_ios(environ_cp)

Configures TensorFlow for iOS builds.

funcconfigure.cygpath(path)

Convert path from posix to windows.

funcconfigure.get_python_major_version(python_bin_path)

Get the python major version.

funcconfigure.get_python_path(environ_cp, python_bin_path)

Get the python site package paths.

funcconfigure.retrieve_clang_version(clang_executable)

Retrieve installed clang version.

funcconfigure.set_clang_cuda_compiler_path(environ_cp)

Set CLANG_CUDA_COMPILER_PATH.

funcconfigure.set_gcc_host_compiler_path(environ_cp)

Set GCC_HOST_COMPILER_PATH.

funcconfigure.set_hermetic_cuda_version(environ_cp)

Set HERMETIC_CUDA_VERSION.

funcconfigure.set_hermetic_cudnn_version(environ_cp)

Set HERMETIC_CUDNN_VERSION.

funcconfigure.set_other_cuda_vars(environ_cp)

Set other CUDA related variables.

funcconfigure.set_system_libs_flag(environ_cp)

Set system libs flags.

funcconfigure.set_tf_cuda_clang(environ_cp)

set TF_CUDA_CLANG action_env.

funcconfigure.set_tf_download_clang(environ_cp)

Set TF_DOWNLOAD_CLANG action_env.

funcconfigure.set_windows_build_flags(environ_cp)

Set Windows specific build options.

funcconfigure.setup_python(environ_cp)

Setup python related env variables.

funcconfigure.symlink_force(target, link_name)

Force symlink, equivalent of 'ln -sf'.

classtensorflow.c.experimental.saved_model.internal.testdata.gen_saved_models.Module

A module with an UninitializedVariable.

classtensorflow.c.experimental.saved_model.internal.testdata.gen_saved_models.SubModule

A module with an UninitializedVariable.

functensorflow.compiler.mlir.quantization.common.python.testing.get_size_ratio(path_a:str, path_b:str) -> float

Return the size ratio of the given paths.

class

tensorflow.compiler.mlir.quantization.stablehlo.python.integration_test.quantize_model_test_base.QuantizedModelTest.EinsumModel

Einsum class.

func

tensorflow.compiler.mlir.quantization.stablehlo.python.quantization.quantize_saved_model(src_saved_model_path:str, dst_saved_model_path:str, config:qc.QuantizationConfig) -> None

Quantizes a saved model.

class

tensorflow.compiler.mlir.quantization.tensorflow.python.integration_test.quantize_model_test_base.QuantizedModelTest.EinsumModel

Einsum class.

class

tensorflow.compiler.mlir.quantization.tensorflow.python.integration_test.quantize_model_test_base.QuantizedModelTest.WhileModel

A model with a while op.

classtensorflow.compiler.mlir.quantization.tensorflow.python.representative_dataset.RepresentativeDatasetLoader

Representative dataset loader.

classtensorflow.compiler.mlir.quantization.tensorflow.python.representative_dataset.RepresentativeDatasetSaver

Representative dataset saver.

func

tensorflow.compiler.mlir.quantization.tensorflow.python.representative_dataset.get_num_samples(repr_ds:RepresentativeDataset) -> Optional[int]

Returns the number of samples if known.

func

tensorflow.compiler.mlir.quantization.tensorflow.python.save_model.create_empty_output_dir(output_directory:str, overwrite:bool=True) -> None

Creates the `output_directory`.

classtensorflow.compiler.mlir.tfr.python.tfr_gen.SymbolTable

Symbol Table for python code.

classtensorflow.compiler.mlir.tfr.python.tfr_gen.TFRGen

Visit the AST and generate MLIR TFR functions.

classtensorflow.compiler.mlir.tfr.python.tfr_gen.TFRTypes

All the supported types.

classtensorflow.compiler.mlir.tfr.python.tfr_gen.TfrGen

Transforms Python objects into TFR MLIR source code.

classtensorflow.core.function.capture.capture_container.MutationAwareDict

A dict with a mutation flag.

classtensorflow.core.function.polymorphism.function_cache.FunctionCache

A container for managing functions.

method

tensorflow.core.function.polymorphism.function_cache.FunctionCache.add(fn:Any, context:Optional[FunctionContext]=None) -> None

Adds a new function using its function_type.

methodtensorflow.core.function.polymorphism.function_cache.FunctionCache.values()

Returns a list of all functions held by this cache.

classtensorflow.core.function.polymorphism.function_type.FunctionType

Represents the type of a TensorFlow function.

method

tensorflow.core.function.polymorphism.function_type.FunctionType.most_specific_common_subtype(others:Sequence['FunctionType']) -> Optional['FunctionType']

Returns a common subtype (if exists).

methodtensorflow.core.function.polymorphism.function_type.FunctionType.unpack_captures(captures) -> List[core.Tensor]

Unpacks captures to flat tensors.

classtensorflow.core.function.polymorphism.function_type.Parameter

Represents a parameter to a function.

methodtensorflow.core.function.polymorphism.function_type.Parameter.from_proto(proto:Any) -> 'Parameter'

Generate a Parameter from the proto representation.

methodtensorflow.core.function.polymorphism.function_type.Parameter.is_subtype_of(other:'Parameter') -> bool

Returns True if self is a supertype of other Parameter.

method

tensorflow.core.function.polymorphism.function_type.Parameter.most_specific_common_supertype(others:Sequence['Parameter']) -> Optional['Parameter']

Returns a common supertype (if exists).

functensorflow.core.function.polymorphism.function_type.sanitize_arg_name(name:str) -> str

Sanitizes function argument names.

classtensorflow.core.function.polymorphism.type_dispatch.TypeDispatchTable

Type dispatch table implementation.

method

tensorflow.core.function.polymorphism.type_dispatch.TypeDispatchTable.add_target(target:function_type.FunctionType) -> None

Adds a new target type.

methodtensorflow.core.function.polymorphism.type_dispatch.TypeDispatchTable.clear() -> None

Deletes all targets in the table.

methodtensorflow.core.function.polymorphism.type_dispatch.TypeDispatchTable.delete(target:function_type.FunctionType) -> None

Deletes a target in the table if it exists.

classtensorflow.core.function.trace_type.default_types.Attrs

Represents a class annotated by attr.s.

method

tensorflow.core.function.trace_type.default_types.Attrs.most_specific_common_supertype(others:Sequence[trace.TraceType]) -> Optional['Attrs']

See base class.

classtensorflow.core.function.trace_type.default_types.Dict

Represents a dictionary of TraceType objects.

methodtensorflow.core.function.trace_type.default_types.Dict.is_subtype_of(other:trace.TraceType) -> bool

See base class.

method

tensorflow.core.function.trace_type.default_types.Dict.most_specific_common_supertype(types:Sequence[trace.TraceType]) -> Optional['Dict']

See base class.

classtensorflow.core.function.trace_type.default_types.List

Represents a list of TraceType objects.

method

tensorflow.core.function.trace_type.default_types.List.most_specific_common_supertype(others:Sequence[trace.TraceType]) -> Optional['Tuple']

See base class.

classtensorflow.core.function.trace_type.default_types.Literal

Represents a Literal type like bool, int or string.

classtensorflow.core.function.trace_type.default_types.NamedTuple

Represents a NamedTuple of TraceType objects.

method

tensorflow.core.function.trace_type.default_types.NamedTuple.most_specific_common_supertype(others:Sequence[trace.TraceType]) -> Optional['NamedTuple']

See base class.

classtensorflow.core.function.trace_type.default_types.Tuple

Represents a tuple of TraceType objects.

method

tensorflow.core.function.trace_type.default_types.Tuple.most_specific_common_supertype(others:Sequence[trace.TraceType]) -> Optional['Tuple']

See base class.

classtensorflow.core.function.trace_type.default_types.Weakref

Represents weakref of an arbitrary Python object.

method

tensorflow.core.function.trace_type.serialization.Serializable.experimental_from_proto(proto:message.Message) -> 'Serializable'

Returns an instance based on a proto.

functensorflow.core.function.trace_type.serialization.serialize(to_serialize:Serializable) -> SerializedTraceType

Converts Serializable to a proto SerializedTraceType.

classtensorflow.core.function.trace_type.trace_type_builder.InternalCastContext

Default casting behaviors.

classtensorflow.core.tfrt.mlrt.kernel.testdata.gen_checkpoint.ToyModule

A toy module for testing checkpoing loading.

functensorflow.dtensor.python.accelerator_util.shutdown_accelerator_system() -> None

Shuts down the accelerator system.

functensorflow.dtensor.python.api.check_layout(tensor:tensor_lib.Tensor, layout:layout_lib.Layout) -> None

Asserts that the layout of the DTensor is `layout`.

functensorflow.dtensor.python.api.device_name() -> str

Returns the singleton DTensor device's name.

functensorflow.dtensor.python.api.fetch_layout(tensor:tensor_lib.Tensor) -> layout_lib.Layout

Fetches the layout of a DTensor.

functensorflow.dtensor.python.api.is_dtensor(tensor) -> bool

Check whether the input tensor is a DTensor.

functensorflow.dtensor.python.api.pack(tensors:Sequence[Any], layout:layout_lib.Layout) -> Any

Packs `tf.Tensor` components into a DTensor.

func

tensorflow.dtensor.python.api.relayout(tensor:tensor_lib.Tensor, layout:layout_lib.Layout, name:Optional[str]=None) -> tensor_lib.Tensor

Changes the layout of `tensor`.

functensorflow.dtensor.python.api.reset_dtensor_device(is_async:bool) -> None

Resets the Eager execution device for DTensor.

functensorflow.dtensor.python.api.unpack(tensor:Any) -> Sequence[Any]

Unpacks a DTensor into `tf.Tensor` components.

functensorflow.dtensor.python.config.client_id() -> int

Returns this client's ID.

functensorflow.dtensor.python.config.is_gpu_present() -> bool

Returns true if TPU devices are present.

functensorflow.dtensor.python.config.is_local_mode() -> bool

Returns true if DTensor shall run in local mode.

functensorflow.dtensor.python.config.is_tpu_present() -> bool

Returns true if TPU devices are present.

functensorflow.dtensor.python.config.num_clients() -> int

Returns the number of clients in this DTensor cluster.

functensorflow.dtensor.python.heartbeat.start(period:int) -> threading.Event

Starts a persistent thread exchanging heartbeats between workers.

classtensorflow.dtensor.python.input_util.DTensorDataset

A dataset of DTensors.

classtensorflow.dtensor.python.input_util.TFDataServiceConfig

Specifies the tf.data service configuration to use.

classtensorflow.dtensor.python.layout.Layout

Represents the layout information of a DTensor.

methodtensorflow.dtensor.python.layout.Layout.batch_sharded(mesh:Mesh, batch_dim:str, rank:int, axis:int=0) -> 'Layout'

Returns a layout sharded on batch dimension.

methodtensorflow.dtensor.python.layout.Layout.delete(dims:List[int]) -> 'Layout'

Returns the layout with the give dimensions deleted.

methodtensorflow.dtensor.python.layout.Layout.from_device(device:str) -> 'Layout'

Constructs a single device layout from a single device mesh.

methodtensorflow.dtensor.python.layout.Layout.from_proto(layout_proto:layout_pb2.LayoutProto) -> 'Layout'

Creates an instance from a LayoutProto.

methodtensorflow.dtensor.python.layout.Layout.from_single_device_mesh(mesh:Mesh) -> 'Layout'

Constructs a single device layout from a single device mesh.

methodtensorflow.dtensor.python.layout.Layout.from_string(layout_str:str) -> 'Layout'

Creates an instance from a human-readable string.

methodtensorflow.dtensor.python.layout.Layout.inner_sharded(mesh:Mesh, inner_dim:str, rank:int) -> 'Layout'

Returns a layout sharded on inner dimension.

methodtensorflow.dtensor.python.layout.Layout.replicated(mesh:Mesh, rank:int) -> 'Layout'

Returns a replicated layout of rank `rank`.

methodtensorflow.dtensor.python.layout.Layout.to_parted() -> 'Layout'

Returns a "parted" layout from a static layout.

methodtensorflow.dtensor.python.layout.Mesh.coords(device_idx:int) -> tensor.Tensor

Converts the device index into a tensor of mesh coordinates.

methodtensorflow.dtensor.python.layout.Mesh.from_device(device:str) -> 'Mesh'

Constructs a single device mesh from a device string.

methodtensorflow.dtensor.python.layout.Mesh.from_proto(proto:layout_pb2.MeshProto) -> 'Mesh'

Construct a mesh instance from input `proto`.

methodtensorflow.dtensor.python.layout.Mesh.global_device_ids() -> np.ndarray

Returns a global device list as an array.

methodtensorflow.dtensor.python.layout.Mesh.host_mesh() -> 'Mesh'

Returns a host mesh.

methodtensorflow.dtensor.python.layout.Mesh.local_device_locations() -> List[Dict[str, int]]

Returns a list of local device locations.

methodtensorflow.dtensor.python.layout.Mesh.strides() -> List[int]

Returns the strides tensor array for this mesh.

func

tensorflow.dtensor.python.mesh_util.barrier(mesh:layout.Mesh, barrier_name:Optional[str]=None, timeout_in_ms:Optional[int]=None)

Runs a barrier on the mesh.

functensorflow.dtensor.python.numpy_util.to_numpy(tensor:TensorLike) -> np.ndarray

Copy `input` DTensor to an equivalent local numpy array.

func

tensorflow.dtensor.python.numpy_util.unpacked_to_numpy(unpacked:List[TensorLike], layout:layout_lib.Layout) -> np.ndarray

Heals local Tensor components to a numpy array.

functensorflow.dtensor.python.tpu_util.shutdown_tpu_system()

Shuts down the TPU system.

classtensorflow.lite.python.analyzer.ModelAnalyzer

Provides a collection of TFLite model analyzer tools.

classtensorflow.lite.python.convert_phase.Component

Enum class defining name of the converter components.

classtensorflow.lite.python.convert_phase.ConverterError

Raised when an error occurs during model conversion.

classtensorflow.lite.python.interpreter.Delegate

Python wrapper class to manage TfLiteDelegate objects.

methodtensorflow.lite.python.interpreter.Interpreter.invoke()

Invoke the interpreter.

classtensorflow.lite.python.interpreter.OpResolverType

Different types of op resolvers for Tensorflow Lite.

classtensorflow.lite.python.lite.RepresentativeDataset

Representative dataset used to optimize the model.

classtensorflow.lite.python.lite.TFLiteConverter

Convert a TensorFlow model into `output_format`.

classtensorflow.lite.python.lite.TFLiteConverterV2

Converts a TensorFlow model into TensorFlow Lite model.

classtensorflow.lite.python.lite.TocoConverter

Convert a TensorFlow model into `output_format`.

classtensorflow.lite.python.lite_v2_test_util.ModelTest

Base test class for TensorFlow Lite 2.x model tests.

classtensorflow.lite.python.lite_v2_test_util.ModelTest.BasicModel

Basic model with multiple functions.

classtensorflow.lite.python.lite_v2_test_util.ModelTest.ConvWrapper

A Wrapper for simulating QAT on Conv2D layers.

classtensorflow.lite.python.lite_v2_test_util.ModelTest.SimpleModelWithOneVariable

Basic model with 1 variable.

classtensorflow.lite.python.metrics.metrics_interface.TFLiteMetricsInterface

Abstract class for TFLiteMetrics.

classtensorflow.lite.python.metrics.metrics_portable.TFLiteMetrics

TFLite metrics helper.

classtensorflow.lite.python.op_hint.OpHint

A class that helps build tflite function invocations.

functensorflow.lite.python.tflite_convert.run_main(_)

Main in tflite_convert.py.

functensorflow.lite.python.util.freeze_graph(sess, input_tensors, output_tensors)

Returns a frozen GraphDef.

functensorflow.lite.python.util.get_dequantize_opcode_idx(model)

Returns the quantize op idx.

functensorflow.lite.python.util.get_quantize_opcode_idx(model)

Returns the quantize op idx.

functensorflow.lite.python.util.get_tensor_name(tensor)

Returns name of the input tensor.

functensorflow.lite.python.util.is_frozen_graph(sess)

Determines if the graph is frozen.

classtensorflow.lite.toco.logging.gen_html.HTMLGenerator

Utility class to generate an HTML report.

classtensorflow.python.autograph.converters.variables.VariableAccessTransformer

Rewrites basic symbol reads.

classtensorflow.python.autograph.core.ag_ctx.NullCtx

Helper substitute for contextlib.nullcontext.

classtensorflow.python.autograph.core.config_lib.Convert

Indicates that this module should be converted.

classtensorflow.python.autograph.core.config_lib.DoNotConvert

Indicates that this module should be not converted.

classtensorflow.python.autograph.core.config_lib.Rule

Base class for conversion rules.

classtensorflow.python.autograph.core.converter.Base

All converters should inherit from this class.

classtensorflow.python.autograph.core.converter.ConversionOptions

Immutable container for global conversion flags.

methodtensorflow.python.autograph.core.converter.Feature.all()

Returns a tuple that enables all options.

classtensorflow.python.autograph.core.converter_testing.TestCase

Base class for unit tests in this module.

classtensorflow.python.autograph.impl.api.AutoGraphError

Base class for all AutoGraph exceptions.

classtensorflow.python.autograph.impl.api.ConversionError

Raised during the conversion process.

classtensorflow.python.autograph.impl.api.PyToTF

The TensorFlow AutoGraph transformer.

classtensorflow.python.autograph.impl.api.StackTraceMapper

Remaps generated code to code it originated from.

classtensorflow.python.autograph.impl.api.StagingError

Raised during the staging (i.e.

functensorflow.python.autograph.impl.api.decorator(f)

Decorator implementation.

functensorflow.python.autograph.operators.data_structures.list_append(list_, x)

The list append function.

functensorflow.python.autograph.operators.data_structures.list_pop(list_, i, opts)

The list pop function.

functensorflow.python.autograph.operators.data_structures.list_stack(list_, opts)

The list stack function.

functensorflow.python.autograph.operators.data_structures.new_list(iterable=None)

The list constructor.

functensorflow.python.autograph.operators.logical.and_(a, b)

Functional form of "and".

functensorflow.python.autograph.operators.logical.eq(a, b)

Functional form of "equal".

functensorflow.python.autograph.operators.logical.not_(a)

Functional form of "not".

functensorflow.python.autograph.operators.logical.not_eq(a, b)

Functional form of "not-equal".

functensorflow.python.autograph.operators.logical.or_(a, b)

Functional form of "or".

functensorflow.python.autograph.operators.slices.get_item(target, i, opts)

The slice read operator (i.e.

functensorflow.python.autograph.operators.slices.set_item(target, i, x)

The slice write operator (i.e.

classtensorflow.python.autograph.operators.variables.Undefined

Represents an undefined symbol in Python.

classtensorflow.python.autograph.operators.variables.UndefinedReturnValue

Represents a return value that is undefined.

functensorflow.python.autograph.operators.variables.ld(v)

Load variable operator.

classtensorflow.python.autograph.pyct.anno.Basic

Container for basic annotation keys.

classtensorflow.python.autograph.pyct.anno.NoValue

Base class for different types of AST annotations.

classtensorflow.python.autograph.pyct.anno.Static

Container for static analysis annotation keys.

classtensorflow.python.autograph.pyct.ast_util.CleanCopier

NodeTransformer-like visitor that copies an AST.

functensorflow.python.autograph.pyct.ast_util.copy_clean(node, preserve_annos=None)

Creates a deep copy of an AST.

functensorflow.python.autograph.pyct.ast_util.matches(node, pattern)

Basic pattern matcher for AST.

functensorflow.python.autograph.pyct.ast_util.parallel_walk(node, other)

Walks two ASTs in parallel.

functensorflow.python.autograph.pyct.ast_util.rename_symbols(node, name_map)

Renames symbols in an AST.

classtensorflow.python.autograph.pyct.cache.CodeObjectCache

A function cache based on code objects.

classtensorflow.python.autograph.pyct.cfg.AstToCfg

Converts an AST to CFGs.

classtensorflow.python.autograph.pyct.cfg.Graph

A Control Flow Graph.

methodtensorflow.python.autograph.pyct.cfg.Graph.as_dot()

Print CFG in DOT format.

classtensorflow.python.autograph.pyct.cfg.GraphBuilder

Builder that constructs a CFG from a given AST.

methodtensorflow.python.autograph.pyct.cfg.GraphBuilder.enter_loop_section(section_id, entry_node)

Enters a loop section.

methodtensorflow.python.autograph.pyct.cfg.GraphBuilder.enter_section(section_id)

Enters a regular section.

methodtensorflow.python.autograph.pyct.cfg.GraphBuilder.exit_loop_section(section_id)

Exits a loop section.

methodtensorflow.python.autograph.pyct.cfg.GraphBuilder.exit_section(section_id)

Exits a regular section.

classtensorflow.python.autograph.pyct.cfg.GraphVisitor

Base class for a CFG visitors.

methodtensorflow.python.autograph.pyct.cfg.GraphVisitor.visit_node(node)

Visitor function.

classtensorflow.python.autograph.pyct.cfg.Node

A node in the CFG.

classtensorflow.python.autograph.pyct.common_transformers.anf.ASTEdgePattern

A pattern defining a type of AST edge.

classtensorflow.python.autograph.pyct.errors.InaccessibleSourceCodeError

Raised when inspect can not access source code.

classtensorflow.python.autograph.pyct.errors.PyCTError

Base class for all exceptions.

functensorflow.python.autograph.pyct.inspect_utils.getdefiningclass(m, owner_class)

Resolves the class (e.g.

classtensorflow.python.autograph.pyct.naming.Namer

Symbol name generator.

classtensorflow.python.autograph.pyct.origin_info.LineLocation

Similar to Location, but without column information.

classtensorflow.python.autograph.pyct.origin_info.Location

Encodes code location information.

classtensorflow.python.autograph.pyct.pretty_printer.PrettyPrinter

Print AST nodes.

classtensorflow.python.autograph.pyct.qual_names.Literal

Represents a Python numeric literal.

classtensorflow.python.autograph.pyct.qual_names.QN

Represents a qualified name.

methodtensorflow.python.autograph.pyct.qual_names.QN.ast()

AST representation.

methodtensorflow.python.autograph.pyct.qual_names.QN.ssf()

Simple symbol form.

classtensorflow.python.autograph.pyct.qual_names.QnResolver

Annotates nodes with QN information.

classtensorflow.python.autograph.pyct.qual_names.Symbol

Represents a Python symbol.

classtensorflow.python.autograph.pyct.static_analysis.activity.ActivityAnalyzer

Annotates nodes with local scope information.

methodtensorflow.python.autograph.pyct.static_analysis.activity.Scope.finalize()

Freezes this scope.

classtensorflow.python.autograph.pyct.templates.ContextAdjuster

Adjusts the ctx field of nodes to ensure consistency.

classtensorflow.python.autograph.pyct.templates.ReplaceTransformer

Replace AST nodes.

classtensorflow.python.autograph.pyct.transformer.Context

Contains information about a source code transformation.

classtensorflow.python.autograph.pyct.transformer.EntityInfo

Contains information about a Python entity.

classtensorflow.python.autograph.pyct.transpiler.GenericTranspiler

A generic transpiler for Python functions.

methodtensorflow.python.autograph.pyct.transpiler.GenericTranspiler.transform_function(fn, user_context)

Transforms a function.

classtensorflow.python.autograph.pyct.transpiler.PyToPy

A generic Python-to-Python transpiler.

methodtensorflow.python.autograph.pyct.transpiler.PyToPy.transform_function(fn, user_context)

Transforms a function.

methodtensorflow.python.autograph.utils.type_registry.TypeRegistry.lookup(obj)

Looks up 'obj'.

classtensorflow.python.checkpoint.async_checkpoint_helper.AsyncCheckpointHelper

Helper class for async checkpoint.

classtensorflow.python.checkpoint.checkpoint.Checkpoint

Manages saving/restoring trackable values to disk.

classtensorflow.python.checkpoint.checkpoint.CheckpointV1

Groups trackable objects, saving and restoring them.

classtensorflow.python.checkpoint.checkpoint_adapter.AbstractCheckpointAdapter

Abstract API for checkpoint adapter.

classtensorflow.python.checkpoint.checkpoint_adapter.ReshardCallback

API to reshard a checkpoint value during restore.

classtensorflow.python.checkpoint.checkpoint_options.CheckpointOptions

Options for constructing a Checkpoint.

classtensorflow.python.checkpoint.checkpoint_view.CheckpointView

Gathers and serializes a checkpoint view.

classtensorflow.python.checkpoint.functional_saver.MultiDeviceSaver

Saves checkpoints directly from multiple devices.

classtensorflow.python.checkpoint.graph_view.ObjectGraphView

Gathers and serializes an object graph.

classtensorflow.python.checkpoint.trackable_view.TrackableView

Gathers and serializes a trackable view.

classtensorflow.python.client.session.BaseSession

A class for interacting with a TensorFlow computation.

methodtensorflow.python.client.session.BaseSession.close()

Closes this session.

classtensorflow.python.client.session.Session

A class for running TensorFlow operations.

classtensorflow.python.client.timeline.StepStatsAnalysis

Stores the step stats analysis output.

classtensorflow.python.compiler.tensorrt.model_tests.model_handler.ModelConfig

Configurations for test models.

classtensorflow.python.compiler.tensorrt.model_tests.model_handler.ModelHandlerV1

Runs a model in TF1.

classtensorflow.python.compiler.tensorrt.model_tests.model_handler.ModelHandlerV2

Runs a model in TF2.

func

tensorflow.python.compiler.tensorrt.model_tests.model_handler.load_meta_graph(saved_model_dir:str, saved_model_tags:str, saved_model_signature_key:str) -> meta_graph_pb2.MetaGraphDef

Loads a `tf.MetaGraphDef` in TF1.

classtensorflow.python.compiler.tensorrt.model_tests.result_analyzer.ResultAnalyzer

Analyzes ModelHandlerManager results.

func

tensorflow.python.compiler.tensorrt.model_tests.result_analyzer.analyze_test_latency(test_results:model_handler.TestResultCollection, use_cpu_baseline:bool) -> DataFrame

Analyzes test latency.

func

tensorflow.python.compiler.tensorrt.model_tests.result_analyzer.analyze_test_numerics(test_results:model_handler.TestResultCollection, use_cpu_baseline:bool) -> (DataFrame, str)

Analyzes test numerics.

func

tensorflow.python.compiler.tensorrt.model_tests.result_analyzer.extract_test_info(test_results:model_handler.TestResultCollection) -> DataFrame

Extracts the test information.

classtensorflow.python.compiler.tensorrt.trt_convert.TrtConversionParams

Parameters that are used for TF-TRT conversion.

classtensorflow.python.data.experimental.ops.grouping.Reducer

A reducer is used for reducing a set of elements.

classtensorflow.python.data.experimental.ops.random_ops.RandomDatasetV1

A `Dataset` of pseudorandom values.

classtensorflow.python.data.experimental.ops.random_ops.RandomDatasetV2

A `Dataset` of pseudorandom values.

classtensorflow.python.data.experimental.ops.writers.TFRecordWriter

Writes a dataset to a TFRecord file.

classtensorflow.python.data.experimental.service.server_lib.DispatchServer

An in-process tf.data service dispatch server.

methodtensorflow.python.data.experimental.service.server_lib.DispatchServer.join() -> None

Blocks until the server has shut down.

methodtensorflow.python.data.experimental.service.server_lib.DispatchServer.start()

Starts this server.

methodtensorflow.python.data.experimental.service.server_lib.DispatchServer.stop() -> None

Stops the server.

classtensorflow.python.data.experimental.service.server_lib.WorkerServer

An in-process tf.data service worker server.

methodtensorflow.python.data.experimental.service.server_lib.WorkerServer.join() -> None

Blocks until the server has shut down.

methodtensorflow.python.data.experimental.service.server_lib.WorkerServer.start() -> None

Starts this server.

methodtensorflow.python.data.experimental.service.server_lib.WorkerServer.stop() -> None

Stops the server.

classtensorflow.python.data.kernel_tests.checkpoint_test_base.CheckpointTestBase

Base test class for checkpointing datasets.

classtensorflow.python.data.kernel_tests.tf_record_test_base.TFRecordTestBase

Base class for TFRecord-based tests.

classtensorflow.python.data.ops.cache_op.CacheDataset

A `Dataset` that caches elements of its input.

classtensorflow.python.data.ops.dataset_ops.DatasetSource

Abstract class representing a dataset with no inputs.

classtensorflow.python.data.ops.dataset_ops.DatasetSpec

Type specification for `tf.data.Dataset`.

methodtensorflow.python.data.ops.dataset_ops.DatasetSpec.element_spec()

The inner element spec.

methodtensorflow.python.data.ops.dataset_ops.DatasetSpec.is_subtype_of(other)

See base class.

methodtensorflow.python.data.ops.dataset_ops.DatasetSpec.most_specific_common_supertype(others)

See base class.

classtensorflow.python.data.ops.dataset_ops.DatasetV1

Represents a potentially large set of elements.

classtensorflow.python.data.ops.dataset_ops.DatasetV2

Represents a potentially large set of elements.

methodtensorflow.python.data.ops.dataset_ops.DatasetV2.cache(filename='', name=None) -> 'DatasetV2'

Caches the elements in this dataset.

methodtensorflow.python.data.ops.dataset_ops.DatasetV2.enumerate(start=0, name=None) -> 'DatasetV2'

Enumerates the elements of this dataset.

methodtensorflow.python.data.ops.dataset_ops.DatasetV2.ignore_errors(log_warning=False, name=None) -> 'DatasetV2'

Drops elements that cause errors.

method

tensorflow.python.data.ops.dataset_ops.DatasetV2.load(path, element_spec=None, compression=None, reader_func=None, wait=False) -> 'DatasetV2'

Loads a previously saved dataset.

methodtensorflow.python.data.ops.dataset_ops.DatasetV2.range(*args, **kwargs) -> 'DatasetV2'

Creates a `Dataset` of a step-separated range of values.

method

tensorflow.python.data.ops.dataset_ops.DatasetV2.window(size, shift=None, stride=1, drop_remainder=False, name=None) -> 'DatasetV2'

Returns a dataset of "windows".

methodtensorflow.python.data.ops.dataset_ops.DatasetV2.zip(*datasets=None, *name=None, *args) -> 'DatasetV2'

Creates a `Dataset` by zipping together the given datasets.

classtensorflow.python.data.ops.dataset_ops.UnaryDataset

Abstract class representing a dataset with one input.

functensorflow.python.data.ops.dataset_ops.apply_rewrite(dataset, rewrite)

Applies a rewrite to a dataset.

func

tensorflow.python.data.ops.dataset_ops.make_initializable_iterator(dataset:DatasetV1, shared_name=None) -> iterator_ops.Iterator

Creates an iterator for elements of `dataset`.

functensorflow.python.data.ops.debug_mode.enable_debug_mode()

Enables debug mode for tf.data.

classtensorflow.python.data.ops.iterator_ops.Iterator

Represents the state of iterating through a `Dataset`.

methodtensorflow.python.data.ops.iterator_ops.Iterator.get_next(name=None)

Returns the next element.

classtensorflow.python.data.ops.iterator_ops.IteratorBase

Represents an iterator of a `tf.data.Dataset`.

methodtensorflow.python.data.ops.iterator_ops.IteratorBase.get_next()

Returns the next element.

classtensorflow.python.data.ops.iterator_ops.IteratorSpec

Type specification for `tf.data.Iterator`.

classtensorflow.python.data.ops.multi_device_iterator_ops.MultiDeviceIterator

An iterator over multiple devices.

classtensorflow.python.data.ops.multi_device_iterator_ops.OwnedMultiDeviceIterator

An iterator over multiple devices.

classtensorflow.python.data.ops.optional_ops.Optional

Represents a value that may or may not be present.

classtensorflow.python.data.ops.optional_ops.OptionalSpec

Type specification for `tf.experimental.Optional`.

classtensorflow.python.data.ops.options.AutoShardPolicy

Represents the type of auto-sharding to use.

classtensorflow.python.data.ops.options.AutotuneAlgorithm

Represents the type of autotuning algorithm to use.

classtensorflow.python.data.ops.options.DistributeOptions

Represents options for distributed data processing.

classtensorflow.python.data.ops.options.OptimizationOptions

Represents options for dataset optimizations.

classtensorflow.python.data.ops.options.Options

Represents options for `tf.data.Dataset`.

classtensorflow.python.data.ops.options.ServiceOptions

Represents options for tf.data service.

classtensorflow.python.data.ops.options.ThreadingOptions

Represents options for dataset threading.

classtensorflow.python.data.util.options.OptionsBase

Base class for representing a set of tf.data options.

functensorflow.python.data.util.options.create_option(name, ty, docstring, default_factory=lambda: None)

Creates a type-checked property.

functensorflow.python.data.util.sparse.any_sparse(classes)

Checks for sparse tensor.

functensorflow.python.data.util.sparse.serialize_sparse_tensors(tensors)

Serializes sparse tensors.

classtensorflow.python.debug.cli.analyzer_cli.DebugAnalyzer

Analyzer for debug data from dump directories.

classtensorflow.python.debug.cli.base_ui.BaseUI

Base class of tfdbg user interface.

classtensorflow.python.debug.cli.command_parser.Interval

Represents an interval between a start and end value.

classtensorflow.python.debug.cli.debugger_cli_common.CommandHandlerRegistry

Registry of command handlers for CLI.

classtensorflow.python.debug.cli.debugger_cli_common.CommandHistory

Keeps command history and supports lookup.

classtensorflow.python.debug.cli.debugger_cli_common.Menu

A class for text-based menu.

methodtensorflow.python.debug.cli.debugger_cli_common.Menu.append(item)

Append an item to the Menu.

classtensorflow.python.debug.cli.debugger_cli_common.MenuItem

A class for an item in a text-based menu.

classtensorflow.python.debug.cli.debugger_cli_common.RichLine

Rich single-line text.

classtensorflow.python.debug.cli.debugger_cli_common.RichTextLines

Rich multi-line text.

methodtensorflow.python.debug.cli.debugger_cli_common.RichTextLines.append(line, font_attr_segs=None)

Append a single line of text.

classtensorflow.python.debug.cli.debugger_cli_common.TabCompletionRegistry

Registry for tab completion responses.

classtensorflow.python.debug.cli.profile_analyzer_cli.ProfileAnalyzer

Analyzer for profiling data.

classtensorflow.python.debug.cli.profile_analyzer_cli.ProfileDataTableView

Table View of profiling data.

classtensorflow.python.debug.cli.readline_ui.ReadlineUI

Readline-based Command-line UI.

classtensorflow.python.debug.cli.tensor_format.HighlightOptions

Options for highlighting elements of a tensor.

classtensorflow.python.debug.lib.debug_data.DebugDumpDir

Data set from a debug-dump directory on filesystem.

methodtensorflow.python.debug.lib.debug_data.DebugDumpDir.devices()

Get the list of device names.

methodtensorflow.python.debug.lib.debug_data.DebugDumpDir.partition_graphs()

Get the partition graphs.

methodtensorflow.python.debug.lib.debug_data.DebugDumpDir.python_graph()

Get the Python graph.

methodtensorflow.python.debug.lib.debug_data.DebugTensorDatum.debug_op()

Name of the debug op.

methodtensorflow.python.debug.lib.debug_data.DebugTensorDatum.dump_size_bytes()

Size of the dump file.

classtensorflow.python.debug.lib.debug_events_monitors.BaseMonitor

Base class for debug event data monitors.

classtensorflow.python.debug.lib.debug_events_monitors.InfNanAlert

Alert for Infinity and NaN values.

classtensorflow.python.debug.lib.debug_events_monitors.InfNanMonitor

Monitor for Infinity and NaN in tensor values.

classtensorflow.python.debug.lib.debug_events_reader.BaseDigest

Base class for digest.

classtensorflow.python.debug.lib.debug_events_reader.DebugEventsReader

Reader class for a tfdbg v2 DebugEvents directory.

classtensorflow.python.debug.lib.debug_events_writer.DebugEventsWriter

A writer for TF debugging events.

methodtensorflow.python.debug.lib.debug_events_writer.DebugEventsWriter.Close()

Close the writer.

classtensorflow.python.debug.lib.debug_gradients.GradientsDebugger

Gradients Debugger.

classtensorflow.python.debug.lib.debug_graphs.DFSGraphTracer

Graph input tracer using depth-first search.

methodtensorflow.python.debug.lib.debug_graphs.DFSGraphTracer.trace(graph_element_name)

Trace inputs.

classtensorflow.python.debug.lib.debug_graphs.DebugGraph

Represents a debugger-decorated graph.

classtensorflow.python.debug.lib.dumping_callback_test_lib.DumpingCallbackTestBase

Base test-case class for tfdbg v2 callbacks.

classtensorflow.python.debug.lib.grpc_debug_server.EventListenerBaseServicer

Base Python class for gRPC debug server.

classtensorflow.python.debug.lib.profiling.ProfileDatum

Profile data point.

classtensorflow.python.debug.wrappers.framework.BaseDebugWrapperSession

Base class of debug-wrapper session classes.

classtensorflow.python.debug.wrappers.framework.OnRunEndRequest

Request to an on-run-end callback.

classtensorflow.python.debug.wrappers.framework.OnRunEndResponse

Response from an on-run-end callback.

classtensorflow.python.debug.wrappers.framework.OnRunStartRequest

Request to an on-run-start callback.

classtensorflow.python.debug.wrappers.framework.OnRunStartResponse

Request from an on-run-start callback.

classtensorflow.python.debug.wrappers.framework.OnSessionInitRequest

Request to an on-session-init callback.

classtensorflow.python.debug.wrappers.framework.OnSessionInitResponse

Response from an on-session-init callback.

classtensorflow.python.debug.wrappers.framework.WatchOptions

Type for return values of watch_fn.

classtensorflow.python.debug.wrappers.hooks.LocalCLIDebugHook

Command-line-interface debugger hook.

classtensorflow.python.distribute.cluster_resolver.kubernetes_cluster_resolver.ExecutableLocation

Defines where the executable runs on.

classtensorflow.python.distribute.cluster_resolver.kubernetes_cluster_resolver.KubernetesClusterResolver

ClusterResolver for Kubernetes.

classtensorflow.python.distribute.cluster_resolver.tpu.tpu_cluster_resolver.TPUClusterResolver

Cluster Resolver for Google Cloud TPUs.

classtensorflow.python.distribute.collective_util.CommunicationImplementation

Cross device communication implementation.

classtensorflow.python.distribute.collective_util.Hints

Hints for collective operations like AllReduce.

classtensorflow.python.distribute.collective_util.Options

Implementation of OptionsInterface.

classtensorflow.python.distribute.combinations.ClusterCombination

Sets up multi worker tests.

classtensorflow.python.distribute.combinations.DistributionCombination

Sets up distribution strategy for tests.

classtensorflow.python.distribute.combinations.DistributionParameter

Transforms arguments of type `NamedDistribution`.

classtensorflow.python.distribute.combinations.TestEnvironment

Holds the test environment information.

functensorflow.python.distribute.combinations.concat(*combined)

Concats combinations.

classtensorflow.python.distribute.coordinator.cluster_coordinator.Closure

Hold a function to be scheduled and its arguments.

classtensorflow.python.distribute.coordinator.cluster_coordinator.ClosureInputError

Wrapper for errors from resource building.

classtensorflow.python.distribute.coordinator.cluster_coordinator.Cluster

A cluster with workers.

classtensorflow.python.distribute.coordinator.cluster_coordinator.ResourceClosure

A closure that builds a resource on a worker.

classtensorflow.python.distribute.coordinator.cluster_coordinator.Worker

A worker in a cluster.

classtensorflow.python.distribute.coordinator.cluster_coordinator.WorkerPreemptionHandler

Handles worker preemptions.

classtensorflow.python.distribute.coordinator.remote_value.RemoteValueStatus

The status of a `RemoteValue` object.

classtensorflow.python.distribute.coordinator.values.PerWorkerValuesTypeSpec

TypeSpec for PerWorkerValues.

classtensorflow.python.distribute.coordinator.values.RemoteValueImpl

Implementation of `RemoteValue`.

classtensorflow.python.distribute.cross_device_ops.AllReduceCrossDeviceOps

All-reduce implementation of CrossDeviceOps.

classtensorflow.python.distribute.cross_device_ops.NcclAllReduce

NCCL all-reduce implementation of CrossDeviceOps.

classtensorflow.python.distribute.cross_device_utils.CollectiveKeys

Class that manages collective keys.

classtensorflow.python.distribute.cross_device_utils.CollectiveReplicaLauncher

Launch collectives on one replica.

method

tensorflow.python.distribute.cross_device_utils.CollectiveReplicaLauncher.all_gather(input_tensor:core.TensorLike, axis:core.TensorLike, options:Optional[collective_util.Options]=None) -> core.Tensor

All-gather a dense tensor.

functensorflow.python.distribute.device_util.canonicalize(d, default=None)

Canonicalize device string.

classtensorflow.python.distribute.distribute_config.DistributeConfig

A config tuple for distribution strategies.

classtensorflow.python.distribute.distribute_coordinator.CoordinatorMode

Specify how distribute coordinator runs.

classtensorflow.python.distribute.distribute_lib.InputReplicationMode

Replication mode for input function.

classtensorflow.python.distribute.distribute_lib.RunOptions

Run options for `strategy.run`.

classtensorflow.python.distribute.experimental.rpc.rpc_ops.Client

Client class for invoking RPCs to the server.

classtensorflow.python.distribute.failure_handling.failure_handling.BorgTPUTerminationConfig

Configurations for Borg.

classtensorflow.python.distribute.failure_handling.failure_handling.BorgTerminationConfig

Configurations for Borg.

classtensorflow.python.distribute.failure_handling.failure_handling.GcpCpuTerminationConfig

Configurations for GCP CPU VM.

classtensorflow.python.distribute.failure_handling.failure_handling.GcpGpuTerminationConfig

Configurations for GCP GPU VM.

classtensorflow.python.distribute.failure_handling.preemption_watcher.PreemptionWatcher

Watch preemption signal and store it.

classtensorflow.python.distribute.input_lib.DistributedDatasetSpec

Type specification for `DistributedDataset.

classtensorflow.python.distribute.input_lib.DistributedDatasetsFromFunction

Inputs created from dataset function.

classtensorflow.python.distribute.input_lib.DistributedIterator

Input Iterator for a distributed dataset.

classtensorflow.python.distribute.input_lib.DistributedIteratorBase

Common implementation for all input iterators.

classtensorflow.python.distribute.input_lib.DistributedIteratorSpec

Type specification for `DistributedIterator`.

classtensorflow.python.distribute.load_context.LoadContext

A context for loading a model.

classtensorflow.python.distribute.mirrored_strategy.MirroredExtended

Implementation of MirroredStrategy.

classtensorflow.python.distribute.multi_process_lib.AbslForkServerProcess

An absl-compatible Forkserver process.

classtensorflow.python.distribute.multi_worker_test_base.IndependentWorkerTestBase

Testing infra for independent workers.

classtensorflow.python.distribute.multi_worker_test_base.MockOsEnv

A class that allows per-thread TF_CONFIG.

classtensorflow.python.distribute.one_device_strategy.OneDeviceExtended

Implementation of OneDeviceStrategy.

classtensorflow.python.distribute.reduce_util.ReduceOp

Indicates how a set of values should be reduced.

classtensorflow.python.distribute.sharded_variable.ShardedVariableMixin

Mixin for ShardedVariable.

methodtensorflow.python.distribute.sharded_variable.ShardedVariableMixin.name()

The name of this object.

classtensorflow.python.distribute.sharded_variable.ShardedVariableSpec

Type specification for a `ShardedVariable`.

classtensorflow.python.distribute.step_fn.Step

Interface for performing each step of a training algorithm.

classtensorflow.python.distribute.strategy_test_lib.RemoteSingleWorkerMirroredStrategyBase

Tests for a Remote single worker.

classtensorflow.python.distribute.tpu_strategy.TPUExtended

Implementation of TPUStrategy.

classtensorflow.python.distribute.tpu_strategy.TPUStrategy

Synchronous training on TPUs and TPU Pods.

methodtensorflow.python.distribute.tpu_strategy.TPUStrategy.run(fn, args=(), kwargs=None, options=None)

See base class.

classtensorflow.python.distribute.tpu_strategy.TPUStrategyV1

TPU distribution strategy implementation.

classtensorflow.python.distribute.tpu_strategy.TPUStrategyV2

Synchronous training on TPUs and TPU Pods.

classtensorflow.python.distribute.tpu_util.LazyVariableTracker

Class to track uninitialized lazy variables.

classtensorflow.python.distribute.tpu_util.TPUUninitializedVariable

UninitializedVariable component for TPU.

classtensorflow.python.distribute.tpu_values.TPUDistributedVariable

DistributedVariable subclass for TPUStrategy.

classtensorflow.python.distribute.tpu_values.TPUVariableMixin

Mixin for TPU variables.

classtensorflow.python.distribute.v1.input_lib.DatasetIterator

Iterator created from input dataset.

classtensorflow.python.distribute.v1.input_lib.DistributedDatasetsFromFunctionV1

Inputs created from dataset function.

classtensorflow.python.distribute.v1.input_lib.DistributedIteratorV1

Input Iterator for a distributed dataset.

classtensorflow.python.distribute.v1.input_lib.InputFunctionIterator

Iterator created from input function.

classtensorflow.python.distribute.values.DistributedValues

Base class for representing distributed values.

classtensorflow.python.distribute.values.DistributedVarOp

A class that looks like `tf.Operation`.

classtensorflow.python.distribute.values.DistributedVariable

Holds a map from replica to variables.

classtensorflow.python.distribute.values.DistributedVariableTraceType

TraceType of DistributedVariable objects.

classtensorflow.python.distribute.values.PerReplica

Holds a map from replica to unsynchronized values.

methodtensorflow.python.distribute.values.PerReplica.values()

Returns the per replica values.

classtensorflow.python.distribute.values.PerReplicaSpec

Type specification for a `PerReplica`.

classtensorflow.python.distribute.values_v2.DistributedVariable

Represents variables that are replicated.

classtensorflow.python.eager.backprop.GradientTape

Record operations for automatic differentiation.

classtensorflow.python.eager.benchmarks_test_base.MicroBenchmarksBase

Run and report benchmark results.

classtensorflow.python.eager.cancellation.CancellationManager

A mechanism for cancelling blocking computation.

classtensorflow.python.eager.context.Context

Environment in which eager operations execute.

About this data

These signatures were extracted from the public source of tensorflow/tensorflow using Python's ast module. Argument names, default values, type annotations and return types are taken verbatim from the code. Implementation bodies are never stored. See how it works for details.

Back to all 805 libraries