[][src]Crate tensorflow

This crate provides Rust bindings for the TensorFlow machine learning library.

If you aren't sure how to use something, please see the examples folder.

Modules

expr

This module builds computation graphs.

io

A module for reading and writing TFRecords, Tensorflow's preferred on-disk data format.

ops

This module exposes functions for building standard operations.

train

This module supports building and training models.

Structs

AttrMetadata

AttrMetadata describes the value of an attribute on an operation.

BFloat16

BFloat16 provides a Rust type for BFloat16. Note that this is not the same as half::f16. BFloat16 is not an IEEE-754 16-bit float. See https://github.com/tensorflow/tensorflow/blob/master/tensorflow/core/framework/bfloat16.h for details.

Device

Metadata about a device.

FetchToken

An opaque token for retrieving an output from a computation.

Function

Function is a grouping of operations with defined inputs and outputs. Once created and added to graphs, functions can be invoked by creating an operation whose operation type matches the function name.

FunctionOptions

Options that can be passed during function creation.

Graph

Represents a computation graph. Graphs may be shared between sessions. Graphs are thread-safe when used as directed.

ImportGraphDefOptions

ImportGraphDefOptions holds options that can be passed to Graph::import_graph_def.

ImportGraphDefResults

ImportGraphDefResults holds results that are generated by Graph::import_graph_def_with_results().

Input

A Input is one end of a graph edge. It holds an operation and an index into the inputs of that operation.

Library

Dynamically loaded plugins. The C API doesn't provide a way to unload libraries, so nothing happens when this goes out of scope.

MetaGraphDef

Contains data necessary to restart training, run inference. It can be used to serialize/de-serialize memory objects necessary for running computation in a graph when crossing the process boundary. It can be used for long term storage of graphs, cross-language execution of graphs, etc.

Operation

An Operation is a node in a Graph. It is a computation which accepts inputs and produces outputs.

OperationDescription

An OperationDescription is an Operation in the process of being built (i.e. the builder pattern).

OperationIter

Iterator over the operations in a Graph.

Output

A Output is one end of a graph edge. It holds an operation and an index into the outputs of that operation.

OutputName

Names a specific Output in the graph.

QInt8

Quantized type for i8.

QInt16

Quantized type for i16.

QInt32

Quantized type for i32.

QUInt8

Quantized type for u8.

QUInt16

Quantized type for u16.

SaveModelError

Error generated while saving a model.

SavedModelBuilder

Builds a SavedModelSaver, which can be used to save models.

SavedModelBundle

Aggregation type for a saved model bundle.

SavedModelSaver

Creates saved models. Use a SavedModelBuilder to create a SavedModelSaver.

Scope

A Scope object represents a set of related TensorFlow ops that have the same properties such as a common name prefix.

Session

Manages a single graph and execution.

SessionOptions

Options that can be passed during session creation.

SessionRunArgs

Manages the inputs and outputs for a single execution of a graph.

Shape

A Shape is the shape of a tensor. A Shape may be an unknown rank, or it may have a known rank with each dimension being known or unknown.

SignatureDef

SignatureDef defines the signature of a computation supported by a TensorFlow graph.

Status

Holds error information when communicating with back and forth with tensorflow.

Tensor

Holds a multi-dimensional array of elements of a single data type.

TensorInfo

Information about a Tensor necessary for feeding or retrieval.

Variable

Holds state in the form of a tensor that persists across steps.

VariableBuilder

Builds a Variable.

WhileBuilder

A WhileBuilder is used to build a while loop.

Enums

AttrType

Describes the type of the value of an attribute on an operation.

Code

Error values that can be returned.

DataType

Type of a single tensor element.

Constants

CLASSIFY_INPUTS

Classification inputs.

CLASSIFY_METHOD_NAME

Classification method name used in a SignatureDef.

CLASSIFY_OUTPUT_CLASSES

Classification classes output.

CLASSIFY_OUTPUT_SCORES

Classification scores output.

DEFAULT_SERVING_SIGNATURE_DEF_KEY

Key in the signature def map for default serving signatures. The default signature is used in inference requests where a specific signature was not specified.

PREDICT_INPUTS

Predict inputs.

PREDICT_METHOD_NAME

Prediction method name used in a SignatureDef.

PREDICT_OUTPUTS

Predict outputs.

REGRESS_INPUTS

Regression inputs.

REGRESS_METHOD_NAME

Regression method name used in a SignatureDef.

REGRESS_OUTPUTS

Regression outputs.

Traits

TensorType

A Rust type that maps to a DataType.

Functions

get_all_registered_kernels

Returns a serialized KernelList protocol buffer containing KernelDefs for all registered kernels.

get_registered_kernels_for_op

Returns a serialized KernelList protocol buffer containing KernelDefs for all kernels registered for the operation named name.

version

Returns a string describing version information of the TensorFlow library. TensorFlow is using semantic versioning.

Type Definitions

OutputTokenDeprecated

Deprecated alias for FetchToken.

Result

Convenience type for Result with Status as the error type.

StepWithGraphDeprecated

Deprecated alias for SessionRunArgs.