PrimitiveOp

public abstract class PrimitiveOp

Known Direct Subclasses

Abort, All, AllToAll<T>, AnonymousHashTable, AnonymousIteratorV2, AnonymousIteratorV3, AnonymousMemoryCache, AnonymousMultiDeviceIterator, AnonymousMultiDeviceIteratorV3, AnonymousMutableDenseHashTable, AnonymousMutableHashTable, AnonymousMutableHashTableOfTensors, AnonymousRandomSeedGenerator, AnonymousSeedGenerator, Any, and 745 others.

Abort	Raise a exception to abort the process when called.
All	Computes the "logical and" of elements across dimensions of a tensor.
AllToAll<T>	An Op to exchange data across TPU replicas.
AnonymousHashTable	Creates a uninitialized anonymous hash table.
AnonymousIteratorV2	A container for an iterator resource.
AnonymousIteratorV3	A container for an iterator resource.
AnonymousMemoryCache
AnonymousMultiDeviceIterator	A container for a multi device iterator resource.
AnonymousMultiDeviceIteratorV3	A container for a multi device iterator resource.
AnonymousMutableDenseHashTable	Creates an empty anonymous mutable hash table that uses tensors as the backing store.
AnonymousMutableHashTable	Creates an empty anonymous mutable hash table.
AnonymousMutableHashTableOfTensors	Creates an empty anonymous mutable hash table of vector values.
AnonymousRandomSeedGenerator
AnonymousSeedGenerator
Any	Computes the "logical or" of elements across dimensions of a tensor.
ApplyAdagradV2<T>	Update '*var' according to the adagrad scheme.
ApproxTopK<T extends Number>	Returns min/max k values and their indices of the input operand in an approximate manner.
AssertCardinalityDataset
AssertNextDataset	A transformation that asserts which transformations happen next.
AssertPrevDataset	A transformation that asserts which transformations happened previously.
AssertThat	Asserts that the given condition is true.
Assign<T>	Update 'ref' by assigning 'value' to it.
AssignAdd<T>	Update 'ref' by adding 'value' to it.
AssignAddVariableOp	Adds a value to the current value of a variable.
AssignSub<T>	Update 'ref' by subtracting 'value' from it.
AssignSubVariableOp	Subtracts a value from the current value of a variable.
AssignVariableOp	Assigns a new value to a variable.
AssignVariableXlaConcatND	Concats input tensor across all dimensions.
AutoShardDataset	Creates a dataset that shards the input dataset.
BandedTriangularSolve<T>
Barrier	Defines a barrier that persists across different graph executions.
BarrierClose	Closes the given barrier.
BarrierIncompleteSize	Computes the number of incomplete elements in the given barrier.
BarrierInsertMany	For each key, assigns the respective value to the specified component.
BarrierReadySize	Computes the number of complete elements in the given barrier.
BarrierTakeMany	Takes the given number of completed elements from a barrier.
Batch	Batches all input tensors nondeterministically.
BatchMatMulV2<T>	Multiplies slices of two tensors in batches.
BatchMatMulV3<V>	Multiplies slices of two tensors in batches.
BatchToSpace<T>	BatchToSpace for 4-D tensors of type T.
BatchToSpaceNd<T>	BatchToSpace for N-D tensors of type T.
BesselI0<T extends Number>
BesselI1<T extends Number>
BesselJ0<T extends Number>
BesselJ1<T extends Number>
BesselK0<T extends Number>
BesselK0e<T extends Number>
BesselK1<T extends Number>
BesselK1e<T extends Number>
BesselY0<T extends Number>
BesselY1<T extends Number>
Bitcast<U>	Bitcasts a tensor from one type to another without copying data.
BlockLSTM<T extends Number>	Computes the LSTM cell forward propagation for all the time steps.
BlockLSTMGrad<T extends Number>	Computes the LSTM cell backward propagation for the entire time sequence.
BlockLSTMGradV2<T extends Number>	Computes the LSTM cell backward propagation for the entire time sequence.
BlockLSTMV2<T extends Number>	Computes the LSTM cell forward propagation for all the time steps.
BoostedTreesAggregateStats	Aggregates the summary of accumulated stats for the batch.
BoostedTreesBucketize	Bucketize each feature based on bucket boundaries.
BoostedTreesCalculateBestFeatureSplit	Calculates gains for each feature and returns the best possible split information for the feature.
BoostedTreesCalculateBestFeatureSplitV2	Calculates gains for each feature and returns the best possible split information for each node.
BoostedTreesCalculateBestGainsPerFeature	Calculates gains for each feature and returns the best possible split information for the feature.
BoostedTreesCenterBias	Calculates the prior from the training data (the bias) and fills in the first node with the logits' prior.
BoostedTreesCreateEnsemble	Creates a tree ensemble model and returns a handle to it.
BoostedTreesCreateQuantileStreamResource	Create the Resource for Quantile Streams.
BoostedTreesDeserializeEnsemble	Deserializes a serialized tree ensemble config and replaces current tree ensemble.
BoostedTreesEnsembleResourceHandleOp	Creates a handle to a BoostedTreesEnsembleResource
BoostedTreesExampleDebugOutputs	Debugging/model interpretability outputs for each example.
BoostedTreesFlushQuantileSummaries	Flush the quantile summaries from each quantile stream resource.
BoostedTreesGetEnsembleStates	Retrieves the tree ensemble resource stamp token, number of trees and growing statistics.
BoostedTreesMakeQuantileSummaries	Makes the summary of quantiles for the batch.
BoostedTreesMakeStatsSummary	Makes the summary of accumulated stats for the batch.
BoostedTreesPredict	Runs multiple additive regression ensemble predictors on input instances and computes the logits.
BoostedTreesQuantileStreamResourceAddSummaries	Add the quantile summaries to each quantile stream resource.
BoostedTreesQuantileStreamResourceDeserialize	Deserialize bucket boundaries and ready flag into current QuantileAccumulator.
BoostedTreesQuantileStreamResourceFlush	Flush the summaries for a quantile stream resource.
BoostedTreesQuantileStreamResourceGetBucketBoundaries	Generate the bucket boundaries for each feature based on accumulated summaries.
BoostedTreesQuantileStreamResourceHandleOp	Creates a handle to a BoostedTreesQuantileStreamResource.
BoostedTreesSerializeEnsemble	Serializes the tree ensemble to a proto.
BoostedTreesSparseAggregateStats	Aggregates the summary of accumulated stats for the batch.
BoostedTreesSparseCalculateBestFeatureSplit	Calculates gains for each feature and returns the best possible split information for the feature.
BoostedTreesTrainingPredict	Runs multiple additive regression ensemble predictors on input instances and computes the update to cached logits.
BoostedTreesUpdateEnsemble	Updates the tree ensemble by either adding a layer to the last tree being grown or by starting a new tree.
BoostedTreesUpdateEnsembleV2	Updates the tree ensemble by adding a layer to the last tree being grown or by starting a new tree.
BroadcastDynamicShape<T extends Number>	Return the shape of s0 op s1 with broadcast.
BroadcastGradientArgs<T extends Number>	Return the reduction indices for computing gradients of s0 op s1 with broadcast.
BroadcastTo<T>	Broadcast an array for a compatible shape.
Bucketize	Bucketizes 'input' based on 'boundaries'.
CSRSparseMatrixComponents<T>	Reads out the CSR components at batch `index`.
CSRSparseMatrixToDense<T>	Convert a (possibly batched) CSRSparseMatrix to dense.
CSRSparseMatrixToSparseTensor<T>	Converts a (possibly batched) CSRSparesMatrix to a SparseTensor.
CSVDataset
CSVDatasetV2
CTCLossV2	Calculates the CTC Loss (log probability) for each batch entry.
CacheDatasetV2
CheckNumericsV2<T extends Number>	Checks a tensor for NaN, -Inf and +Inf values.
ChooseFastestDataset
ClipByValue<T>	Clips tensor values to a specified min and max.
CollateTPUEmbeddingMemory	An op that merges the string-encoded memory config protos from all hosts.
CollectiveAllToAllV2<T extends Number>	Mutually exchanges multiple tensors of identical type and shape.
CollectiveAllToAllV3<T extends Number>	Mutually exchanges multiple tensors of identical type and shape.
CollectiveAssignGroupV2	Assign group keys based on group assignment.
CollectiveBcastRecvV2<U>	Receives a tensor value broadcast from another device.
CollectiveBcastSendV2<T>	Broadcasts a tensor value to one or more other devices.
CollectiveGather<T extends Number>	Mutually accumulates multiple tensors of identical type and shape.
CollectiveGatherV2<T extends Number>	Mutually accumulates multiple tensors of identical type and shape.
CollectiveInitializeCommunicator	Initializes a group for collective operations.
CollectivePermute<T>	An Op to permute tensors across replicated TPU instances.
CollectiveReduceScatterV2<T extends Number>	Mutually reduces multiple tensors of identical type and shape and scatters the result.
CollectiveReduceV2<T extends Number>	Mutually reduces multiple tensors of identical type and shape.
CollectiveReduceV3<T extends Number>	Mutually reduces multiple tensors of identical type and shape.
CombinedNonMaxSuppression	Greedily selects a subset of bounding boxes in descending order of score, This operation performs non_max_suppression on the inputs per batch, across all classes.
CompositeTensorVariantFromComponents	Encodes an `ExtensionType` value into a `variant` scalar Tensor.
CompositeTensorVariantToComponents	Decodes a `variant` scalar Tensor into an `ExtensionType` value.
CompressElement	Compresses a dataset element.
ComputeBatchSize	Computes the static batch size of a dataset sans partial batches.
ComputeDedupDataTupleMask	An op computes tuple mask of deduplication data from embedding core.
Concat<T>	Concatenates tensors along one dimension.
ConfigureAndInitializeGlobalTPU	An op that sets up the centralized structures for a distributed TPU system.
ConfigureDistributedTPU	Sets up the centralized structures for a distributed TPU system.
ConfigureTPUEmbedding	Sets up TPUEmbedding in a distributed TPU system.
ConfigureTPUEmbeddingHost	An op that configures the TPUEmbedding software on a host.
ConfigureTPUEmbeddingMemory	An op that configures the TPUEmbedding software on a host.
ConnectTPUEmbeddingHosts	An op that sets up communication between TPUEmbedding host software instances after ConfigureTPUEmbeddingHost has been called on each host.
Constant<T>	An operator producing a constant value.
ConsumeMutexLock	This op consumes a lock created by `MutexLock`.
ControlTrigger	Does nothing.
Conv<T extends Number>	Computes a N-D convolution given (N+1+batch_dims)-D `input` and (N+2)-D `filter` tensors.
Conv2DBackpropFilterV2<T extends Number>	Computes the gradients of convolution with respect to the filter.
Conv2DBackpropInputV2<T extends Number>	Computes the gradients of convolution with respect to the input.
Copy<T>	Copy a tensor from CPU-to-CPU or GPU-to-GPU.
CopyHost<T>	Copy a tensor to host.
CopyToMesh<T>
CopyToMeshGrad<T>
CountUpTo<T extends Number>	Increments 'ref' until it reaches 'limit'.
CrossReplicaSum<T extends Number>	An Op to sum inputs across replicated TPU instances.
CudnnRNNBackpropV3<T extends Number>	Backprop step of CudnnRNNV3.
CudnnRNNCanonicalToParamsV2<T extends Number>	Converts CudnnRNN params from canonical form to usable form.
CudnnRNNParamsToCanonicalV2<T extends Number>	Retrieves CudnnRNN params in canonical form.
CudnnRNNV3<T extends Number>	A RNN backed by cuDNN.
CumulativeLogsumexp<T extends Number>	Compute the cumulative product of the tensor `x` along `axis`.
DTensorRestoreV2
DTensorSetGlobalTPUArray	An op that informs a host of the global ids of all the of TPUs in the system.
DataServiceDataset	Creates a dataset that reads data from the tf.data service.
DataServiceDatasetV2	Creates a dataset that reads data from the tf.data service.
DatasetCardinality	Returns the cardinality of `input_dataset`.
DatasetFromGraph	Creates a dataset from the given `graph_def`.
DatasetToGraphV2	Returns a serialized GraphDef representing `input_dataset`.
Dawsn<T extends Number>
DebugGradientIdentity<T>	Identity op for gradient debugging.
DebugGradientRefIdentity<T>	Identity op for gradient debugging.
DebugIdentity<T>	Provides an identity mapping of the non-Ref type input tensor for debugging.
DebugIdentityV2<T>	Debug Identity V2 Op.
DebugIdentityV3<T>	Provides an identity mapping of the non-Ref type input tensor for debugging.
DebugNanCount	Debug NaN Value Counter Op.
DebugNumericSummary	Debug Numeric Summary Op.
DebugNumericSummaryV2<U extends Number>	Debug Numeric Summary V2 Op.
DecodeImage<T extends Number>	Function for decode_bmp, decode_gif, decode_jpeg, and decode_png.
DecodePaddedRaw<T extends Number>	Reinterpret the bytes of a string as a vector of numbers.
DecodeProto	The op extracts fields from a serialized protocol buffers message into tensors.
DeepCopy<T>	Makes a copy of `x`.
DeleteIterator	A container for an iterator resource.
DeleteMemoryCache
DeleteMultiDeviceIterator	A container for an iterator resource.
DeleteRandomSeedGenerator
DeleteSeedGenerator
DeleteSessionTensor	Delete the tensor specified by its handle in the session.
DenseBincount<U extends Number>	Counts the number of occurrences of each value in an integer array.
DenseCountSparseOutput<U extends Number>	Performs sparse-output bin counting for a tf.tensor input.
DenseToCSRSparseMatrix	Converts a dense tensor to a (possibly batched) CSRSparseMatrix.
DestroyResourceOp	Deletes the resource specified by the handle.
DestroyTemporaryVariable<T>	Destroys the temporary variable and returns its final value.
DeviceIndex	Return the index of device the op runs.
DirectedInterleaveDataset	A substitute for `InterleaveDataset` on a fixed list of `N` datasets.
DisableCopyOnRead	Turns off the copy-on-read mode.
DistributedSave
DrawBoundingBoxesV2<T extends Number>	Draw bounding boxes on a batch of images.
DummyIterationCounter
DummyMemoryCache
DummySeedGenerator
DynamicEnqueueTPUEmbeddingArbitraryTensorBatch	Eases the porting of code that uses tf.nn.embedding_lookup_sparse().
DynamicEnqueueTPUEmbeddingRaggedTensorBatch
DynamicPartition<T>	Partitions `data` into `num_partitions` tensors using indices from `partitions`.
DynamicStitch<T>	Interleave the values from the `data` tensors into a single tensor.
EditDistance	Computes the (possibly normalized) Levenshtein Edit Distance.
Eig<U>	Computes the eigen decomposition of one or more square matrices.
Einsum<T>	Tensor contraction according to Einstein summation convention.
Empty<T>	Creates a tensor with the given shape.
EmptyTensorList	Creates and returns an empty tensor list.
EmptyTensorMap	Creates and returns an empty tensor map.
EncodeProto	The op serializes protobuf messages provided in the input tensors.
EnqueueTPUEmbeddingArbitraryTensorBatch	Eases the porting of code that uses tf.nn.embedding_lookup_sparse().
EnqueueTPUEmbeddingBatch	An op that enqueues a list of input batch tensors to TPUEmbedding.
EnqueueTPUEmbeddingIntegerBatch	An op that enqueues a list of input batch tensors to TPUEmbedding.
EnqueueTPUEmbeddingRaggedTensorBatch	Eases the porting of code that uses tf.nn.embedding_lookup().
EnqueueTPUEmbeddingSparseBatch	An op that enqueues TPUEmbedding input indices from a SparseTensor.
EnqueueTPUEmbeddingSparseTensorBatch	Eases the porting of code that uses tf.nn.embedding_lookup_sparse().
EnsureShape<T>	Ensures that the tensor's shape matches the expected shape.
Enter<T>	Creates or finds a child frame, and makes `data` available to the child frame.
Erfinv<T extends Number>
EuclideanNorm<T>	Computes the euclidean norm of elements across dimensions of a tensor.
ExecuteTPUEmbeddingPartitioner	An op that executes the TPUEmbedding partitioner on the central configuration device and computes the HBM size (in bytes) required for TPUEmbedding operation.
Exit<T>	Exits the current frame to its parent frame.
ExpandDims<T>	Inserts a dimension of 1 into a tensor's shape.
ExperimentalAutoShardDataset	Creates a dataset that shards the input dataset.
ExperimentalBytesProducedStatsDataset	Records the bytes size of each element of `input_dataset` in a StatsAggregator.
ExperimentalChooseFastestDataset
ExperimentalDatasetCardinality	Returns the cardinality of `input_dataset`.
ExperimentalDatasetToTFRecord	Writes the given dataset to the given file using the TFRecord format.
ExperimentalDenseToSparseBatchDataset	Creates a dataset that batches input elements into a SparseTensor.
ExperimentalLatencyStatsDataset	Records the latency of producing `input_dataset` elements in a StatsAggregator.
ExperimentalMatchingFilesDataset
ExperimentalMaxIntraOpParallelismDataset	Creates a dataset that overrides the maximum intra-op parallelism.
ExperimentalParseExampleDataset	Transforms `input_dataset` containing `Example` protos as vectors of DT_STRING into a dataset of `Tensor` or `SparseTensor` objects representing the parsed features.
ExperimentalPrivateThreadPoolDataset	Creates a dataset that uses a custom thread pool to compute `input_dataset`.
ExperimentalRandomDataset	Creates a Dataset that returns pseudorandom numbers.
ExperimentalRebatchDataset	Creates a dataset that changes the batch size.
ExperimentalSetStatsAggregatorDataset
ExperimentalSlidingWindowDataset	Creates a dataset that passes a sliding window over `input_dataset`.
ExperimentalSqlDataset	Creates a dataset that executes a SQL query and emits rows of the result set.
ExperimentalStatsAggregatorHandle	Creates a statistics manager resource.
ExperimentalStatsAggregatorSummary	Produces a summary of any statistics recorded by the given statistics manager.
ExperimentalUnbatchDataset	A dataset that splits the elements of its input into multiple elements.
Expint<T extends Number>
ExtractGlimpseV2	Extracts a glimpse from the input tensor.
ExtractVolumePatches<T extends Number>	Extract `patches` from `input` and put them in the `"depth"` output dimension.
FFTND<T>	ND fast Fourier transform.
FileSystemSetConfiguration	Set configuration of the file system.
Fill<U>	Creates a tensor filled with a scalar value.
FinalizeDataset	Creates a dataset by applying `tf.data.Options` to `input_dataset`.
FinalizeTPUEmbedding	An op that finalizes the TPUEmbedding configuration.
Fingerprint	Generates fingerprint values.
FresnelCos<T extends Number>
FresnelSin<T extends Number>
FusedBatchNormGradV3<T extends Number, U extends Number>	Gradient for batch normalization.
FusedBatchNormV3<T extends Number, U extends Number>	Batch normalization.
GRUBlockCell<T extends Number>	Computes the GRU cell forward propagation for 1 time step.
GRUBlockCellGrad<T extends Number>	Computes the GRU cell back-propagation for 1 time step.
Gather<T>	Gather slices from `params` axis `axis` according to `indices`.
GatherNd<T>	Gather slices from `params` into a Tensor with shape specified by `indices`.
GenerateBoundingBoxProposals	This op produces Region of Interests from given bounding boxes(bbox_deltas) encoded wrt anchors according to eq.2 in arXiv:1506.01497 The op selects top `pre_nms_topn` scoring boxes, decodes them with respect to anchors, applies non-maximal suppression on overlapping boxes with higher than `nms_threshold` intersection-over-union (iou) value, discarding boxes where shorter side is less than `min_size`.
GetElementAtIndex	Gets the element at the specified index in a dataset.
GetOptions	Returns the `tf.data.Options` attached to `input_dataset`.
GetSessionHandle	Store the input tensor in the state of the current session.
GetSessionTensor<T>	Get the value of the tensor specified by its handle.
GuaranteeConst<T>	Gives a guarantee to the TF runtime that the input tensor is a constant.
HashTable	Creates a non-initialized hash table.
HistogramFixedWidth<U extends Number>	Return histogram of values.
IFFTND<T>	ND inverse fast Fourier transform.
IRFFTND<U extends Number>	ND inverse real fast Fourier transform.
Identity<T>	Return a tensor with the same shape and contents as the input tensor or value.
IdentityN	Returns a list of tensors with the same shapes and contents as the input tensors.
IgnoreErrorsDataset	Creates a dataset that contains the elements of `input_dataset` ignoring errors.
ImageProjectiveTransformV2<T extends Number>	Applies the given transform to each of the images.
ImageProjectiveTransformV3<T extends Number>	Applies the given transform to each of the images.
ImmutableConst<T>	Returns immutable tensor from memory region.
InfeedDequeue<T>	A placeholder op for a value that will be fed into the computation.
InfeedDequeueTuple	Fetches multiple values from infeed as an XLA tuple.
InfeedEnqueue	An op which feeds a single Tensor value into the computation.
InfeedEnqueuePrelinearizedBuffer	An op which enqueues prelinearized buffer into TPU infeed.
InfeedEnqueueTuple	Feeds multiple Tensor values into the computation as an XLA tuple.
InitializeTable	Table initializer that takes two tensors for keys and values respectively.
InitializeTableFromDataset
InitializeTableFromTextFile	Initializes a table from a text file.
InplaceAdd<T>	Adds v into specified rows of x.
InplaceSub<T>	Subtracts `v` into specified rows of `x`.
InplaceUpdate<T>	Updates specified rows 'i' with values 'v'.
IsBoostedTreesEnsembleInitialized	Checks whether a tree ensemble has been initialized.
IsBoostedTreesQuantileStreamResourceInitialized	Checks whether a quantile stream has been initialized.
IsTPUEmbeddingInitialized	Whether TPU Embedding is initialized in a distributed TPU system.
IsVariableInitialized	Checks whether a tensor has been initialized.
IsotonicRegression<U extends Number>	Solves a batch of isotonic regression problems.
IteratorGetDevice	Returns the name of the device on which `resource` has been placed.
KMC2ChainInitialization	Returns the index of a data point that should be added to the seed set.
KmeansPlusPlusInitialization	Selects num_to_sample rows of input using the KMeans++ criterion.
KthOrderStatistic	Computes the Kth order statistic of a data set.
LMDBDataset	Creates a dataset that emits the key-value pairs in one or more LMDB files.
LSTMBlockCell<T extends Number>	Computes the LSTM cell forward propagation for 1 time step.
LSTMBlockCellGrad<T extends Number>	Computes the LSTM cell backward propagation for 1 timestep.
LinSpace<T extends Number>	Generates values in an interval.
ListDataset	Creates a dataset that emits each of `tensors` once.
LoadAllTPUEmbeddingParameters	An op that loads optimization parameters into embedding memory.
LoadTPUEmbeddingADAMParameters	Load ADAM embedding parameters.
LoadTPUEmbeddingAdadeltaParameters	Load Adadelta embedding parameters.
LoadTPUEmbeddingAdagradMomentumParameters	Load Adagrad Momentum embedding parameters.
LoadTPUEmbeddingAdagradParameters	Load Adagrad embedding parameters.
LoadTPUEmbeddingCenteredRMSPropParameters	Load centered RMSProp embedding parameters.
LoadTPUEmbeddingFTRLParameters	Load FTRL embedding parameters.
LoadTPUEmbeddingFrequencyEstimatorParameters	Load frequency estimator embedding parameters.
LoadTPUEmbeddingMDLAdagradLightParameters	Load MDL Adagrad Light embedding parameters.
LoadTPUEmbeddingMomentumParameters	Load Momentum embedding parameters.
LoadTPUEmbeddingProximalAdagradParameters	Load proximal Adagrad embedding parameters.
LoadTPUEmbeddingProximalYogiParameters
LoadTPUEmbeddingRMSPropParameters	Load RMSProp embedding parameters.
LoadTPUEmbeddingStochasticGradientDescentParameters	Load SGD embedding parameters.
LookupTableExport<T, U>	Outputs all keys and values in the table.
LookupTableFind<U>	Looks up keys in a table, outputs the corresponding values.
LookupTableImport	Replaces the contents of the table with the specified keys and values.
LookupTableInsert	Updates the table to associates keys with values.
LookupTableRemove	Removes keys and its associated values from a table.
LookupTableSize	Computes the number of elements in the given table.
LoopCond	Forwards the input to the output.
LowerBound<U extends Number>	Applies lower_bound(sorted_search_values, values) along each row.
Lu<T, U extends Number>	Computes the LU decomposition of one or more square matrices.
MakeUnique	Make all elements in the non-Batch dimension unique, but \"close\" to their initial value.
MapClear	Op removes all elements in the underlying container.
MapIncompleteSize	Op returns the number of incomplete elements in the underlying container.
MapPeek	Op peeks at the values at the specified key.
MapSize	Op returns the number of elements in the underlying container.
MapStage	Stage (key, values) in the underlying container which behaves like a hashtable.
MapUnstage	Op removes and returns the values associated with the key from the underlying container.
MapUnstageNoKey	Op removes and returns a random (key, value) from the underlying container.
MatrixDiagPartV2<T>	Returns the batched diagonal part of a batched tensor.
MatrixDiagPartV3<T>	Returns the batched diagonal part of a batched tensor.
MatrixDiagV2<T>	Returns a batched diagonal tensor with given batched diagonal values.
MatrixDiagV3<T>	Returns a batched diagonal tensor with given batched diagonal values.
MatrixSetDiagV2<T>	Returns a batched matrix tensor with new batched diagonal values.
MatrixSetDiagV3<T>	Returns a batched matrix tensor with new batched diagonal values.
Max<T>	Computes the maximum of elements across dimensions of a tensor.
MaxIntraOpParallelismDataset	Creates a dataset that overrides the maximum intra-op parallelism.
Merge<T>	Forwards the value of an available tensor from `inputs` to `output`.
MergeDedupData	An op merges elements of integer and float tensors into deduplication data as XLA tuple.
Min<T>	Computes the minimum of elements across dimensions of a tensor.
MirrorPad<T>	Pads a tensor with mirrored values.
MirrorPadGrad<T>	Gradient op for `MirrorPad` op.
MlirPassthroughOp	Wraps an arbitrary MLIR computation expressed as a module with a main() function.
MulNoNan<T>	Returns x * y element-wise.
MutableDenseHashTable	Creates an empty hash table that uses tensors as the backing store.
MutableHashTable	Creates an empty hash table.
MutableHashTableOfTensors	Creates an empty hash table.
Mutex	Creates a Mutex resource that can be locked by `MutexLock`.
MutexLock	Locks a mutex resource.
NcclAllReduce<T extends Number>	Outputs a tensor containing the reduction across all input tensors.
NcclBroadcast<T extends Number>	Sends `input` to all devices that are connected to the output.
NcclReduce<T extends Number>	Reduces `input` from `num_devices` using `reduction` to a single device.
Ndtri<T extends Number>
NearestNeighbors	Selects the k nearest centers for each point.
NextAfter<T extends Number>	Returns the next representable value of `x1` in the direction of `x2`, element-wise.
NextIteration<T>	Makes its input available to the next iteration.
NoOp	Does nothing.
NonDeterministicInts<U>	Non-deterministically generates some integers.
NonMaxSuppressionV5<T extends Number>	Greedily selects a subset of bounding boxes in descending order of score, pruning away boxes that have high intersection-over-union (IOU) overlap with previously selected boxes.
NonSerializableDataset
OneHot<U>	Returns a one-hot tensor.
OnesLike<T>	Returns a tensor of ones with the same shape and type as x.
OptimizeDatasetV2	Creates a dataset by applying related optimizations to `input_dataset`.
OptionsDataset	Creates a dataset by attaching tf.data.Options to `input_dataset`.
OrderedMapClear	Op removes all elements in the underlying container.
OrderedMapIncompleteSize	Op returns the number of incomplete elements in the underlying container.
OrderedMapPeek	Op peeks at the values at the specified key.
OrderedMapSize	Op returns the number of elements in the underlying container.
OrderedMapStage	Stage (key, values) in the underlying container which behaves like a ordered associative container.
OrderedMapUnstage	Op removes and returns the values associated with the key from the underlying container.
OrderedMapUnstageNoKey	Op removes and returns the (key, value) element with the smallest key from the underlying container.
OutfeedDequeue<T>	Retrieves a single tensor from the computation outfeed.
OutfeedDequeueTuple	Retrieve multiple values from the computation outfeed.
OutfeedDequeueTupleV2	Retrieve multiple values from the computation outfeed.
OutfeedDequeueV2<T>	Retrieves a single tensor from the computation outfeed.
OutfeedEnqueue	Enqueue a Tensor on the computation outfeed.
OutfeedEnqueueTuple	Enqueue multiple Tensor values on the computation outfeed.
Pad<T>	Pads a tensor.
ParallelBatchDataset
ParallelConcat<T>	Concatenates a list of `N` tensors along the first dimension.
ParallelDynamicStitch<T>	Interleave the values from the `data` tensors into a single tensor.
ParseExampleDatasetV2	Transforms `input_dataset` containing `Example` protos as vectors of DT_STRING into a dataset of `Tensor` or `SparseTensor` objects representing the parsed features.
ParseExampleV2	Transforms a vector of tf.Example protos (as strings) into typed tensors.
ParseSequenceExampleV2	Transforms a vector of tf.io.SequenceExample protos (as strings) into typed tensors.
Placeholder<T>	A placeholder op for a value that will be fed into the computation.
PlaceholderWithDefault<T>	A placeholder op that passes through `input` when its output is not fed.
Prelinearize	An op which linearizes one Tensor value to an opaque variant tensor.
PrelinearizeTuple	An op which linearizes multiple Tensor values to an opaque variant tensor.
Print	Prints a string scalar.
PrivateThreadPoolDataset	Creates a dataset that uses a custom thread pool to compute `input_dataset`.
Prod<T>	Computes the product of elements across dimensions of a tensor.
QuantizeAndDequantizeV4<T extends Number>	Quantizes then dequantizes a tensor.
QuantizeAndDequantizeV4Grad<T extends Number>	Returns the gradient of `QuantizeAndDequantizeV4`.
QuantizedConcat<T>	Concatenates quantized tensors along one dimension.
QuantizedConcatV2<T>
QuantizedConv2DAndRelu<V>
QuantizedConv2DAndReluAndRequantize<V>
QuantizedConv2DAndRequantize<V>
QuantizedConv2DPerChannel<V>	Computes QuantizedConv2D per channel.
QuantizedConv2DWithBias<V>
QuantizedConv2DWithBiasAndRelu<V>
QuantizedConv2DWithBiasAndReluAndRequantize<W>
QuantizedConv2DWithBiasAndRequantize<W>
QuantizedConv2DWithBiasSignedSumAndReluAndRequantize<X>
QuantizedConv2DWithBiasSumAndRelu<V>
QuantizedConv2DWithBiasSumAndReluAndRequantize<X>
QuantizedDepthwiseConv2D<V>	Computes quantized depthwise Conv2D.
QuantizedDepthwiseConv2DWithBias<V>	Computes quantized depthwise Conv2D with Bias.
QuantizedDepthwiseConv2DWithBiasAndRelu<V>	Computes quantized depthwise Conv2D with Bias and Relu.
QuantizedDepthwiseConv2DWithBiasAndReluAndRequantize<W>	Computes quantized depthwise Conv2D with Bias, Relu and Requantize.
QuantizedMatMulWithBias<W>	Performs a quantized matrix multiplication of `a` by the matrix `b` with bias add.
QuantizedMatMulWithBiasAndDequantize<W extends Number>
QuantizedMatMulWithBiasAndRelu<V>	Perform a quantized matrix multiplication of `a` by the matrix `b` with bias add and relu fusion.
QuantizedMatMulWithBiasAndReluAndRequantize<W>	Perform a quantized matrix multiplication of `a` by the matrix `b` with bias add and relu and requantize fusion.
QuantizedMatMulWithBiasAndRequantize<W>
QuantizedReshape<T>	Reshapes a quantized tensor as per the Reshape op.
RFFTND<U>	ND fast real Fourier transform.
RaggedBincount<U extends Number>	Counts the number of occurrences of each value in an integer array.
RaggedCountSparseOutput<U extends Number>	Performs sparse-output bin counting for a ragged tensor input.
RaggedCross<T, U extends Number>	Generates a feature cross from a list of tensors, and returns it as a RaggedTensor.
RaggedFillEmptyRows<T>
RaggedFillEmptyRowsGrad<T>
RaggedGather<T extends Number, U>	Gather ragged slices from `params` axis `0` according to `indices`.
RaggedRange<U extends Number, T extends Number>	Returns a `RaggedTensor` containing the specified sequences of numbers.
RaggedTensorFromVariant<U extends Number, T>	Decodes a `variant` Tensor into a `RaggedTensor`.
RaggedTensorToSparse<U>	Converts a `RaggedTensor` into a `SparseTensor` with the same values.
RaggedTensorToTensor<U>	Create a dense tensor from a ragged tensor, possibly altering its shape.
RaggedTensorToVariant	Encodes a `RaggedTensor` into a `variant` Tensor.
RaggedTensorToVariantGradient<U>	Helper used to compute the gradient for `RaggedTensorToVariant`.
RandomDatasetV2	Creates a Dataset that returns pseudorandom numbers.
RandomIndexShuffle<T extends Number>	Outputs the position of `value` in a permutation of [0, ..., max_index].
Range<T extends Number>	Creates a sequence of numbers.
Rank	Returns the rank of a tensor.
ReadVariableOp<T>	Reads the value of a variable.
ReadVariableXlaSplitND<T>	Splits resource variable input tensor across all dimensions.
RebatchDataset	Creates a dataset that changes the batch size.
RebatchDatasetV2	Creates a dataset that changes the batch size.
Recv<T>	Receives the named tensor from send_device on recv_device.
RecvTPUEmbeddingActivations	An op that receives embedding activations on the TPU.
ReduceAll	Computes the "logical and" of elements across dimensions of a tensor.
ReduceAny	Computes the "logical or" of elements across dimensions of a tensor.
ReduceMax<T>	Computes the maximum of elements across dimensions of a tensor.
ReduceMin<T>	Computes the minimum of elements across dimensions of a tensor.
ReduceProd<T>	Computes the product of elements across dimensions of a tensor.
ReduceSum<T>	Computes the sum of elements across dimensions of a tensor.
RefEnter<T>	Creates or finds a child frame, and makes `data` available to the child frame.
RefExit<T>	Exits the current frame to its parent frame.
RefIdentity<T>	Return the same ref tensor as the input ref tensor.
RefMerge<T>	Forwards the value of an available tensor from `inputs` to `output`.
RefNextIteration<T>	Makes its input available to the next iteration.
RefSelect<T>	Forwards the `index`th element of `inputs` to `output`.
RefSwitch<T>	Forwards the ref tensor `data` to the output port determined by `pred`.
RegisterDataset	Registers a dataset with the tf.data service.
RegisterDatasetV2	Registers a dataset with the tf.data service.
Relayout<T>
RelayoutLike<T>
RequantizationRangePerChannel	Computes requantization range per channel.
RequantizePerChannel<U>	Requantizes input with min and max values known per channel.
Reshape<T>	Reshapes a tensor.
ResourceAccumulatorApplyGradient	Applies a gradient to a given accumulator.
ResourceAccumulatorNumAccumulated	Returns the number of gradients aggregated in the given accumulators.
ResourceAccumulatorSetGlobalStep	Updates the accumulator with a new value for global_step.
ResourceAccumulatorTakeGradient<T>	Extracts the average gradient in the given ConditionalAccumulator.
ResourceApplyAdagradV2	Update '*var' according to the adagrad scheme.
ResourceApplyAdamWithAmsgrad	Update '*var' according to the Adam algorithm.
ResourceApplyKerasMomentum	Update '*var' according to the momentum scheme.
ResourceConditionalAccumulator	A conditional accumulator for aggregating gradients.
ResourceCountUpTo<T extends Number>	Increments variable pointed to by 'resource' until it reaches 'limit'.
ResourceGather<U>	Gather slices from the variable pointed to by `resource` according to `indices`.
ResourceGatherNd<U>
ResourceScatterAdd	Adds sparse updates to the variable referenced by `resource`.
ResourceScatterDiv	Divides sparse updates into the variable referenced by `resource`.
ResourceScatterMax	Reduces sparse updates into the variable referenced by `resource` using the `max` operation.
ResourceScatterMin	Reduces sparse updates into the variable referenced by `resource` using the `min` operation.
ResourceScatterMul	Multiplies sparse updates into the variable referenced by `resource`.
ResourceScatterNdAdd	Applies sparse addition to individual values or slices in a Variable.
ResourceScatterNdMax
ResourceScatterNdMin
ResourceScatterNdSub	Applies sparse subtraction to individual values or slices in a Variable.
ResourceScatterNdUpdate	Applies sparse `updates` to individual values or slices within a given variable according to `indices`.
ResourceScatterSub	Subtracts sparse updates from the variable referenced by `resource`.
ResourceScatterUpdate	Assigns sparse updates to the variable referenced by `resource`.
ResourceSparseApplyAdagradV2	Update relevant entries in 'var' and 'accum' according to the adagrad scheme.
ResourceSparseApplyKerasMomentum	Update relevant entries in 'var' and 'accum' according to the momentum scheme.
ResourceStridedSliceAssign	Assign `value` to the sliced l-value reference of `ref`.
RetrieveAllTPUEmbeddingParameters	An op that retrieves optimization parameters from embedding to host memory.
RetrieveTPUEmbeddingADAMParameters	Retrieve ADAM embedding parameters.
RetrieveTPUEmbeddingAdadeltaParameters	Retrieve Adadelta embedding parameters.
RetrieveTPUEmbeddingAdagradMomentumParameters	Retrieve Adagrad Momentum embedding parameters.
RetrieveTPUEmbeddingAdagradParameters	Retrieve Adagrad embedding parameters.
RetrieveTPUEmbeddingCenteredRMSPropParameters	Retrieve centered RMSProp embedding parameters.
RetrieveTPUEmbeddingFTRLParameters	Retrieve FTRL embedding parameters.
RetrieveTPUEmbeddingFrequencyEstimatorParameters	Retrieve frequency estimator embedding parameters.
RetrieveTPUEmbeddingMDLAdagradLightParameters	Retrieve MDL Adagrad Light embedding parameters.
RetrieveTPUEmbeddingMomentumParameters	Retrieve Momentum embedding parameters.
RetrieveTPUEmbeddingProximalAdagradParameters	Retrieve proximal Adagrad embedding parameters.
RetrieveTPUEmbeddingProximalYogiParameters
RetrieveTPUEmbeddingRMSPropParameters	Retrieve RMSProp embedding parameters.
RetrieveTPUEmbeddingStochasticGradientDescentParameters	Retrieve SGD embedding parameters.
Reverse<T>	Reverses specific dimensions of a tensor.
ReverseSequence<T>	Reverses variable length slices.
RewriteDataset
RiscAbs<T extends Number>
RiscAdd<T extends Number>	Returns x + y element-wise.
RiscBinaryArithmetic<T extends Number>
RiscBinaryComparison
RiscBitcast<U>
RiscBroadcast<T>
RiscCast<U>
RiscCeil<T extends Number>
RiscCholesky<T extends Number>
RiscConcat<T>
RiscConv<T extends Number>
RiscCos<T extends Number>
RiscDiv<T extends Number>
RiscDot<T extends Number>
RiscExp<T extends Number>
RiscFft<T>
RiscFloor<T extends Number>
RiscGather<T>
RiscImag<U extends Number>
RiscIsFinite
RiscLog<T extends Number>
RiscLogicalAnd
RiscLogicalNot
RiscLogicalOr
RiscMax<T extends Number>	Returns max(x, y) element-wise.
RiscMin<T extends Number>
RiscMul<T extends Number>
RiscNeg<T extends Number>
RiscPad<T extends Number>
RiscPool<T extends Number>
RiscPow<T extends Number>
RiscRandomUniform
RiscReal<U extends Number>
RiscReduce<T extends Number>
RiscRem<T extends Number>
RiscReshape<T extends Number>
RiscReverse<T extends Number>
RiscScatter<U extends Number>
RiscShape<U extends Number>
RiscSign<T extends Number>
RiscSlice<T extends Number>
RiscSort<T extends Number>
RiscSqueeze<T>
RiscSub<T extends Number>
RiscTranspose<T>
RiscTriangularSolve<T extends Number>
RiscUnary<T extends Number>
RngReadAndSkip	Advance the counter of a counter-based RNG.
RngSkip	Advance the counter of a counter-based RNG.
Roll<T>	Rolls the elements of a tensor along an axis.
SamplingDataset	Creates a dataset that takes a Bernoulli sample of the contents of another dataset.
ScaleAndTranslate
ScaleAndTranslateGrad<T extends Number>
ScatterAdd<T>	Adds sparse updates to a variable reference.
ScatterDiv<T>	Divides a variable reference by sparse updates.
ScatterMax<T extends Number>	Reduces sparse updates into a variable reference using the `max` operation.
ScatterMin<T extends Number>	Reduces sparse updates into a variable reference using the `min` operation.
ScatterMul<T>	Multiplies sparse updates into a variable reference.
ScatterNd<U>	Scatters `updates` into a tensor of shape `shape` according to `indices`.
ScatterNdAdd<T>	Applies sparse addition to individual values or slices in a Variable.
ScatterNdMax<T>	Computes element-wise maximum.
ScatterNdMin<T>	Computes element-wise minimum.
ScatterNdNonAliasingAdd<T>	Applies sparse addition to `input` using individual values or slices from `updates` according to indices `indices`.
ScatterNdSub<T>	Applies sparse subtraction to individual values or slices in a Variable.
ScatterNdUpdate<T>	Applies sparse `updates` to individual values or slices within a given variable according to `indices`.
ScatterSub<T>	Subtracts sparse updates to a variable reference.
ScatterUpdate<T>	Applies sparse updates to a variable reference.
SegmentMaxV2<T extends Number>	Computes the maximum along segments of a tensor.
SegmentMinV2<T extends Number>	Computes the minimum along segments of a tensor.
SegmentProdV2<T>	Computes the product along segments of a tensor.
SegmentSumV2<T>	Computes the sum along segments of a tensor.
SelectV2<T>
Send	Sends the named tensor from send_device to recv_device.
SendTPUEmbeddingGradients	Performs gradient updates of embedding tables.
SetDiff1d<T, U extends Number>	Computes the difference between two lists of numbers or strings.
SetSize	Number of unique elements along last dimension of input `set`.
Shape<U extends Number>	Returns the shape of a tensor.
ShapeN<U extends Number>	Returns shape of tensors.
ShardDataset	Creates a `Dataset` that includes only 1/`num_shards` of this dataset.
ShuffleAndRepeatDatasetV2
ShuffleDatasetV2
ShuffleDatasetV3
ShutdownDistributedTPU	Shuts down a running distributed TPU system.
ShutdownTPUSystem	An op that shuts down the TPU system.
Size<U extends Number>	Returns the size of a tensor.
Skipgram	Parses a text file and creates a batch of examples.
SleepDataset
Slice<T>	Return a slice from 'input'.
SlidingWindowDataset	Creates a dataset that passes a sliding window over `input_dataset`.
Snapshot<T>	Returns a copy of the input tensor.
SnapshotChunkDataset
SnapshotDataset	Creates a dataset that will write to / read from a snapshot.
SnapshotDatasetReader
SnapshotNestedDatasetReader
SobolSample<T extends Number>	Generates points from the Sobol sequence.
SpaceToBatchNd<T>	SpaceToBatch for N-D tensors of type T.
SparseApplyAdagradV2<T>	Update relevant entries in 'var' and 'accum' according to the adagrad scheme.
SparseBincount<U extends Number>	Counts the number of occurrences of each value in an integer array.
SparseCountSparseOutput<U extends Number>	Performs sparse-output bin counting for a sparse tensor input.
SparseCrossHashed	Generates sparse cross from a list of sparse and dense tensors.
SparseCrossV2	Generates sparse cross from a list of sparse and dense tensors.
SparseMatrixAdd	Sparse addition of two CSR matrices, C = alpha * A + beta * B.
SparseMatrixMatMul<T>	Matrix-multiplies a sparse matrix with a dense matrix.
SparseMatrixMul	Element-wise multiplication of a sparse matrix with a dense tensor.
SparseMatrixNNZ	Returns the number of nonzeroes of `sparse_matrix`.
SparseMatrixOrderingAMD	Computes the Approximate Minimum Degree (AMD) ordering of `input`.
SparseMatrixSoftmax	Calculates the softmax of a CSRSparseMatrix.
SparseMatrixSoftmaxGrad	Calculates the gradient of the SparseMatrixSoftmax op.
SparseMatrixSparseCholesky	Computes the sparse Cholesky decomposition of `input`.
SparseMatrixSparseMatMul	Sparse-matrix-multiplies two CSR matrices `a` and `b`.
SparseMatrixTranspose	Transposes the inner (matrix) dimensions of a CSRSparseMatrix.
SparseMatrixZeros	Creates an all-zeros CSRSparseMatrix with shape `dense_shape`.
SparseSegmentMeanGradV2<T extends Number, U extends Number>	Computes gradients for SparseSegmentMean.
SparseSegmentSqrtNGradV2<T extends Number, U extends Number>	Computes gradients for SparseSegmentSqrtN.
SparseSegmentSumGrad<T extends Number>	Computes gradients for SparseSegmentSum.
SparseSegmentSumGradV2<T extends Number, U extends Number>	Computes gradients for SparseSegmentSum.
SparseTensorToCSRSparseMatrix	Converts a SparseTensor to a (possibly batched) CSRSparseMatrix.
Spence<T extends Number>
Split<T>	Splits a tensor into `num_split` tensors along one dimension.
SplitDedupData<T extends Number, U extends Number>	An op splits input deduplication data XLA tuple into integer and floating point tensors.
SplitV<T>	Splits a tensor into `num_split` tensors along one dimension.
Squeeze<T>	Removes dimensions of size 1 from the shape of a tensor.
Stack<T>	Packs a list of `N` rank-`R` tensors into one rank-`(R+1)` tensor.
Stage	Stage values similar to a lightweight Enqueue.
StageClear	Op removes all elements in the underlying container.
StagePeek	Op peeks at the values at the specified index.
StageSize	Op returns the number of elements in the underlying container.
StatefulRandomBinomial<V extends Number>
StatefulStandardNormal<U>	Outputs random values from a normal distribution.
StatefulStandardNormalV2<U>	Outputs random values from a normal distribution.
StatefulTruncatedNormal<U>	Outputs random values from a truncated normal distribution.
StatefulUniform<U>	Outputs random values from a uniform distribution.
StatefulUniformFullInt<U>	Outputs random integers from a uniform distribution.
StatefulUniformInt<U>	Outputs random integers from a uniform distribution.
StatelessParameterizedTruncatedNormal<V extends Number>
StatelessRandomBinomial<W extends Number>	Outputs deterministic pseudorandom random numbers from a binomial distribution.
StatelessRandomGammaV2<V extends Number>	Outputs deterministic pseudorandom random numbers from a gamma distribution.
StatelessRandomGammaV3<U extends Number>	Outputs deterministic pseudorandom random numbers from a gamma distribution.
StatelessRandomGetAlg	Picks the best counter-based RNG algorithm based on device.
StatelessRandomGetKeyCounter	Scrambles seed into key and counter, using the best algorithm based on device.
StatelessRandomGetKeyCounterAlg	Picks the best algorithm based on device, and scrambles seed into key and counter.
StatelessRandomNormalV2<U extends Number>	Outputs deterministic pseudorandom values from a normal distribution.
StatelessRandomPoisson<W extends Number>	Outputs deterministic pseudorandom random numbers from a Poisson distribution.
StatelessRandomUniformFullInt<V extends Number>	Outputs deterministic pseudorandom random integers from a uniform distribution.
StatelessRandomUniformFullIntV2<U extends Number>	Outputs deterministic pseudorandom random integers from a uniform distribution.
StatelessRandomUniformIntV2<U extends Number>	Outputs deterministic pseudorandom random integers from a uniform distribution.
StatelessRandomUniformV2<U extends Number>	Outputs deterministic pseudorandom random values from a uniform distribution.
StatelessSampleDistortedBoundingBox<T extends Number>	Generate a randomly distorted bounding box for an image deterministically.
StatelessShuffle<T>	Randomly and deterministically shuffles a tensor along its first dimension.
StatelessTruncatedNormalV2<U extends Number>	Outputs deterministic pseudorandom values from a truncated normal distribution.
StatsAggregatorHandleV2
StatsAggregatorSetSummaryWriter	Set a summary_writer_interface to record statistics using given stats_aggregator.
StochasticCastToInt<U extends Number>	Stochastically cast a given tensor from floats to ints.
StopGradient<T>	Stops gradient computation.
StridedSlice<T>	Return a strided slice from `input`.
StridedSliceAssign<T>	Assign `value` to the sliced l-value reference of `ref`.
StridedSliceGrad<U>	Returns the gradient of `StridedSlice`.
StringLower	Converts all uppercase characters into their respective lowercase replacements.
StringNGrams<T extends Number>	Creates ngrams from ragged string data.
StringUpper	Converts all lowercase characters into their respective uppercase replacements.
Sum<T>	Computes the sum of elements across dimensions of a tensor.
SwitchCond<T>	Forwards `data` to the output port determined by `pred`.
SyncDevice	Synchronizes the device this op is run on.
TFRecordDatasetV2	Creates a dataset that emits the records from one or more TFRecord files.
TPUCompilationResult	Returns the result of a TPU compilation.
TPUCompileSucceededAssert	Asserts that compilation succeeded.
TPUEmbeddingActivations	An op enabling differentiation of TPU Embeddings.
TPUExecute	Op that loads and executes a TPU program on a TPU device.
TPUExecuteAndUpdateVariables	Op that executes a program with optional in-place variable updates.
TPUOrdinalSelector	A TPU core selector Op.
TPUPartitionedInput<T>	An op that groups a list of partitioned inputs together.
TPUPartitionedInputV2<T>	An op that groups a list of partitioned inputs together.
TPUPartitionedOutput<T>	An op that demultiplexes a tensor to be sharded by XLA to a list of partitioned outputs outside the XLA computation.
TPUPartitionedOutputV2<T>	An op that demultiplexes a tensor to be sharded by XLA to a list of partitioned outputs outside the XLA computation.
TPUReplicateMetadata	Metadata indicating how the TPU computation should be replicated.
TPUReplicatedInput<T>	Connects N inputs to an N-way replicated TPU computation.
TPUReplicatedOutput<T>	Connects N outputs from an N-way replicated TPU computation.
TPUReshardVariables	Op that reshards on-device TPU variables to specified state.
TPURoundRobin	Round-robin load balancing on TPU cores.
TemporaryVariable<T>	Returns a tensor that may be mutated, but only persists within a single step.
TensorArray	An array of Tensors of given size.
TensorArrayClose	Delete the TensorArray from its resource container.
TensorArrayConcat<T>	Concat the elements from the TensorArray into value `value`.
TensorArrayGather<T>	Gather specific elements from the TensorArray into output `value`.
TensorArrayGrad	Creates a TensorArray for storing the gradients of values in the given handle.
TensorArrayGradWithShape	Creates a TensorArray for storing multiple gradients of values in the given handle.
TensorArrayPack<T>
TensorArrayRead<T>	Read an element from the TensorArray into output `value`.
TensorArrayScatter	Scatter the data from the input value into specific TensorArray elements.
TensorArraySize	Get the current size of the TensorArray.
TensorArraySplit	Split the data from the input value into TensorArray elements.
TensorArrayUnpack
TensorArrayWrite	Push an element onto the tensor_array.
TensorListConcat<T>	Concats all tensors in the list along the 0th dimension.
TensorListConcatLists
TensorListConcatV2<U>	Concats all tensors in the list along the 0th dimension.
TensorListElementShape<T extends Number>	The shape of the elements of the given list, as a tensor.
TensorListFromTensor	Creates a TensorList which, when stacked, has the value of `tensor`.
TensorListGather<T>	Creates a Tensor by indexing into the TensorList.
TensorListGetItem<T>
TensorListLength	Returns the number of tensors in the input tensor list.
TensorListPopBack<T>	Returns the last element of the input list as well as a list with all but that element.
TensorListPushBack	Returns a list which has the passed-in `Tensor` as last element and the other elements of the given list in `input_handle`.
TensorListPushBackBatch
TensorListReserve	List of the given size with empty elements.
TensorListResize	Resizes the list.
TensorListScatter	Creates a TensorList by indexing into a Tensor.
TensorListScatterIntoExistingList	Scatters tensor at indices in an input list.
TensorListScatterV2	Creates a TensorList by indexing into a Tensor.
TensorListSetItem
TensorListSplit	Splits a tensor into a list.
TensorListStack<T>	Stacks all tensors in the list.
TensorMapErase	Returns a tensor map with item from given key erased.
TensorMapHasKey	Returns whether the given key exists in the map.
TensorMapInsert	Returns a map that is the 'input_handle' with the given key-value pair inserted.
TensorMapLookup<U>	Returns the value from a given key in a tensor map.
TensorMapSize	Returns the number of tensors in the input tensor map.
TensorMapStackKeys<T>	Returns a Tensor stack of all keys in a tensor map.
TensorScatterAdd<T>	Adds sparse `updates` to an existing tensor according to `indices`.
TensorScatterMax<T>	Apply a sparse update to a tensor taking the element-wise maximum.
TensorScatterMin<T>
TensorScatterSub<T>	Subtracts sparse `updates` from an existing tensor according to `indices`.
TensorScatterUpdate<T>	Scatter `updates` into an existing tensor according to `indices`.
TensorStridedSliceUpdate<T>	Assign `value` to the sliced l-value reference of `input`.
ThreadPoolDataset	Creates a dataset that uses a custom thread pool to compute `input_dataset`.
ThreadPoolHandle	Creates a dataset that uses a custom thread pool to compute `input_dataset`.
Tile<T>	Constructs a tensor by tiling a given tensor.
Timestamp	Provides the time since epoch in seconds.
ToBool	Converts a tensor to a scalar predicate.
TopKUnique	Returns the TopK unique values in the array in sorted order.
TopKWithUnique	Returns the TopK values in the array in sorted order.
TpuHandleToProtoKey	Converts XRT's uid handles to TensorFlow-friendly input format.
TridiagonalMatMul<T>	Calculate product with tridiagonal matrix.
TridiagonalSolve<T>	Solves tridiagonal systems of equations.
Unbatch<T>	Reverses the operation of Batch for a single output Tensor.
UnbatchGrad<T>	Gradient of Unbatch.
UncompressElement	Uncompresses a compressed dataset element.
UnicodeDecode<T extends Number>	Decodes each string in `input` into a sequence of Unicode code points.
UnicodeEncode	Encode a tensor of ints into unicode strings.
UniformDequantize<U extends Number>	Perform dequantization on the quantized Tensor `input`.
UniformQuantize<U>	Perform quantization on Tensor `input`.
UniformQuantizedAdd<T>	Perform quantized add of quantized Tensor `lhs` and quantized Tensor `rhs` to make quantized `output`.
UniformQuantizedClipByValue<T>	Perform clip by value on the quantized Tensor `operand`.
UniformQuantizedConvolution<U>	Perform quantized convolution of quantized Tensor `lhs` and quantized Tensor `rhs`.
UniformQuantizedConvolutionHybrid<V extends Number>	Perform hybrid quantized convolution of float Tensor `lhs` and quantized Tensor `rhs`.
UniformQuantizedDot<U>	Perform quantized dot of quantized Tensor `lhs` and quantized Tensor `rhs` to make quantized `output`.
UniformQuantizedDotHybrid<V extends Number>	Perform hybrid quantized dot of float Te