| 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. |
| ComputeDedupDataSize |
An op computes the size of the deduplication data from embedding core and returns the updated config. |
| 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. |
| ConvertToCooTensor |
|
| 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. |
| GetMinibatchSplitsWithPhysicalReplica |
|
| GetMinibatchesInCsrWithPhysicalReplica |
|
| 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. |
| GlobalIterId |
|
| 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. |
| ListSnapshotChunksDataset |
|
| 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. |
| StoreMinibatchStatisticsInFdo |
|
| 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. |
| TPUAnnotateTensorsWithDynamicShape |
|
| TPUCompilationResult |
Returns the result of a TPU compilation. |
| TPUCompileSucceededAssert |
Asserts that compilation succeeded. |
| TPUCopyWithDynamicShape |
Op that copies host tensor to device with dynamic shape support. |
| 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 Tensor `lhs` and quantized Tensor `rhs`. |
| UniformRequantize<U> |
Given quantized tensor `input`, requantize it with new quantization parameters. |
| Unique<T, V extends Number> |
Finds unique elements along an axis of a tensor. |
| UniqueDataset |
Creates a dataset that contains the unique elements of `input_dataset`. |
| UniqueWithCounts<T, V extends Number> |
Finds unique elements along an axis of a tensor. |
| UnravelIndex<T extends Number> |
Converts an array of flat indices into a tuple of coordinate arrays. |
| UnsortedSegmentJoin |
|
| Unstack<T> |
Unpacks a given dimension of a rank-`R` tensor into `num` rank-`(R-1)` tensors. |
| Unstage |
Op is similar to a lightweight Dequeue. |
| UnwrapDatasetVariant |
|
| UpperBound<U extends Number> |
Applies upper_bound(sorted_search_values, values) along each row. |
| VarHandleOp |
Creates a handle to a Variable resource. |
| VarIsInitializedOp |
Checks whether a resource handle-based variable has been initialized. |
| Variable<T> |
Holds state in the form of a tensor that persists across steps. |
| VariableShape<T extends Number> |
Returns the shape of the variable pointed to by `resource`. |
| Where |
Returns locations of nonzero / true values in a tensor. |
| Where3<T> |
Selects elements from `x` or `y`, depending on `condition`. |
| WindowOp |
|
| WorkerHeartbeat |
Worker heartbeat op. |
| WrapDatasetVariant |
|
| WriteRawProtoSummary |
Writes a serialized proto summary. |
| XlaConcatND<T> |
Concats input tensor across all dimensions. |
| XlaRecvFromHost<T> |
An op to receive a tensor from the host. |
| XlaRecvTPUEmbeddingActivations |
An op that receives embedding activations on the TPU. |
| XlaRecvTPUEmbeddingDeduplicationData |
Receives deduplication data (indices and weights) from the embedding core. |
| XlaSendTPUEmbeddingGradients |
An op that performs gradient updates of embedding tables. |
| XlaSendToHost |
An op to send a tensor to the host. |
| XlaSparseCoreAdagrad |
|
| XlaSparseCoreAdagradMomentum |
|
| XlaSparseCoreAdam |
|
| XlaSparseCoreFtrl |
|
| XlaSparseCoreSgd |
|
| XlaSparseDenseMatmul |
|
| XlaSparseDenseMatmulGradWithAdagradAndCsrInput |
|
| XlaSparseDenseMatmulGradWithAdagradMomentumAndCsrInput |
|
| XlaSparseDenseMatmulGradWithAdamAndCsrInput |
|
| XlaSparseDenseMatmulGradWithFtrlAndCsrInput |
|
| XlaSparseDenseMatmulGradWithSgdAndCsrInput |
|
| XlaSparseDenseMatmulWithCsrInput |
|
| XlaSplitND<T> |
Splits input tensor across all dimensions. |
| Xlog1py<T> |
Returns 0 if x == 0, and x * log1p(y) otherwise, elementwise. |
| ZerosLike<T> |
Returns a tensor of zeros with the same shape and type as x. |
