op

package

v2.16.1+incompatible Latest Latest Go to latest Published: Mar 6, 2024 License: Apache-2.0, BSD-2-Clause Imports: 3 Imported by: 110

Details

Valid go.mod file

The Go module system was introduced in Go 1.11 and is the official dependency management solution for Go.
Redistributable license

Redistributable licenses place minimal restrictions on how software can be used, modified, and redistributed.
Tagged version

Modules with tagged versions give importers more predictable builds.
Stable version

When a project reaches major version v1 it is considered stable.
Learn more about best practices

Repository

github.com/tensorflow/tensorflow

Links

Open Source Insights

Documentation ¶

Overview ¶

Package op defines functions for adding TensorFlow operations to a Graph.

Functions for adding an operation to a graph take a Scope object as the first argument. The Scope object encapsulates a graph and a set of properties (such as a name prefix) for all operations being added to the graph.

WARNING: The API in this package has not been finalized and can change without notice.

Example ¶

// This example creates a Graph that multiplies a constant matrix with
// a matrix to be provided during graph execution (via
// tensorflow.Session).
s := NewScope()
input := Placeholder(s, tf.Float) // Matrix to be provided to Session.Run
output := MatMul(s,
	Const(s, [][]float32{{10}, {20}}), // Constant 2x1 matrix
	input,
	MatMulTransposeB(true))
if s.Err() != nil {
	panic(s.Err())
}
// Shape of the product: The number of rows is fixed by m1, but the
// number of columns will depend on m2, which is unknown.
fmt.Println(output.Shape())

Output:

[2, ?]

Index ¶

func Abort(scope *Scope, optional ...AbortAttr) (o *tf.Operation)
func Abs(scope *Scope, x tf.Output) (y tf.Output)
func AccumulateNV2(scope *Scope, inputs []tf.Output, shape tf.Shape) (sum tf.Output)
func Acos(scope *Scope, x tf.Output) (y tf.Output)
func Acosh(scope *Scope, x tf.Output) (y tf.Output)
func Add(scope *Scope, x tf.Output, y tf.Output) (z tf.Output)
func AddManySparseToTensorsMap(scope *Scope, sparse_indices tf.Output, sparse_values tf.Output, ...) (sparse_handles tf.Output)
func AddN(scope *Scope, inputs []tf.Output) (sum tf.Output)
func AddSparseToTensorsMap(scope *Scope, sparse_indices tf.Output, sparse_values tf.Output, ...) (sparse_handle tf.Output)
func AddV2(scope *Scope, x tf.Output, y tf.Output) (z tf.Output)
func AdjustContrast(scope *Scope, images tf.Output, contrast_factor tf.Output, min_value tf.Output, ...) (output tf.Output)
func AdjustContrastv2(scope *Scope, images tf.Output, contrast_factor tf.Output) (output tf.Output)
func AdjustHue(scope *Scope, images tf.Output, delta tf.Output) (output tf.Output)
func AdjustSaturation(scope *Scope, images tf.Output, scale tf.Output) (output tf.Output)
func All(scope *Scope, input tf.Output, axis tf.Output, optional ...AllAttr) (output tf.Output)
func AllCandidateSampler(scope *Scope, true_classes tf.Output, num_true int64, num_sampled int64, ...) (sampled_candidates tf.Output, true_expected_count tf.Output, ...)
func AllToAll(scope *Scope, input tf.Output, group_assignment tf.Output, ...) (output tf.Output)
func Angle(scope *Scope, input tf.Output, optional ...AngleAttr) (output tf.Output)
func AnonymousHashTable(scope *Scope, key_dtype tf.DataType, value_dtype tf.DataType) (table_handle tf.Output)
func AnonymousIterator(scope *Scope, output_types []tf.DataType, output_shapes []tf.Shape) (handle tf.Output)
func AnonymousIteratorV2(scope *Scope, output_types []tf.DataType, output_shapes []tf.Shape) (handle tf.Output, deleter tf.Output)
func AnonymousIteratorV3(scope *Scope, output_types []tf.DataType, output_shapes []tf.Shape) (handle tf.Output)
func AnonymousMultiDeviceIterator(scope *Scope, devices []string, output_types []tf.DataType, ...) (handle tf.Output, deleter tf.Output)
func AnonymousMultiDeviceIteratorV3(scope *Scope, devices []string, output_types []tf.DataType, ...) (handle tf.Output)
func AnonymousMutableDenseHashTable(scope *Scope, empty_key tf.Output, deleted_key tf.Output, ...) (table_handle tf.Output)
func AnonymousMutableHashTable(scope *Scope, key_dtype tf.DataType, value_dtype tf.DataType) (table_handle tf.Output)
func AnonymousMutableHashTableOfTensors(scope *Scope, key_dtype tf.DataType, value_dtype tf.DataType, ...) (table_handle tf.Output)
func Any(scope *Scope, input tf.Output, axis tf.Output, optional ...AnyAttr) (output tf.Output)
func ApproxTopK(scope *Scope, input tf.Output, k int64, optional ...ApproxTopKAttr) (values tf.Output, indices tf.Output)
func ApproximateEqual(scope *Scope, x tf.Output, y tf.Output, optional ...ApproximateEqualAttr) (z tf.Output)
func ArgMax(scope *Scope, input tf.Output, dimension tf.Output, optional ...ArgMaxAttr) (output tf.Output)
func ArgMin(scope *Scope, input tf.Output, dimension tf.Output, optional ...ArgMinAttr) (output tf.Output)
func AsString(scope *Scope, input tf.Output, optional ...AsStringAttr) (output tf.Output)
func Asin(scope *Scope, x tf.Output) (y tf.Output)
func Asinh(scope *Scope, x tf.Output) (y tf.Output)
func Assert(scope *Scope, condition tf.Output, data []tf.Output, optional ...AssertAttr) (o *tf.Operation)
func AssertNextDataset(scope *Scope, input_dataset tf.Output, transformations tf.Output, ...) (handle tf.Output)
func AssertPrevDataset(scope *Scope, input_dataset tf.Output, transformations tf.Output, ...) (handle tf.Output)
func AssignAddVariableOp(scope *Scope, resource tf.Output, value tf.Output) (o *tf.Operation)
func AssignSubVariableOp(scope *Scope, resource tf.Output, value tf.Output) (o *tf.Operation)
func AssignVariableOp(scope *Scope, resource tf.Output, value tf.Output, ...) (o *tf.Operation)
func AssignVariableXlaConcatND(scope *Scope, resource tf.Output, inputs []tf.Output, num_concats []int64, ...) (o *tf.Operation)
func Atan(scope *Scope, x tf.Output) (y tf.Output)
func Atan2(scope *Scope, y tf.Output, x tf.Output) (z tf.Output)
func Atanh(scope *Scope, x tf.Output) (y tf.Output)
func AudioSpectrogram(scope *Scope, input tf.Output, window_size int64, stride int64, ...) (spectrogram tf.Output)
func AudioSummary(scope *Scope, tag tf.Output, tensor tf.Output, sample_rate float32, ...) (summary tf.Output)
func AudioSummaryV2(scope *Scope, tag tf.Output, tensor tf.Output, sample_rate tf.Output, ...) (summary tf.Output)
func AutoShardDataset(scope *Scope, input_dataset tf.Output, num_workers tf.Output, index tf.Output, ...) (handle tf.Output)
func AvgPool(scope *Scope, value tf.Output, ksize []int64, strides []int64, padding string, ...) (output tf.Output)
func AvgPool3D(scope *Scope, input tf.Output, ksize []int64, strides []int64, padding string, ...) (output tf.Output)
func AvgPool3DGrad(scope *Scope, orig_input_shape tf.Output, grad tf.Output, ksize []int64, ...) (output tf.Output)
func AvgPoolGrad(scope *Scope, orig_input_shape tf.Output, grad tf.Output, ksize []int64, ...) (output tf.Output)
func Batch(scope *Scope, in_tensors []tf.Output, num_batch_threads int64, ...) (batched_tensors []tf.Output, batch_index tf.Output, id tf.Output)
func BatchDataset(scope *Scope, input_dataset tf.Output, batch_size tf.Output, ...) (handle tf.Output)
func BatchDatasetV2(scope *Scope, input_dataset tf.Output, batch_size tf.Output, ...) (handle tf.Output)
func BatchMatMul(scope *Scope, x tf.Output, y tf.Output, optional ...BatchMatMulAttr) (output tf.Output)
func BatchMatMulV2(scope *Scope, x tf.Output, y tf.Output, optional ...BatchMatMulV2Attr) (output tf.Output)
func BatchMatMulV3(scope *Scope, x tf.Output, y tf.Output, Tout tf.DataType, ...) (output tf.Output)
func BatchNormWithGlobalNormalization(scope *Scope, t tf.Output, m tf.Output, v tf.Output, beta tf.Output, ...) (result tf.Output)
func BatchNormWithGlobalNormalizationGrad(scope *Scope, t tf.Output, m tf.Output, v tf.Output, gamma tf.Output, ...) (dx tf.Output, dm tf.Output, dv tf.Output, db tf.Output, dg tf.Output)
func BatchToSpace(scope *Scope, input tf.Output, crops tf.Output, block_size int64) (output tf.Output)
func BatchToSpaceND(scope *Scope, input tf.Output, block_shape tf.Output, crops tf.Output) (output tf.Output)
func Betainc(scope *Scope, a tf.Output, b tf.Output, x tf.Output) (z tf.Output)
func BiasAdd(scope *Scope, value tf.Output, bias tf.Output, optional ...BiasAddAttr) (output tf.Output)
func BiasAddGrad(scope *Scope, out_backprop tf.Output, optional ...BiasAddGradAttr) (output tf.Output)
func BiasAddV1(scope *Scope, value tf.Output, bias tf.Output) (output tf.Output)
func Bincount(scope *Scope, arr tf.Output, size tf.Output, weights tf.Output) (bins tf.Output)
func Bitcast(scope *Scope, input tf.Output, type_ tf.DataType) (output tf.Output)
func BitwiseAnd(scope *Scope, x tf.Output, y tf.Output) (z tf.Output)
func BitwiseOr(scope *Scope, x tf.Output, y tf.Output) (z tf.Output)
func BitwiseXor(scope *Scope, x tf.Output, y tf.Output) (z tf.Output)
func BlockLSTM(scope *Scope, seq_len_max tf.Output, x tf.Output, cs_prev tf.Output, ...) (i tf.Output, cs tf.Output, f tf.Output, o tf.Output, ci tf.Output, ...)
func BlockLSTMGrad(scope *Scope, seq_len_max tf.Output, x tf.Output, cs_prev tf.Output, ...) (x_grad tf.Output, cs_prev_grad tf.Output, h_prev_grad tf.Output, ...)
func BlockLSTMGradV2(scope *Scope, seq_len_max tf.Output, x tf.Output, cs_prev tf.Output, ...) (x_grad tf.Output, cs_prev_grad tf.Output, h_prev_grad tf.Output, ...)
func BlockLSTMV2(scope *Scope, seq_len_max tf.Output, x tf.Output, cs_prev tf.Output, ...) (i tf.Output, cs tf.Output, f tf.Output, o tf.Output, ci tf.Output, ...)
func BoostedTreesAggregateStats(scope *Scope, node_ids tf.Output, gradients tf.Output, hessians tf.Output, ...) (stats_summary tf.Output)
func BoostedTreesBucketize(scope *Scope, float_values []tf.Output, bucket_boundaries []tf.Output) (buckets []tf.Output)
func BoostedTreesCalculateBestFeatureSplit(scope *Scope, node_id_range tf.Output, stats_summary tf.Output, l1 tf.Output, ...) (node_ids tf.Output, gains tf.Output, feature_dimensions tf.Output, ...)
func BoostedTreesCalculateBestFeatureSplitV2(scope *Scope, node_id_range tf.Output, stats_summaries_list []tf.Output, ...) (node_ids tf.Output, gains tf.Output, feature_ids tf.Output, ...)
func BoostedTreesCalculateBestGainsPerFeature(scope *Scope, node_id_range tf.Output, stats_summary_list []tf.Output, ...) (node_ids_list []tf.Output, gains_list []tf.Output, thresholds_list []tf.Output, ...)
func BoostedTreesCenterBias(scope *Scope, tree_ensemble_handle tf.Output, mean_gradients tf.Output, ...) (continue_centering tf.Output)
func BoostedTreesCreateEnsemble(scope *Scope, tree_ensemble_handle tf.Output, stamp_token tf.Output, ...) (o *tf.Operation)
func BoostedTreesCreateQuantileStreamResource(scope *Scope, quantile_stream_resource_handle tf.Output, epsilon tf.Output, ...) (o *tf.Operation)
func BoostedTreesDeserializeEnsemble(scope *Scope, tree_ensemble_handle tf.Output, stamp_token tf.Output, ...) (o *tf.Operation)
func BoostedTreesEnsembleResourceHandleOp(scope *Scope, optional ...BoostedTreesEnsembleResourceHandleOpAttr) (resource tf.Output)
func BoostedTreesExampleDebugOutputs(scope *Scope, tree_ensemble_handle tf.Output, bucketized_features []tf.Output, ...) (examples_debug_outputs_serialized tf.Output)
func BoostedTreesFlushQuantileSummaries(scope *Scope, quantile_stream_resource_handle tf.Output, num_features int64) (summaries []tf.Output)
func BoostedTreesGetEnsembleStates(scope *Scope, tree_ensemble_handle tf.Output) (stamp_token tf.Output, num_trees tf.Output, num_finalized_trees tf.Output, ...)
func BoostedTreesMakeQuantileSummaries(scope *Scope, float_values []tf.Output, example_weights tf.Output, ...) (summaries []tf.Output)
func BoostedTreesMakeStatsSummary(scope *Scope, node_ids tf.Output, gradients tf.Output, hessians tf.Output, ...) (stats_summary tf.Output)
func BoostedTreesPredict(scope *Scope, tree_ensemble_handle tf.Output, bucketized_features []tf.Output, ...) (logits tf.Output)
func BoostedTreesQuantileStreamResourceAddSummaries(scope *Scope, quantile_stream_resource_handle tf.Output, summaries []tf.Output) (o *tf.Operation)
func BoostedTreesQuantileStreamResourceDeserialize(scope *Scope, quantile_stream_resource_handle tf.Output, ...) (o *tf.Operation)
func BoostedTreesQuantileStreamResourceFlush(scope *Scope, quantile_stream_resource_handle tf.Output, num_buckets tf.Output, ...) (o *tf.Operation)
func BoostedTreesQuantileStreamResourceGetBucketBoundaries(scope *Scope, quantile_stream_resource_handle tf.Output, num_features int64) (bucket_boundaries []tf.Output)
func BoostedTreesQuantileStreamResourceHandleOp(scope *Scope, optional ...BoostedTreesQuantileStreamResourceHandleOpAttr) (resource tf.Output)
func BoostedTreesSerializeEnsemble(scope *Scope, tree_ensemble_handle tf.Output) (stamp_token tf.Output, tree_ensemble_serialized tf.Output)
func BoostedTreesSparseAggregateStats(scope *Scope, node_ids tf.Output, gradients tf.Output, hessians tf.Output, ...) (stats_summary_indices tf.Output, stats_summary_values tf.Output, ...)
func BoostedTreesSparseCalculateBestFeatureSplit(scope *Scope, node_id_range tf.Output, stats_summary_indices tf.Output, ...) (node_ids tf.Output, gains tf.Output, feature_dimensions tf.Output, ...)
func BoostedTreesTrainingPredict(scope *Scope, tree_ensemble_handle tf.Output, cached_tree_ids tf.Output, ...) (partial_logits tf.Output, tree_ids tf.Output, node_ids tf.Output)
func BoostedTreesUpdateEnsemble(scope *Scope, tree_ensemble_handle tf.Output, feature_ids tf.Output, ...) (o *tf.Operation)
func BoostedTreesUpdateEnsembleV2(scope *Scope, tree_ensemble_handle tf.Output, feature_ids []tf.Output, ...) (o *tf.Operation)
func BroadcastArgs(scope *Scope, s0 tf.Output, s1 tf.Output) (r0 tf.Output)
func BroadcastGradientArgs(scope *Scope, s0 tf.Output, s1 tf.Output) (r0 tf.Output, r1 tf.Output)
func BroadcastTo(scope *Scope, input tf.Output, shape tf.Output) (output tf.Output)
func Bucketize(scope *Scope, input tf.Output, boundaries []float32) (output tf.Output)
func BytesProducedStatsDataset(scope *Scope, input_dataset tf.Output, tag tf.Output, ...) (handle tf.Output)
func CSRSparseMatrixComponents(scope *Scope, csr_sparse_matrix tf.Output, index tf.Output, type_ tf.DataType) (row_ptrs tf.Output, col_inds tf.Output, values tf.Output)
func CSRSparseMatrixToDense(scope *Scope, sparse_input tf.Output, type_ tf.DataType) (dense_output tf.Output)
func CSRSparseMatrixToSparseTensor(scope *Scope, sparse_matrix tf.Output, type_ tf.DataType) (indices tf.Output, values tf.Output, dense_shape tf.Output)
func CTCBeamSearchDecoder(scope *Scope, inputs tf.Output, sequence_length tf.Output, beam_width int64, ...) (decoded_indices []tf.Output, decoded_values []tf.Output, ...)
func CTCGreedyDecoder(scope *Scope, inputs tf.Output, sequence_length tf.Output, ...) (decoded_indices tf.Output, decoded_values tf.Output, decoded_shape tf.Output, ...)
func CTCLoss(scope *Scope, inputs tf.Output, labels_indices tf.Output, ...) (loss tf.Output, gradient tf.Output)
func CTCLossV2(scope *Scope, inputs tf.Output, labels_indices tf.Output, ...) (loss tf.Output, gradient tf.Output)
func CacheDataset(scope *Scope, input_dataset tf.Output, filename tf.Output, ...) (handle tf.Output)
func Cast(scope *Scope, x tf.Output, DstT tf.DataType, optional ...CastAttr) (y tf.Output)
func Ceil(scope *Scope, x tf.Output) (y tf.Output)
func CheckNumerics(scope *Scope, tensor tf.Output, message string) (output tf.Output)
func CheckNumericsV2(scope *Scope, tensor tf.Output, message string) (output tf.Output)
func Cholesky(scope *Scope, input tf.Output) (output tf.Output)
func CholeskyGrad(scope *Scope, l tf.Output, grad tf.Output) (output tf.Output)
func ClipByValue(scope *Scope, t tf.Output, clip_value_min tf.Output, clip_value_max tf.Output) (output tf.Output)
func CollateTPUEmbeddingMemory(scope *Scope, memory_configs []tf.Output) (merged_memory_config tf.Output)
func CollectiveAllToAllV2(scope *Scope, input tf.Output, group_size tf.Output, group_key tf.Output, ...) (data tf.Output)
func CollectiveAllToAllV3(scope *Scope, input tf.Output, communicator tf.Output, ...) (data tf.Output)
func CollectiveAssignGroupV2(scope *Scope, group_assignment tf.Output, device_index tf.Output, ...) (group_size tf.Output, group_key tf.Output)
func CollectiveBcastRecv(scope *Scope, T tf.DataType, group_size int64, group_key int64, ...) (data tf.Output)
func CollectiveBcastRecvV2(scope *Scope, group_size tf.Output, group_key tf.Output, ...) (data tf.Output)
func CollectiveBcastSend(scope *Scope, input tf.Output, group_size int64, group_key int64, ...) (data tf.Output)
func CollectiveBcastSendV2(scope *Scope, input tf.Output, group_size tf.Output, group_key tf.Output, ...) (data tf.Output)
func CollectiveGather(scope *Scope, input tf.Output, group_size int64, group_key int64, ...) (data tf.Output)
func CollectiveGatherV2(scope *Scope, input tf.Output, group_size tf.Output, group_key tf.Output, ...) (data tf.Output)
func CollectiveInitializeCommunicator(scope *Scope, group_key tf.Output, rank tf.Output, group_size tf.Output, ...) (communicator tf.Output)
func CollectivePermute(scope *Scope, input tf.Output, source_target_pairs tf.Output) (output tf.Output)
func CollectiveReduce(scope *Scope, input tf.Output, group_size int64, group_key int64, ...) (data tf.Output)
func CollectiveReduceScatterV2(scope *Scope, input tf.Output, group_size tf.Output, group_key tf.Output, ...) (data tf.Output)
func CollectiveReduceV2(scope *Scope, input tf.Output, group_size tf.Output, group_key tf.Output, ...) (data tf.Output)
func CollectiveReduceV3(scope *Scope, input tf.Output, communicator tf.Output, ...) (data tf.Output)
func CombinedNonMaxSuppression(scope *Scope, boxes tf.Output, scores tf.Output, ...) (nmsed_boxes tf.Output, nmsed_scores tf.Output, nmsed_classes tf.Output, ...)
func Complex(scope *Scope, real tf.Output, imag tf.Output, optional ...ComplexAttr) (out tf.Output)
func ComplexAbs(scope *Scope, x tf.Output, optional ...ComplexAbsAttr) (y tf.Output)
func CompositeTensorVariantFromComponents(scope *Scope, components []tf.Output, metadata string) (encoded tf.Output)
func CompositeTensorVariantToComponents(scope *Scope, encoded tf.Output, metadata string, Tcomponents []tf.DataType) (components []tf.Output)
func CompressElement(scope *Scope, components []tf.Output) (compressed tf.Output)
func ComputeAccidentalHits(scope *Scope, true_classes tf.Output, sampled_candidates tf.Output, ...) (indices tf.Output, ids tf.Output, weights tf.Output)
func ComputeBatchSize(scope *Scope, input_dataset tf.Output) (batch_size tf.Output)
func ComputeDedupDataSize(scope *Scope, config string) (num_elements tf.Output)
func ComputeDedupDataTupleMask(scope *Scope, config string) (output_shape tf.Output)
func Concat(scope *Scope, concat_dim tf.Output, values []tf.Output) (output tf.Output)
func ConcatOffset(scope *Scope, concat_dim tf.Output, shape []tf.Output) (offset []tf.Output)
func ConcatV2(scope *Scope, values []tf.Output, axis tf.Output) (output tf.Output)
func ConcatenateDataset(scope *Scope, input_dataset tf.Output, another_dataset tf.Output, ...) (handle tf.Output)
func ConfigureAndInitializeGlobalTPU(scope *Scope, optional ...ConfigureAndInitializeGlobalTPUAttr) (output tf.Output)
func ConfigureDistributedTPU(scope *Scope, optional ...ConfigureDistributedTPUAttr) (topology tf.Output)
func ConfigureTPUEmbedding(scope *Scope, config string) (o *tf.Operation)
func ConfigureTPUEmbeddingHost(scope *Scope, common_config tf.Output, memory_config tf.Output, config string) (network_config tf.Output)
func ConfigureTPUEmbeddingMemory(scope *Scope, common_config tf.Output) (memory_config tf.Output)
func Conj(scope *Scope, input tf.Output) (output tf.Output)
func ConjugateTranspose(scope *Scope, x tf.Output, perm tf.Output) (y tf.Output)
func ConnectTPUEmbeddingHosts(scope *Scope, network_configs []tf.Output) (o *tf.Operation)
func Const(scope *Scope, value interface{}) (output tf.Output)
func ConsumeMutexLock(scope *Scope, mutex_lock tf.Output) (o *tf.Operation)
func ControlTrigger(scope *Scope) (o *tf.Operation)
func Conv(scope *Scope, input tf.Output, filter tf.Output, strides []int64, ...) (output tf.Output)
func Conv2D(scope *Scope, input tf.Output, filter tf.Output, strides []int64, ...) (output tf.Output)
func Conv2DBackpropFilter(scope *Scope, input tf.Output, filter_sizes tf.Output, out_backprop tf.Output, ...) (output tf.Output)
func Conv2DBackpropFilterV2(scope *Scope, input tf.Output, filter tf.Output, out_backprop tf.Output, ...) (output tf.Output)
func Conv2DBackpropInput(scope *Scope, input_sizes tf.Output, filter tf.Output, out_backprop tf.Output, ...) (output tf.Output)
func Conv2DBackpropInputV2(scope *Scope, input tf.Output, filter tf.Output, out_backprop tf.Output, ...) (output tf.Output)
func Conv3D(scope *Scope, input tf.Output, filter tf.Output, strides []int64, ...) (output tf.Output)
func Conv3DBackpropFilter(scope *Scope, input tf.Output, filter tf.Output, out_backprop tf.Output, ...) (output tf.Output)
func Conv3DBackpropFilterV2(scope *Scope, input tf.Output, filter_sizes tf.Output, out_backprop tf.Output, ...) (output tf.Output)
func Conv3DBackpropInput(scope *Scope, input tf.Output, filter tf.Output, out_backprop tf.Output, ...) (output tf.Output)
func Conv3DBackpropInputV2(scope *Scope, input_sizes tf.Output, filter tf.Output, out_backprop tf.Output, ...) (output tf.Output)
func Copy(scope *Scope, input tf.Output, optional ...CopyAttr) (output tf.Output)
func CopyHost(scope *Scope, input tf.Output, optional ...CopyHostAttr) (output tf.Output)
func Cos(scope *Scope, x tf.Output) (y tf.Output)
func Cosh(scope *Scope, x tf.Output) (y tf.Output)
func CropAndResize(scope *Scope, image tf.Output, boxes tf.Output, box_ind tf.Output, ...) (crops tf.Output)
func CropAndResizeGradBoxes(scope *Scope, grads tf.Output, image tf.Output, boxes tf.Output, ...) (output tf.Output)
func CropAndResizeGradImage(scope *Scope, grads tf.Output, boxes tf.Output, box_ind tf.Output, ...) (output tf.Output)
func Cross(scope *Scope, a tf.Output, b tf.Output) (product tf.Output)
func CrossReplicaSum(scope *Scope, input tf.Output, group_assignment tf.Output) (output tf.Output)
func CudnnRNN(scope *Scope, input tf.Output, input_h tf.Output, input_c tf.Output, ...) (output tf.Output, output_h tf.Output, output_c tf.Output, ...)
func CudnnRNNBackprop(scope *Scope, input tf.Output, input_h tf.Output, input_c tf.Output, ...) (input_backprop tf.Output, input_h_backprop tf.Output, ...)
func CudnnRNNBackpropV2(scope *Scope, input tf.Output, input_h tf.Output, input_c tf.Output, ...) (input_backprop tf.Output, input_h_backprop tf.Output, ...)
func CudnnRNNBackpropV3(scope *Scope, input tf.Output, input_h tf.Output, input_c tf.Output, ...) (input_backprop tf.Output, input_h_backprop tf.Output, ...)
func CudnnRNNCanonicalToParams(scope *Scope, num_layers tf.Output, num_units tf.Output, input_size tf.Output, ...) (params tf.Output)
func CudnnRNNCanonicalToParamsV2(scope *Scope, num_layers tf.Output, num_units tf.Output, input_size tf.Output, ...) (params tf.Output)
func CudnnRNNParamsSize(scope *Scope, num_layers tf.Output, num_units tf.Output, input_size tf.Output, ...) (params_size tf.Output)
func CudnnRNNParamsToCanonical(scope *Scope, num_layers tf.Output, num_units tf.Output, input_size tf.Output, ...) (weights []tf.Output, biases []tf.Output)
func CudnnRNNParamsToCanonicalV2(scope *Scope, num_layers tf.Output, num_units tf.Output, input_size tf.Output, ...) (weights []tf.Output, biases []tf.Output)
func CudnnRNNV2(scope *Scope, input tf.Output, input_h tf.Output, input_c tf.Output, ...) (output tf.Output, output_h tf.Output, output_c tf.Output, ...)
func CudnnRNNV3(scope *Scope, input tf.Output, input_h tf.Output, input_c tf.Output, ...) (output tf.Output, output_h tf.Output, output_c tf.Output, ...)
func Cumprod(scope *Scope, x tf.Output, axis tf.Output, optional ...CumprodAttr) (out tf.Output)
func Cumsum(scope *Scope, x tf.Output, axis tf.Output, optional ...CumsumAttr) (out tf.Output)
func CumulativeLogsumexp(scope *Scope, x tf.Output, axis tf.Output, optional ...CumulativeLogsumexpAttr) (out tf.Output)
func DTensorSetGlobalTPUArray(scope *Scope, topology tf.Output) (o *tf.Operation)
func DataFormatDimMap(scope *Scope, x tf.Output, optional ...DataFormatDimMapAttr) (y tf.Output)
func DataFormatVecPermute(scope *Scope, x tf.Output, optional ...DataFormatVecPermuteAttr) (y tf.Output)
func DataServiceDataset(scope *Scope, dataset_id tf.Output, processing_mode tf.Output, ...) (handle tf.Output)
func DataServiceDatasetV2(scope *Scope, dataset_id tf.Output, processing_mode tf.Output, ...) (handle tf.Output)
func DatasetCardinality(scope *Scope, input_dataset tf.Output, optional ...DatasetCardinalityAttr) (cardinality tf.Output)
func DatasetFingerprint(scope *Scope, input_dataset tf.Output) (fingerprint tf.Output)
func DatasetFromGraph(scope *Scope, graph_def tf.Output) (handle tf.Output)
func DatasetToGraph(scope *Scope, input_dataset tf.Output, optional ...DatasetToGraphAttr) (graph tf.Output)
func DatasetToGraphV2(scope *Scope, input_dataset tf.Output, optional ...DatasetToGraphV2Attr) (graph tf.Output)
func DatasetToSingleElement(scope *Scope, dataset tf.Output, output_types []tf.DataType, ...) (components []tf.Output)
func DatasetToTFRecord(scope *Scope, input_dataset tf.Output, filename tf.Output, ...) (o *tf.Operation)
func DebugGradientIdentity(scope *Scope, input tf.Output) (output tf.Output)
func DebugIdentity(scope *Scope, input tf.Output, optional ...DebugIdentityAttr) (output tf.Output)
func DebugIdentityV2(scope *Scope, input tf.Output, optional ...DebugIdentityV2Attr) (output tf.Output)
func DebugIdentityV3(scope *Scope, input tf.Output, optional ...DebugIdentityV3Attr) (output tf.Output)
func DebugNanCount(scope *Scope, input tf.Output, optional ...DebugNanCountAttr) (output tf.Output)
func DebugNumericSummary(scope *Scope, input tf.Output, optional ...DebugNumericSummaryAttr) (output tf.Output)
func DebugNumericSummaryV2(scope *Scope, input tf.Output, optional ...DebugNumericSummaryV2Attr) (output tf.Output)
func DecodeAndCropJpeg(scope *Scope, contents tf.Output, crop_window tf.Output, ...) (image tf.Output)
func DecodeBase64(scope *Scope, input tf.Output) (output tf.Output)
func DecodeBmp(scope *Scope, contents tf.Output, optional ...DecodeBmpAttr) (image tf.Output)
func DecodeCSV(scope *Scope, records tf.Output, record_defaults []tf.Output, ...) (output []tf.Output)
func DecodeCompressed(scope *Scope, bytes tf.Output, optional ...DecodeCompressedAttr) (output tf.Output)
func DecodeGif(scope *Scope, contents tf.Output) (image tf.Output)
func DecodeImage(scope *Scope, contents tf.Output, optional ...DecodeImageAttr) (image tf.Output)
func DecodeJSONExample(scope *Scope, json_examples tf.Output) (binary_examples tf.Output)
func DecodeJpeg(scope *Scope, contents tf.Output, optional ...DecodeJpegAttr) (image tf.Output)
func DecodePaddedRaw(scope *Scope, input_bytes tf.Output, fixed_length tf.Output, ...) (output tf.Output)
func DecodePng(scope *Scope, contents tf.Output, optional ...DecodePngAttr) (image tf.Output)
func DecodeProtoV2(scope *Scope, bytes tf.Output, message_type string, field_names []string, ...) (sizes tf.Output, values []tf.Output)
func DecodeRaw(scope *Scope, bytes tf.Output, out_type tf.DataType, optional ...DecodeRawAttr) (output tf.Output)
func DecodeWav(scope *Scope, contents tf.Output, optional ...DecodeWavAttr) (audio tf.Output, sample_rate tf.Output)
func DeepCopy(scope *Scope, x tf.Output) (y tf.Output)
func DeleteIterator(scope *Scope, handle tf.Output, deleter tf.Output) (o *tf.Operation)
func DeleteMultiDeviceIterator(scope *Scope, multi_device_iterator tf.Output, iterators []tf.Output, ...) (o *tf.Operation)
func DeleteSessionTensor(scope *Scope, handle tf.Output) (o *tf.Operation)
func DenseBincount(scope *Scope, input tf.Output, size tf.Output, weights tf.Output, ...) (output tf.Output)
func DenseCountSparseOutput(scope *Scope, values tf.Output, weights tf.Output, binary_output bool, ...) (output_indices tf.Output, output_values tf.Output, ...)
func DenseToCSRSparseMatrix(scope *Scope, dense_input tf.Output, indices tf.Output) (sparse_output tf.Output)
func DenseToDenseSetOperation(scope *Scope, set1 tf.Output, set2 tf.Output, set_operation string, ...) (result_indices tf.Output, result_values tf.Output, result_shape tf.Output)
func DenseToSparseBatchDataset(scope *Scope, input_dataset tf.Output, batch_size tf.Output, ...) (handle tf.Output)
func DenseToSparseSetOperation(scope *Scope, set1 tf.Output, set2_indices tf.Output, set2_values tf.Output, ...) (result_indices tf.Output, result_values tf.Output, result_shape tf.Output)
func DepthToSpace(scope *Scope, input tf.Output, block_size int64, optional ...DepthToSpaceAttr) (output tf.Output)
func DepthwiseConv2dNative(scope *Scope, input tf.Output, filter tf.Output, strides []int64, ...) (output tf.Output)
func DepthwiseConv2dNativeBackpropFilter(scope *Scope, input tf.Output, filter_sizes tf.Output, out_backprop tf.Output, ...) (output tf.Output)
func DepthwiseConv2dNativeBackpropInput(scope *Scope, input_sizes tf.Output, filter tf.Output, out_backprop tf.Output, ...) (output tf.Output)
func Dequantize(scope *Scope, input tf.Output, min_range tf.Output, max_range tf.Output, ...) (output tf.Output)
func DeserializeIterator(scope *Scope, resource_handle tf.Output, serialized tf.Output) (o *tf.Operation)
func DeserializeManySparse(scope *Scope, serialized_sparse tf.Output, dtype tf.DataType) (sparse_indices tf.Output, sparse_values tf.Output, sparse_shape tf.Output)
func DeserializeSparse(scope *Scope, serialized_sparse tf.Output, dtype tf.DataType) (sparse_indices tf.Output, sparse_values tf.Output, sparse_shape tf.Output)
func DestroyResourceOp(scope *Scope, resource tf.Output, optional ...DestroyResourceOpAttr) (o *tf.Operation)
func DeviceIndex(scope *Scope, device_names []string) (index tf.Output)
func Diag(scope *Scope, diagonal tf.Output) (output tf.Output)
func DiagPart(scope *Scope, input tf.Output) (diagonal tf.Output)
func Digamma(scope *Scope, x tf.Output) (y tf.Output)
func Dilation2D(scope *Scope, input tf.Output, filter tf.Output, strides []int64, ...) (output tf.Output)
func Dilation2DBackpropFilter(scope *Scope, input tf.Output, filter tf.Output, out_backprop tf.Output, ...) (filter_backprop tf.Output)
func Dilation2DBackpropInput(scope *Scope, input tf.Output, filter tf.Output, out_backprop tf.Output, ...) (in_backprop tf.Output)
func DirectedInterleaveDataset(scope *Scope, selector_input_dataset tf.Output, ...) (handle tf.Output)
func DisableCopyOnRead(scope *Scope, resource tf.Output) (o *tf.Operation)
func Div(scope *Scope, x tf.Output, y tf.Output) (z tf.Output)
func DivNoNan(scope *Scope, x tf.Output, y tf.Output) (z tf.Output)
func DrawBoundingBoxes(scope *Scope, images tf.Output, boxes tf.Output) (output tf.Output)
func DrawBoundingBoxesV2(scope *Scope, images tf.Output, boxes tf.Output, colors tf.Output) (output tf.Output)
func DynamicEnqueueTPUEmbeddingArbitraryTensorBatch(scope *Scope, sample_indices_or_row_splits []tf.Output, ...) (o *tf.Operation)
func DynamicPartition(scope *Scope, data tf.Output, partitions tf.Output, num_partitions int64) (outputs []tf.Output)
func DynamicStitch(scope *Scope, indices []tf.Output, data []tf.Output) (merged tf.Output)
func EagerPyFunc(scope *Scope, input []tf.Output, token string, Tout []tf.DataType, ...) (output []tf.Output)
func EditDistance(scope *Scope, hypothesis_indices tf.Output, hypothesis_values tf.Output, ...) (output tf.Output)
func Eig(scope *Scope, input tf.Output, Tout tf.DataType, optional ...EigAttr) (e tf.Output, v tf.Output)
func Einsum(scope *Scope, inputs []tf.Output, equation string) (output tf.Output)
func Elu(scope *Scope, features tf.Output) (activations tf.Output)
func EluGrad(scope *Scope, gradients tf.Output, outputs tf.Output) (backprops tf.Output)
func Empty(scope *Scope, shape tf.Output, dtype tf.DataType, optional ...EmptyAttr) (output tf.Output)
func EmptyTensorList(scope *Scope, element_shape tf.Output, max_num_elements tf.Output, ...) (handle tf.Output)
func EmptyTensorMap(scope *Scope) (handle tf.Output)
func EncodeBase64(scope *Scope, input tf.Output, optional ...EncodeBase64Attr) (output tf.Output)
func EncodeJpeg(scope *Scope, image tf.Output, optional ...EncodeJpegAttr) (contents tf.Output)
func EncodeJpegVariableQuality(scope *Scope, images tf.Output, quality tf.Output) (contents tf.Output)
func EncodePng(scope *Scope, image tf.Output, optional ...EncodePngAttr) (contents tf.Output)
func EncodeProto(scope *Scope, sizes tf.Output, values []tf.Output, field_names []string, ...) (bytes tf.Output)
func EncodeWav(scope *Scope, audio tf.Output, sample_rate tf.Output) (contents tf.Output)
func EnqueueTPUEmbeddingArbitraryTensorBatch(scope *Scope, sample_indices_or_row_splits []tf.Output, ...) (o *tf.Operation)
func EnqueueTPUEmbeddingBatch(scope *Scope, batch []tf.Output, mode_override tf.Output, ...) (o *tf.Operation)
func EnqueueTPUEmbeddingIntegerBatch(scope *Scope, batch []tf.Output, mode_override tf.Output, ...) (o *tf.Operation)
func EnqueueTPUEmbeddingRaggedTensorBatch(scope *Scope, sample_splits []tf.Output, embedding_indices []tf.Output, ...) (o *tf.Operation)
func EnqueueTPUEmbeddingSparseBatch(scope *Scope, sample_indices []tf.Output, embedding_indices []tf.Output, ...) (o *tf.Operation)
func EnqueueTPUEmbeddingSparseTensorBatch(scope *Scope, sample_indices []tf.Output, embedding_indices []tf.Output, ...) (o *tf.Operation)
func EnsureShape(scope *Scope, input tf.Output, shape tf.Shape) (output tf.Output)
func Enter(scope *Scope, data tf.Output, frame_name string, optional ...EnterAttr) (output tf.Output)
func Equal(scope *Scope, x tf.Output, y tf.Output, optional ...EqualAttr) (z tf.Output)
func Erf(scope *Scope, x tf.Output) (y tf.Output)
func Erfc(scope *Scope, x tf.Output) (y tf.Output)
func EuclideanNorm(scope *Scope, input tf.Output, axis tf.Output, optional ...EuclideanNormAttr) (output tf.Output)
func ExecuteTPUEmbeddingPartitioner(scope *Scope, config string) (common_config tf.Output)
func Exit(scope *Scope, data tf.Output) (output tf.Output)
func Exp(scope *Scope, x tf.Output) (y tf.Output)
func ExpandDims(scope *Scope, input tf.Output, axis tf.Output) (output tf.Output)
func ExperimentalAutoShardDataset(scope *Scope, input_dataset tf.Output, num_workers tf.Output, index tf.Output, ...) (handle tf.Output)
func ExperimentalBytesProducedStatsDataset(scope *Scope, input_dataset tf.Output, tag tf.Output, ...) (handle tf.Output)
func ExperimentalDatasetCardinality(scope *Scope, input_dataset tf.Output) (cardinality tf.Output)
func ExperimentalDatasetToTFRecord(scope *Scope, input_dataset tf.Output, filename tf.Output, ...) (o *tf.Operation)
func ExperimentalDenseToSparseBatchDataset(scope *Scope, input_dataset tf.Output, batch_size tf.Output, ...) (handle tf.Output)
func ExperimentalDirectedInterleaveDataset(scope *Scope, selector_input_dataset tf.Output, ...) (handle tf.Output)
func ExperimentalIgnoreErrorsDataset(scope *Scope, input_dataset tf.Output, output_types []tf.DataType, ...) (handle tf.Output)
func ExperimentalIteratorGetDevice(scope *Scope, resource tf.Output) (device tf.Output)
func ExperimentalLatencyStatsDataset(scope *Scope, input_dataset tf.Output, tag tf.Output, ...) (handle tf.Output)
func ExperimentalMaxIntraOpParallelismDataset(scope *Scope, input_dataset tf.Output, max_intra_op_parallelism tf.Output, ...) (handle tf.Output)
func ExperimentalParseExampleDataset(scope *Scope, input_dataset tf.Output, num_parallel_calls tf.Output, ...) (handle tf.Output)
func ExperimentalPrivateThreadPoolDataset(scope *Scope, input_dataset tf.Output, num_threads tf.Output, ...) (handle tf.Output)
func ExperimentalRandomDataset(scope *Scope, seed tf.Output, seed2 tf.Output, output_types []tf.DataType, ...) (handle tf.Output)
func ExperimentalRebatchDataset(scope *Scope, input_dataset tf.Output, num_replicas tf.Output, ...) (handle tf.Output)
func ExperimentalSlidingWindowDataset(scope *Scope, input_dataset tf.Output, window_size tf.Output, ...) (handle tf.Output)
func ExperimentalSqlDataset(scope *Scope, driver_name tf.Output, data_source_name tf.Output, ...) (handle tf.Output)
func ExperimentalStatsAggregatorHandle(scope *Scope, optional ...ExperimentalStatsAggregatorHandleAttr) (handle tf.Output)
func ExperimentalStatsAggregatorSummary(scope *Scope, iterator tf.Output) (summary tf.Output)
func ExperimentalThreadPoolDataset(scope *Scope, input_dataset tf.Output, thread_pool tf.Output, ...) (handle tf.Output)
func ExperimentalThreadPoolHandle(scope *Scope, num_threads int64, display_name string, ...) (handle tf.Output)
func ExperimentalUnbatchDataset(scope *Scope, input_dataset tf.Output, output_types []tf.DataType, ...) (handle tf.Output)
func ExperimentalUniqueDataset(scope *Scope, input_dataset tf.Output, output_types []tf.DataType, ...) (handle tf.Output)
func Expm1(scope *Scope, x tf.Output) (y tf.Output)
func ExtractGlimpse(scope *Scope, input tf.Output, size tf.Output, offsets tf.Output, ...) (glimpse tf.Output)
func ExtractGlimpseV2(scope *Scope, input tf.Output, size tf.Output, offsets tf.Output, ...) (glimpse tf.Output)
func ExtractImagePatches(scope *Scope, images tf.Output, ksizes []int64, strides []int64, rates []int64, ...) (patches tf.Output)
func ExtractJpegShape(scope *Scope, contents tf.Output, optional ...ExtractJpegShapeAttr) (image_shape tf.Output)
func ExtractVolumePatches(scope *Scope, input tf.Output, ksizes []int64, strides []int64, padding string) (patches tf.Output)
func FFT(scope *Scope, input tf.Output) (output tf.Output)
func FFT2D(scope *Scope, input tf.Output) (output tf.Output)
func FFT3D(scope *Scope, input tf.Output) (output tf.Output)
func FFTND(scope *Scope, input tf.Output, fft_length tf.Output, axes tf.Output) (output tf.Output)
func FIFOQueueV2(scope *Scope, component_types []tf.DataType, optional ...FIFOQueueV2Attr) (handle tf.Output)
func Fact(scope *Scope) (fact tf.Output)
func FakeParam(scope *Scope, dtype tf.DataType, shape tf.Shape) (output tf.Output)
func FakeQuantWithMinMaxArgs(scope *Scope, inputs tf.Output, optional ...FakeQuantWithMinMaxArgsAttr) (outputs tf.Output)
func FakeQuantWithMinMaxArgsGradient(scope *Scope, gradients tf.Output, inputs tf.Output, ...) (backprops tf.Output)
func FakeQuantWithMinMaxVars(scope *Scope, inputs tf.Output, min tf.Output, max tf.Output, ...) (outputs tf.Output)
func FakeQuantWithMinMaxVarsGradient(scope *Scope, gradients tf.Output, inputs tf.Output, min tf.Output, ...) (backprops_wrt_input tf.Output, backprop_wrt_min tf.Output, ...)
func FakeQuantWithMinMaxVarsPerChannel(scope *Scope, inputs tf.Output, min tf.Output, max tf.Output, ...) (outputs tf.Output)
func FakeQuantWithMinMaxVarsPerChannelGradient(scope *Scope, gradients tf.Output, inputs tf.Output, min tf.Output, ...) (backprops_wrt_input tf.Output, backprop_wrt_min tf.Output, ...)
func FileSystemSetConfiguration(scope *Scope, scheme tf.Output, key tf.Output, value tf.Output) (o *tf.Operation)
func Fill(scope *Scope, dims tf.Output, value tf.Output) (output tf.Output)
func FilterByLastComponentDataset(scope *Scope, input_dataset tf.Output, output_types []tf.DataType, ...) (output tf.Output)
func FinalizeDataset(scope *Scope, input_dataset tf.Output, output_types []tf.DataType, ...) (handle tf.Output)
func FinalizeTPUEmbedding(scope *Scope, common_config tf.Output, memory_config tf.Output) (o *tf.Operation)
func Fingerprint(scope *Scope, data tf.Output, method tf.Output) (fingerprint tf.Output)
func FixedLengthRecordDataset(scope *Scope, filenames tf.Output, header_bytes tf.Output, ...) (handle tf.Output)
func FixedLengthRecordReaderV2(scope *Scope, record_bytes int64, optional ...FixedLengthRecordReaderV2Attr) (reader_handle tf.Output)
func FixedUnigramCandidateSampler(scope *Scope, true_classes tf.Output, num_true int64, num_sampled int64, ...) (sampled_candidates tf.Output, true_expected_count tf.Output, ...)
func Floor(scope *Scope, x tf.Output) (y tf.Output)
func FloorDiv(scope *Scope, x tf.Output, y tf.Output) (z tf.Output)
func FloorMod(scope *Scope, x tf.Output, y tf.Output) (z tf.Output)
func FractionalAvgPool(scope *Scope, value tf.Output, pooling_ratio []float32, ...) (output tf.Output, row_pooling_sequence tf.Output, ...)
func FractionalAvgPoolGrad(scope *Scope, orig_input_tensor_shape tf.Output, out_backprop tf.Output, ...) (output tf.Output)
func FractionalMaxPool(scope *Scope, value tf.Output, pooling_ratio []float32, ...) (output tf.Output, row_pooling_sequence tf.Output, ...)
func FractionalMaxPoolGrad(scope *Scope, orig_input tf.Output, orig_output tf.Output, ...) (output tf.Output)
func FusedBatchNorm(scope *Scope, x tf.Output, scale tf.Output, offset tf.Output, mean tf.Output, ...) (y tf.Output, batch_mean tf.Output, batch_variance tf.Output, ...)
func FusedBatchNormGrad(scope *Scope, y_backprop tf.Output, x tf.Output, scale tf.Output, ...) (x_backprop tf.Output, scale_backprop tf.Output, offset_backprop tf.Output, ...)
func FusedBatchNormGradV2(scope *Scope, y_backprop tf.Output, x tf.Output, scale tf.Output, ...) (x_backprop tf.Output, scale_backprop tf.Output, offset_backprop tf.Output, ...)
func FusedBatchNormGradV3(scope *Scope, y_backprop tf.Output, x tf.Output, scale tf.Output, ...) (x_backprop tf.Output, scale_backprop tf.Output, offset_backprop tf.Output, ...)
func FusedBatchNormV2(scope *Scope, x tf.Output, scale tf.Output, offset tf.Output, mean tf.Output, ...) (y tf.Output, batch_mean tf.Output, batch_variance tf.Output, ...)
func FusedBatchNormV3(scope *Scope, x tf.Output, scale tf.Output, offset tf.Output, mean tf.Output, ...) (y tf.Output, batch_mean tf.Output, batch_variance tf.Output, ...)
func FusedPadConv2D(scope *Scope, input tf.Output, paddings tf.Output, filter tf.Output, ...) (output tf.Output)
func FusedResizeAndPadConv2D(scope *Scope, input tf.Output, size tf.Output, paddings tf.Output, ...) (output tf.Output)
func GRUBlockCell(scope *Scope, x tf.Output, h_prev tf.Output, w_ru tf.Output, w_c tf.Output, ...) (r tf.Output, u tf.Output, c tf.Output, h tf.Output)
func GRUBlockCellGrad(scope *Scope, x tf.Output, h_prev tf.Output, w_ru tf.Output, w_c tf.Output, ...) (d_x tf.Output, d_h_prev tf.Output, d_c_bar tf.Output, d_r_bar_u_bar tf.Output)
func Gather(scope *Scope, params tf.Output, indices tf.Output, optional ...GatherAttr) (output tf.Output)
func GatherNd(scope *Scope, params tf.Output, indices tf.Output) (output tf.Output)
func GatherV2(scope *Scope, params tf.Output, indices tf.Output, axis tf.Output, ...) (output tf.Output)
func GenerateBoundingBoxProposals(scope *Scope, scores tf.Output, bbox_deltas tf.Output, image_info tf.Output, ...) (rois tf.Output, roi_probabilities tf.Output)
func GenerateVocabRemapping(scope *Scope, new_vocab_file tf.Output, old_vocab_file tf.Output, ...) (remapping tf.Output, num_present tf.Output)
func GetElementAtIndex(scope *Scope, dataset tf.Output, index tf.Output, output_types []tf.DataType, ...) (components []tf.Output)
func GetOptions(scope *Scope, input_dataset tf.Output) (serialized_options tf.Output)
func GetSessionHandle(scope *Scope, value tf.Output) (handle tf.Output)
func GetSessionHandleV2(scope *Scope, value tf.Output) (handle tf.Output)
func GetSessionTensor(scope *Scope, handle tf.Output, dtype tf.DataType) (value tf.Output)
func Gradients(scope *Scope, y []tf.Output, x []tf.Output, dx ...tf.Output) (output []tf.Output)
func Greater(scope *Scope, x tf.Output, y tf.Output) (z tf.Output)
func GreaterEqual(scope *Scope, x tf.Output, y tf.Output) (z tf.Output)
func GuaranteeConst(scope *Scope, input tf.Output) (output tf.Output)
func HSVToRGB(scope *Scope, images tf.Output) (output tf.Output)
func HashTableV2(scope *Scope, key_dtype tf.DataType, value_dtype tf.DataType, ...) (table_handle tf.Output)
func HistogramFixedWidth(scope *Scope, values tf.Output, value_range tf.Output, nbins tf.Output, ...) (out tf.Output)
func HistogramSummary(scope *Scope, tag tf.Output, values tf.Output) (summary tf.Output)
func HostConst(scope *Scope, value tf.Tensor, dtype tf.DataType) (output tf.Output)
func IFFT(scope *Scope, input tf.Output) (output tf.Output)
func IFFT2D(scope *Scope, input tf.Output) (output tf.Output)
func IFFT3D(scope *Scope, input tf.Output) (output tf.Output)
func IFFTND(scope *Scope, input tf.Output, fft_length tf.Output, axes tf.Output) (output tf.Output)
func IRFFT(scope *Scope, input tf.Output, fft_length tf.Output, optional ...IRFFTAttr) (output tf.Output)
func IRFFT2D(scope *Scope, input tf.Output, fft_length tf.Output, optional ...IRFFT2DAttr) (output tf.Output)
func IRFFT3D(scope *Scope, input tf.Output, fft_length tf.Output, optional ...IRFFT3DAttr) (output tf.Output)
func IRFFTND(scope *Scope, input tf.Output, fft_length tf.Output, axes tf.Output, ...) (output tf.Output)
func Identity(scope *Scope, input tf.Output) (output tf.Output)
func IdentityN(scope *Scope, input []tf.Output) (output []tf.Output)
func IdentityReaderV2(scope *Scope, optional ...IdentityReaderV2Attr) (reader_handle tf.Output)
func Igamma(scope *Scope, a tf.Output, x tf.Output) (z tf.Output)
func IgammaGradA(scope *Scope, a tf.Output, x tf.Output) (z tf.Output)
func Igammac(scope *Scope, a tf.Output, x tf.Output) (z tf.Output)
func IgnoreErrorsDataset(scope *Scope, input_dataset tf.Output, output_types []tf.DataType, ...) (handle tf.Output)
func Imag(scope *Scope, input tf.Output, optional ...ImagAttr) (output tf.Output)
func ImageProjectiveTransformV2(scope *Scope, images tf.Output, transforms tf.Output, output_shape tf.Output, ...) (transformed_images tf.Output)
func ImageProjectiveTransformV3(scope *Scope, images tf.Output, transforms tf.Output, output_shape tf.Output, ...) (transformed_images tf.Output)
func ImageSummary(scope *Scope, tag tf.Output, tensor tf.Output, optional ...ImageSummaryAttr) (summary tf.Output)
func ImmutableConst(scope *Scope, dtype tf.DataType, shape tf.Shape, memory_region_name string) (tensor tf.Output)
func InTopK(scope *Scope, predictions tf.Output, targets tf.Output, k int64) (precision tf.Output)
func InTopKV2(scope *Scope, predictions tf.Output, targets tf.Output, k tf.Output) (precision tf.Output)
func InfeedDequeue(scope *Scope, dtype tf.DataType, shape tf.Shape) (output tf.Output)
func InfeedDequeueTuple(scope *Scope, dtypes []tf.DataType, shapes []tf.Shape) (outputs []tf.Output)
func InfeedEnqueue(scope *Scope, input tf.Output, optional ...InfeedEnqueueAttr) (o *tf.Operation)
func InfeedEnqueuePrelinearizedBuffer(scope *Scope, input tf.Output, ...) (o *tf.Operation)
func InfeedEnqueueTuple(scope *Scope, inputs []tf.Output, shapes []tf.Shape, ...) (o *tf.Operation)
func InitializeTableFromTextFileV2(scope *Scope, table_handle tf.Output, filename tf.Output, key_index int64, ...) (o *tf.Operation)
func InitializeTableV2(scope *Scope, table_handle tf.Output, keys tf.Output, values tf.Output) (o *tf.Operation)
func InplaceAdd(scope *Scope, x tf.Output, i tf.Output, v tf.Output) (y tf.Output)
func InplaceSub(scope *Scope, x tf.Output, i tf.Output, v tf.Output) (y tf.Output)
func InplaceUpdate(scope *Scope, x tf.Output, i tf.Output, v tf.Output) (y tf.Output)
func Inv(scope *Scope, x tf.Output) (y tf.Output)
func InvGrad(scope *Scope, y tf.Output, dy tf.Output) (z tf.Output)
func Invert(scope *Scope, x tf.Output) (y tf.Output)
func InvertPermutation(scope *Scope, x tf.Output) (y tf.Output)
func IsBoostedTreesEnsembleInitialized(scope *Scope, tree_ensemble_handle tf.Output) (is_initialized tf.Output)
func IsBoostedTreesQuantileStreamResourceInitialized(scope *Scope, quantile_stream_resource_handle tf.Output) (is_initialized tf.Output)
func IsFinite(scope *Scope, x tf.Output) (y tf.Output)
func IsInf(scope *Scope, x tf.Output) (y tf.Output)
func IsNan(scope *Scope, x tf.Output) (y tf.Output)
func IsTPUEmbeddingInitialized(scope *Scope, optional ...IsTPUEmbeddingInitializedAttr) (is_tpu_embedding_initialized tf.Output)
func IsotonicRegression(scope *Scope, input tf.Output, optional ...IsotonicRegressionAttr) (output tf.Output, segments tf.Output)
func Iterator(scope *Scope, shared_name string, container string, output_types []tf.DataType, ...) (handle tf.Output)
func IteratorFromStringHandle(scope *Scope, string_handle tf.Output, ...) (resource_handle tf.Output)
func IteratorGetDevice(scope *Scope, resource tf.Output) (device tf.Output)
func IteratorGetNext(scope *Scope, iterator tf.Output, output_types []tf.DataType, ...) (components []tf.Output)
func IteratorGetNextAsOptional(scope *Scope, iterator tf.Output, output_types []tf.DataType, ...) (optional tf.Output)
func IteratorGetNextSync(scope *Scope, iterator tf.Output, output_types []tf.DataType, ...) (components []tf.Output)
func IteratorToStringHandle(scope *Scope, resource_handle tf.Output) (string_handle tf.Output)
func KMC2ChainInitialization(scope *Scope, distances tf.Output, seed tf.Output) (index tf.Output)
func KmeansPlusPlusInitialization(scope *Scope, points tf.Output, num_to_sample tf.Output, seed tf.Output, ...) (samples tf.Output)
func KthOrderStatistic(scope *Scope, input tf.Output, k int64) (output tf.Output)
func L2Loss(scope *Scope, t tf.Output) (output tf.Output)
func LMDBDataset(scope *Scope, filenames tf.Output, output_types []tf.DataType, ...) (handle tf.Output)
func LRN(scope *Scope, input tf.Output, optional ...LRNAttr) (output tf.Output)
func LRNGrad(scope *Scope, input_grads tf.Output, input_image tf.Output, ...) (output tf.Output)
func LSTMBlockCell(scope *Scope, x tf.Output, cs_prev tf.Output, h_prev tf.Output, w tf.Output, ...) (i tf.Output, cs tf.Output, f tf.Output, o tf.Output, ci tf.Output, ...)
func LSTMBlockCellGrad(scope *Scope, x tf.Output, cs_prev tf.Output, h_prev tf.Output, w tf.Output, ...) (cs_prev_grad tf.Output, dicfo tf.Output, wci_grad tf.Output, ...)
func LatencyStatsDataset(scope *Scope, input_dataset tf.Output, tag tf.Output, ...) (handle tf.Output)
func LeakyRelu(scope *Scope, features tf.Output, optional ...LeakyReluAttr) (activations tf.Output)
func LeakyReluGrad(scope *Scope, gradients tf.Output, features tf.Output, ...) (backprops tf.Output)
func LearnedUnigramCandidateSampler(scope *Scope, true_classes tf.Output, num_true int64, num_sampled int64, ...) (sampled_candidates tf.Output, true_expected_count tf.Output, ...)
func LeftShift(scope *Scope, x tf.Output, y tf.Output) (z tf.Output)
func Less(scope *Scope, x tf.Output, y tf.Output) (z tf.Output)
func LessEqual(scope *Scope, x tf.Output, y tf.Output) (z tf.Output)
func Lgamma(scope *Scope, x tf.Output) (y tf.Output)
func LinSpace(scope *Scope, start tf.Output, stop tf.Output, num tf.Output) (output tf.Output)
func ListDataset(scope *Scope, tensors []tf.Output, output_types []tf.DataType, ...) (handle tf.Output)
func ListDiff(scope *Scope, x tf.Output, y tf.Output, optional ...ListDiffAttr) (out tf.Output, idx tf.Output)
func LoadAllTPUEmbeddingParameters(scope *Scope, parameters []tf.Output, auxiliary1 []tf.Output, ...) (o *tf.Operation)
func LoadAndRemapMatrix(scope *Scope, ckpt_path tf.Output, old_tensor_name tf.Output, ...) (output_matrix tf.Output)
func LoadTPUEmbeddingADAMParameters(scope *Scope, parameters tf.Output, momenta tf.Output, velocities tf.Output, ...) (o *tf.Operation)
func LoadTPUEmbeddingAdadeltaParameters(scope *Scope, parameters tf.Output, accumulators tf.Output, updates tf.Output, ...) (o *tf.Operation)
func LoadTPUEmbeddingAdagradMomentumParameters(scope *Scope, parameters tf.Output, accumulators tf.Output, momenta tf.Output, ...) (o *tf.Operation)
func LoadTPUEmbeddingAdagradParameters(scope *Scope, parameters tf.Output, accumulators tf.Output, num_shards int64, ...) (o *tf.Operation)
func LoadTPUEmbeddingCenteredRMSPropParameters(scope *Scope, parameters tf.Output, ms tf.Output, mom tf.Output, mg tf.Output, ...) (o *tf.Operation)
func LoadTPUEmbeddingFTRLParameters(scope *Scope, parameters tf.Output, accumulators tf.Output, linears tf.Output, ...) (o *tf.Operation)
func LoadTPUEmbeddingFrequencyEstimatorParameters(scope *Scope, parameters tf.Output, last_hit_step tf.Output, num_shards int64, ...) (o *tf.Operation)
func LoadTPUEmbeddingMDLAdagradLightParameters(scope *Scope, parameters tf.Output, accumulators tf.Output, weights tf.Output, ...) (o *tf.Operation)
func LoadTPUEmbeddingMomentumParameters(scope *Scope, parameters tf.Output, momenta tf.Output, num_shards int64, ...) (o *tf.Operation)
func LoadTPUEmbeddingProximalAdagradParameters(scope *Scope, parameters tf.Output, accumulators tf.Output, num_shards int64, ...) (o *tf.Operation)
func LoadTPUEmbeddingRMSPropParameters(scope *Scope, parameters tf.Output, ms tf.Output, mom tf.Output, ...) (o *tf.Operation)
func LoadTPUEmbeddingStochasticGradientDescentParameters(scope *Scope, parameters tf.Output, num_shards int64, shard_id int64, ...) (o *tf.Operation)
func Log(scope *Scope, x tf.Output) (y tf.Output)
func Log1p(scope *Scope, x tf.Output) (y tf.Output)
func LogMatrixDeterminant(scope *Scope, input tf.Output) (sign tf.Output, log_abs_determinant tf.Output)
func LogSoftmax(scope *Scope, logits tf.Output) (logsoftmax tf.Output)
func LogUniformCandidateSampler(scope *Scope, true_classes tf.Output, num_true int64, num_sampled int64, ...) (sampled_candidates tf.Output, true_expected_count tf.Output, ...)
func LogicalAnd(scope *Scope, x tf.Output, y tf.Output) (z tf.Output)
func LogicalNot(scope *Scope, x tf.Output) (y tf.Output)
func LogicalOr(scope *Scope, x tf.Output, y tf.Output) (z tf.Output)
func LookupTableExportV2(scope *Scope, table_handle tf.Output, Tkeys tf.DataType, Tvalues tf.DataType) (keys tf.Output, values tf.Output)
func LookupTableFindV2(scope *Scope, table_handle tf.Output, keys tf.Output, default_value tf.Output) (values tf.Output)
func LookupTableImportV2(scope *Scope, table_handle tf.Output, keys tf.Output, values tf.Output) (o *tf.Operation)
func LookupTableInsertV2(scope *Scope, table_handle tf.Output, keys tf.Output, values tf.Output) (o *tf.Operation)
func LookupTableRemoveV2(scope *Scope, table_handle tf.Output, keys tf.Output) (o *tf.Operation)
func LookupTableSizeV2(scope *Scope, table_handle tf.Output) (size tf.Output)
func LoopCond(scope *Scope, input tf.Output) (output tf.Output)
func LowerBound(scope *Scope, sorted_inputs tf.Output, values tf.Output, ...) (output tf.Output)
func Lu(scope *Scope, input tf.Output, optional ...LuAttr) (lu tf.Output, p tf.Output)
func MakeIterator(scope *Scope, dataset tf.Output, iterator tf.Output) (o *tf.Operation)
func MakeUnique(scope *Scope, input tf.Output) (output tf.Output)
func MapClear(scope *Scope, dtypes []tf.DataType, optional ...MapClearAttr) (o *tf.Operation)
func MapIncompleteSize(scope *Scope, dtypes []tf.DataType, optional ...MapIncompleteSizeAttr) (size tf.Output)
func MapPeek(scope *Scope, key tf.Output, indices tf.Output, dtypes []tf.DataType, ...) (values []tf.Output)
func MapSize(scope *Scope, dtypes []tf.DataType, optional ...MapSizeAttr) (size tf.Output)
func MapStage(scope *Scope, key tf.Output, indices tf.Output, values []tf.Output, ...) (o *tf.Operation)
func MapUnstage(scope *Scope, key tf.Output, indices tf.Output, dtypes []tf.DataType, ...) (values []tf.Output)
func MapUnstageNoKey(scope *Scope, indices tf.Output, dtypes []tf.DataType, ...) (key tf.Output, values []tf.Output)
func MatMul(scope *Scope, a tf.Output, b tf.Output, optional ...MatMulAttr) (product tf.Output)
func MatchingFiles(scope *Scope, pattern tf.Output) (filenames tf.Output)
func MatrixBandPart(scope *Scope, input tf.Output, num_lower tf.Output, num_upper tf.Output) (band tf.Output)
func MatrixDeterminant(scope *Scope, input tf.Output) (output tf.Output)
func MatrixDiag(scope *Scope, diagonal tf.Output) (output tf.Output)
func MatrixDiagPart(scope *Scope, input tf.Output) (diagonal tf.Output)
func MatrixDiagPartV2(scope *Scope, input tf.Output, k tf.Output, padding_value tf.Output) (diagonal tf.Output)
func MatrixDiagPartV3(scope *Scope, input tf.Output, k tf.Output, padding_value tf.Output, ...) (diagonal tf.Output)
func MatrixDiagV2(scope *Scope, diagonal tf.Output, k tf.Output, num_rows tf.Output, ...) (output tf.Output)
func MatrixDiagV3(scope *Scope, diagonal tf.Output, k tf.Output, num_rows tf.Output, ...) (output tf.Output)
func MatrixExponential(scope *Scope, input tf.Output) (output tf.Output)
func MatrixInverse(scope *Scope, input tf.Output, optional ...MatrixInverseAttr) (output tf.Output)
func MatrixLogarithm(scope *Scope, input tf.Output) (output tf.Output)
func MatrixSetDiag(scope *Scope, input tf.Output, diagonal tf.Output) (output tf.Output)
func MatrixSetDiagV2(scope *Scope, input tf.Output, diagonal tf.Output, k tf.Output) (output tf.Output)
func MatrixSetDiagV3(scope *Scope, input tf.Output, diagonal tf.Output, k tf.Output, ...) (output tf.Output)
func MatrixSolve(scope *Scope, matrix tf.Output, rhs tf.Output, optional ...MatrixSolveAttr) (output tf.Output)
func MatrixSolveLs(scope *Scope, matrix tf.Output, rhs tf.Output, l2_regularizer tf.Output, ...) (output tf.Output)
func MatrixSquareRoot(scope *Scope, input tf.Output) (output tf.Output)
func MatrixTriangularSolve(scope *Scope, matrix tf.Output, rhs tf.Output, ...) (output tf.Output)
func Max(scope *Scope, input tf.Output, axis tf.Output, optional ...MaxAttr) (output tf.Output)
func MaxIntraOpParallelismDataset(scope *Scope, input_dataset tf.Output, max_intra_op_parallelism tf.Output, ...) (handle tf.Output)
func MaxPool(scope *Scope, input tf.Output, ksize []int64, strides []int64, padding string, ...) (output tf.Output)
func MaxPool3D(scope *Scope, input tf.Output, ksize []int64, strides []int64, padding string, ...) (output tf.Output)
func MaxPool3DGrad(scope *Scope, orig_input tf.Output, orig_output tf.Output, grad tf.Output, ...) (output tf.Output)
func MaxPool3DGradGrad(scope *Scope, orig_input tf.Output, orig_output tf.Output, grad tf.Output, ...) (output tf.Output)
func MaxPoolGrad(scope *Scope, orig_input tf.Output, orig_output tf.Output, grad tf.Output, ...) (output tf.Output)
func MaxPoolGradGrad(scope *Scope, orig_input tf.Output, orig_output tf.Output, grad tf.Output, ...) (output tf.Output)
func MaxPoolGradGradV2(scope *Scope, orig_input tf.Output, orig_output tf.Output, grad tf.Output, ...) (output tf.Output)
func MaxPoolGradGradWithArgmax(scope *Scope, input tf.Output, grad tf.Output, argmax tf.Output, ksize []int64, ...) (output tf.Output)
func MaxPoolGradV2(scope *Scope, orig_input tf.Output, orig_output tf.Output, grad tf.Output, ...) (output tf.Output)
func MaxPoolGradWithArgmax(scope *Scope, input tf.Output, grad tf.Output, argmax tf.Output, ksize []int64, ...) (output tf.Output)
func MaxPoolV2(scope *Scope, input tf.Output, ksize tf.Output, strides tf.Output, ...) (output tf.Output)
func MaxPoolWithArgmax(scope *Scope, input tf.Output, ksize []int64, strides []int64, padding string, ...) (output tf.Output, argmax tf.Output)
func Maximum(scope *Scope, x tf.Output, y tf.Output) (z tf.Output)
func Mean(scope *Scope, input tf.Output, axis tf.Output, optional ...MeanAttr) (output tf.Output)
func Merge(scope *Scope, inputs []tf.Output) (output tf.Output, value_index tf.Output)
func MergeDedupData(scope *Scope, integer_tensor tf.Output, float_tensor tf.Output, ...) (output tf.Output)
func MergeSummary(scope *Scope, inputs []tf.Output) (summary tf.Output)
func MergeV2Checkpoints(scope *Scope, checkpoint_prefixes tf.Output, destination_prefix tf.Output, ...) (o *tf.Operation)
func Mfcc(scope *Scope, spectrogram tf.Output, sample_rate tf.Output, ...) (output tf.Output)
func Min(scope *Scope, input tf.Output, axis tf.Output, optional ...MinAttr) (output tf.Output)
func Minimum(scope *Scope, x tf.Output, y tf.Output) (z tf.Output)
func MirrorPad(scope *Scope, input tf.Output, paddings tf.Output, mode string) (output tf.Output)
func MirrorPadGrad(scope *Scope, input tf.Output, paddings tf.Output, mode string) (output tf.Output)
func MlirPassthroughOp(scope *Scope, inputs []tf.Output, mlir_module string, Toutputs []tf.DataType) (outputs []tf.Output)
func Mod(scope *Scope, x tf.Output, y tf.Output) (z tf.Output)
func ModelDataset(scope *Scope, input_dataset tf.Output, output_types []tf.DataType, ...) (handle tf.Output)
func Mul(scope *Scope, x tf.Output, y tf.Output) (z tf.Output)
func MulNoNan(scope *Scope, x tf.Output, y tf.Output) (z tf.Output)
func MultiDeviceIterator(scope *Scope, devices []string, shared_name string, container string, ...) (handle tf.Output)
func MultiDeviceIteratorFromStringHandle(scope *Scope, string_handle tf.Output, ...) (multi_device_iterator tf.Output)
func MultiDeviceIteratorGetNextFromShard(scope *Scope, multi_device_iterator tf.Output, shard_num tf.Output, ...) (components []tf.Output)
func MultiDeviceIteratorInit(scope *Scope, dataset tf.Output, multi_device_iterator tf.Output, ...) (incarnation_id tf.Output)
func MultiDeviceIteratorToStringHandle(scope *Scope, multi_device_iterator tf.Output) (string_handle tf.Output)
func Multinomial(scope *Scope, logits tf.Output, num_samples tf.Output, ...) (output tf.Output)
func MutableDenseHashTableV2(scope *Scope, empty_key tf.Output, deleted_key tf.Output, ...) (table_handle tf.Output)
func MutableHashTableOfTensorsV2(scope *Scope, key_dtype tf.DataType, value_dtype tf.DataType, ...) (table_handle tf.Output)
func MutableHashTableV2(scope *Scope, key_dtype tf.DataType, value_dtype tf.DataType, ...) (table_handle tf.Output)
func MutexLock(scope *Scope, mutex tf.Output) (mutex_lock tf.Output)
func MutexV2(scope *Scope, optional ...MutexV2Attr) (resource tf.Output)
func NcclAllReduce(scope *Scope, input tf.Output, reduction string, num_devices int64, ...) (data tf.Output)
func NcclBroadcast(scope *Scope, input tf.Output, shape tf.Shape) (output tf.Output)
func NcclReduce(scope *Scope, input []tf.Output, reduction string) (data tf.Output)
func NearestNeighbors(scope *Scope, points tf.Output, centers tf.Output, k tf.Output) (nearest_center_indices tf.Output, nearest_center_distances tf.Output)
func Neg(scope *Scope, x tf.Output) (y tf.Output)
func NextAfter(scope *Scope, x1 tf.Output, x2 tf.Output) (output tf.Output)
func NextIteration(scope *Scope, data tf.Output) (output tf.Output)
func NoOp(scope *Scope) (o *tf.Operation)
func NonDeterministicInts(scope *Scope, shape tf.Output, optional ...NonDeterministicIntsAttr) (output tf.Output)
func NonMaxSuppression(scope *Scope, boxes tf.Output, scores tf.Output, max_output_size tf.Output, ...) (selected_indices tf.Output)
func NonMaxSuppressionV2(scope *Scope, boxes tf.Output, scores tf.Output, max_output_size tf.Output, ...) (selected_indices tf.Output)
func NonMaxSuppressionV3(scope *Scope, boxes tf.Output, scores tf.Output, max_output_size tf.Output, ...) (selected_indices tf.Output)
func NonMaxSuppressionV4(scope *Scope, boxes tf.Output, scores tf.Output, max_output_size tf.Output, ...) (selected_indices tf.Output, valid_outputs tf.Output)
func NonMaxSuppressionV5(scope *Scope, boxes tf.Output, scores tf.Output, max_output_size tf.Output, ...) (selected_indices tf.Output, selected_scores tf.Output, valid_outputs tf.Output)
func NonMaxSuppressionWithOverlaps(scope *Scope, overlaps tf.Output, scores tf.Output, max_output_size tf.Output, ...) (selected_indices tf.Output)
func NotEqual(scope *Scope, x tf.Output, y tf.Output, optional ...NotEqualAttr) (z tf.Output)
func NthElement(scope *Scope, input tf.Output, n tf.Output, optional ...NthElementAttr) (values tf.Output)
func OneHot(scope *Scope, indices tf.Output, depth tf.Output, on_value tf.Output, ...) (output tf.Output)
func OnesLike(scope *Scope, x tf.Output) (y tf.Output)
func OptimizeDataset(scope *Scope, input_dataset tf.Output, optimizations tf.Output, ...) (handle tf.Output)
func OptimizeDatasetV2(scope *Scope, input_dataset tf.Output, optimizations_enabled tf.Output, ...) (handle tf.Output)
func OptionalFromValue(scope *Scope, components []tf.Output) (optional tf.Output)
func OptionalGetValue(scope *Scope, optional tf.Output, output_types []tf.DataType, ...) (components []tf.Output)
func OptionalHasValue(scope *Scope, optional tf.Output) (has_value tf.Output)
func OptionalNone(scope *Scope) (optional tf.Output)
func OptionsDataset(scope *Scope, input_dataset tf.Output, serialized_options string, ...) (handle tf.Output)
func OrderedMapClear(scope *Scope, dtypes []tf.DataType, optional ...OrderedMapClearAttr) (o *tf.Operation)
func OrderedMapIncompleteSize(scope *Scope, dtypes []tf.DataType, optional ...OrderedMapIncompleteSizeAttr) (size tf.Output)
func OrderedMapPeek(scope *Scope, key tf.Output, indices tf.Output, dtypes []tf.DataType, ...) (values []tf.Output)
func OrderedMapSize(scope *Scope, dtypes []tf.DataType, optional ...OrderedMapSizeAttr) (size tf.Output)
func OrderedMapStage(scope *Scope, key tf.Output, indices tf.Output, values []tf.Output, ...) (o *tf.Operation)
func OrderedMapUnstage(scope *Scope, key tf.Output, indices tf.Output, dtypes []tf.DataType, ...) (values []tf.Output)
func OrderedMapUnstageNoKey(scope *Scope, indices tf.Output, dtypes []tf.DataType, ...) (key tf.Output, values []tf.Output)
func OutfeedDequeue(scope *Scope, dtype tf.DataType, shape tf.Shape, ...) (output tf.Output)
func OutfeedDequeueTuple(scope *Scope, dtypes []tf.DataType, shapes []tf.Shape, ...) (outputs []tf.Output)
func OutfeedDequeueTupleV2(scope *Scope, device_ordinal tf.Output, dtypes []tf.DataType, ...) (outputs []tf.Output)
func OutfeedDequeueV2(scope *Scope, device_ordinal tf.Output, dtype tf.DataType, shape tf.Shape) (output tf.Output)
func OutfeedEnqueue(scope *Scope, input tf.Output) (o *tf.Operation)
func OutfeedEnqueueTuple(scope *Scope, inputs []tf.Output) (o *tf.Operation)
func Pack(scope *Scope, values []tf.Output, optional ...PackAttr) (output tf.Output)
func Pad(scope *Scope, input tf.Output, paddings tf.Output) (output tf.Output)
func PadV2(scope *Scope, input tf.Output, paddings tf.Output, constant_values tf.Output) (output tf.Output)
func PaddedBatchDataset(scope *Scope, input_dataset tf.Output, batch_size tf.Output, ...) (handle tf.Output)
func PaddedBatchDatasetV2(scope *Scope, input_dataset tf.Output, batch_size tf.Output, ...) (handle tf.Output)
func PaddingFIFOQueueV2(scope *Scope, component_types []tf.DataType, ...) (handle tf.Output)
func ParallelConcat(scope *Scope, values []tf.Output, shape tf.Shape) (output tf.Output)
func ParallelDynamicStitch(scope *Scope, indices []tf.Output, data []tf.Output) (merged tf.Output)
func ParameterizedTruncatedNormal(scope *Scope, shape tf.Output, means tf.Output, stdevs tf.Output, ...) (output tf.Output)
func ParseExample(scope *Scope, serialized tf.Output, names tf.Output, sparse_keys []tf.Output, ...) (sparse_indices []tf.Output, sparse_values []tf.Output, ...)
func ParseExampleDataset(scope *Scope, input_dataset tf.Output, num_parallel_calls tf.Output, ...) (handle tf.Output)
func ParseExampleDatasetV2(scope *Scope, input_dataset tf.Output, num_parallel_calls tf.Output, ...) (handle tf.Output)
func ParseExampleV2(scope *Scope, serialized tf.Output, names tf.Output, sparse_keys tf.Output, ...) (sparse_indices []tf.Output, sparse_values []tf.Output, ...)
func ParseSequenceExample(scope *Scope, serialized tf.Output, debug_name tf.Output, ...) (context_sparse_indices []tf.Output, context_sparse_values []tf.Output, ...)
func ParseSequenceExampleV2(scope *Scope, serialized tf.Output, debug_name tf.Output, ...) (context_sparse_indices []tf.Output, context_sparse_values []tf.Output, ...)
func ParseSingleExample(scope *Scope, serialized tf.Output, dense_defaults []tf.Output, ...) (sparse_indices []tf.Output, sparse_values []tf.Output, ...)
func ParseSingleSequenceExample(scope *Scope, serialized tf.Output, ...) (context_sparse_indices []tf.Output, context_sparse_values []tf.Output, ...)
func ParseTensor(scope *Scope, serialized tf.Output, out_type tf.DataType) (output tf.Output)
func Placeholder(scope *Scope, dtype tf.DataType, optional ...PlaceholderAttr) (output tf.Output)
func PlaceholderV2(scope *Scope, dtype tf.DataType, shape tf.Shape) (output tf.Output)
func PlaceholderWithDefault(scope *Scope, input tf.Output, shape tf.Shape) (output tf.Output)
func Polygamma(scope *Scope, a tf.Output, x tf.Output) (z tf.Output)
func PopulationCount(scope *Scope, x tf.Output) (y tf.Output)
func Pow(scope *Scope, x tf.Output, y tf.Output) (z tf.Output)
func PrefetchDataset(scope *Scope, input_dataset tf.Output, buffer_size tf.Output, ...) (handle tf.Output)
func Prelinearize(scope *Scope, input tf.Output, optional ...PrelinearizeAttr) (output tf.Output)
func PrelinearizeTuple(scope *Scope, inputs []tf.Output, shapes []tf.Shape, ...) (output tf.Output)
func PreventGradient(scope *Scope, input tf.Output, optional ...PreventGradientAttr) (output tf.Output)
func Print(scope *Scope, input tf.Output, data []tf.Output, optional ...PrintAttr) (output tf.Output)
func PrintV2(scope *Scope, input tf.Output, optional ...PrintV2Attr) (o *tf.Operation)
func PriorityQueueV2(scope *Scope, shapes []tf.Shape, optional ...PriorityQueueV2Attr) (handle tf.Output)
func PrivateThreadPoolDataset(scope *Scope, input_dataset tf.Output, num_threads tf.Output, ...) (handle tf.Output)
func Prod(scope *Scope, input tf.Output, axis tf.Output, optional ...ProdAttr) (output tf.Output)
func Qr(scope *Scope, input tf.Output, optional ...QrAttr) (q tf.Output, r tf.Output)
func QuantizeAndDequantize(scope *Scope, input tf.Output, optional ...QuantizeAndDequantizeAttr) (output tf.Output)
func QuantizeAndDequantizeV2(scope *Scope, input tf.Output, input_min tf.Output, input_max tf.Output, ...) (output tf.Output)
func QuantizeAndDequantizeV3(scope *Scope, input tf.Output, input_min tf.Output, input_max tf.Output, ...) (output tf.Output)
func QuantizeAndDequantizeV4(scope *Scope, input tf.Output, input_min tf.Output, input_max tf.Output, ...) (output tf.Output)
func QuantizeAndDequantizeV4Grad(scope *Scope, gradients tf.Output, input tf.Output, input_min tf.Output, ...) (input_backprop tf.Output, input_min_backprop tf.Output, ...)
func QuantizeDownAndShrinkRange(scope *Scope, input tf.Output, input_min tf.Output, input_max tf.Output, ...) (output tf.Output, output_min tf.Output, output_max tf.Output)
func QuantizeV2(scope *Scope, input tf.Output, min_range tf.Output, max_range tf.Output, ...) (output tf.Output, output_min tf.Output, output_max tf.Output)
func QuantizedAdd(scope *Scope, x tf.Output, y tf.Output, min_x tf.Output, max_x tf.Output, ...) (z tf.Output, min_z tf.Output, max_z tf.Output)
func QuantizedAvgPool(scope *Scope, input tf.Output, min_input tf.Output, max_input tf.Output, ...) (output tf.Output, min_output tf.Output, max_output tf.Output)
func QuantizedBatchNormWithGlobalNormalization(scope *Scope, t tf.Output, t_min tf.Output, t_max tf.Output, m tf.Output, ...) (result tf.Output, result_min tf.Output, result_max tf.Output)
func QuantizedBiasAdd(scope *Scope, input tf.Output, bias tf.Output, min_input tf.Output, ...) (output tf.Output, min_out tf.Output, max_out tf.Output)
func QuantizedConcat(scope *Scope, concat_dim tf.Output, values []tf.Output, input_mins []tf.Output, ...) (output tf.Output, output_min tf.Output, output_max tf.Output)
func QuantizedConv2D(scope *Scope, input tf.Output, filter tf.Output, min_input tf.Output, ...) (output tf.Output, min_output tf.Output, max_output tf.Output)
func QuantizedConv2DPerChannel(scope *Scope, input tf.Output, filter tf.Output, min_input tf.Output, ...) (output tf.Output, min_output tf.Output, max_output tf.Output)
func QuantizedDepthwiseConv2D(scope *Scope, input tf.Output, filter tf.Output, min_input tf.Output, ...) (output tf.Output, min_output tf.Output, max_output tf.Output)
func QuantizedDepthwiseConv2DWithBias(scope *Scope, input tf.Output, filter tf.Output, bias tf.Output, ...) (output tf.Output, min_output tf.Output, max_output tf.Output)
func QuantizedDepthwiseConv2DWithBiasAndRelu(scope *Scope, input tf.Output, filter tf.Output, bias tf.Output, ...) (output tf.Output, min_output tf.Output, max_output tf.Output)
func QuantizedDepthwiseConv2DWithBiasAndReluAndRequantize(scope *Scope, input tf.Output, filter tf.Output, bias tf.Output, ...) (output tf.Output, min_output tf.Output, max_output tf.Output)
func QuantizedInstanceNorm(scope *Scope, x tf.Output, x_min tf.Output, x_max tf.Output, ...) (y tf.Output, y_min tf.Output, y_max tf.Output)
func QuantizedMatMul(scope *Scope, a tf.Output, b tf.Output, min_a tf.Output, max_a tf.Output, ...) (out tf.Output, min_out tf.Output, max_out tf.Output)
func QuantizedMatMulWithBias(scope *Scope, a tf.Output, b tf.Output, bias tf.Output, min_a tf.Output, ...) (out tf.Output, min_out tf.Output, max_out tf.Output)
func QuantizedMatMulWithBiasAndRelu(scope *Scope, a tf.Output, b tf.Output, bias tf.Output, min_a tf.Output, ...) (out tf.Output, min_out tf.Output, max_out tf.Output)
func QuantizedMatMulWithBiasAndReluAndRequantize(scope *Scope, a tf.Output, b tf.Output, bias tf.Output, min_a tf.Output, ...) (out tf.Output, min_out tf.Output, max_out tf.Output)
func QuantizedMaxPool(scope *Scope, input tf.Output, min_input tf.Output, max_input tf.Output, ...) (output tf.Output, min_output tf.Output, max_output tf.Output)
func QuantizedMul(scope *Scope, x tf.Output, y tf.Output, min_x tf.Output, max_x tf.Output, ...) (z tf.Output, min_z tf.Output, max_z tf.Output)
func QuantizedRelu(scope *Scope, features tf.Output, min_features tf.Output, ...) (activations tf.Output, min_activations tf.Output, max_activations tf.Output)
func QuantizedRelu6(scope *Scope, features tf.Output, min_features tf.Output, ...) (activations tf.Output, min_activations tf.Output, max_activations tf.Output)
func QuantizedReluX(scope *Scope, features tf.Output, max_value tf.Output, min_features tf.Output, ...) (activations tf.Output, min_activations tf.Output, max_activations tf.Output)
func QuantizedReshape(scope *Scope, tensor tf.Output, shape tf.Output, input_min tf.Output, ...) (output tf.Output, output_min tf.Output, output_max tf.Output)
func QuantizedResizeBilinear(scope *Scope, images tf.Output, size tf.Output, min tf.Output, max tf.Output, ...) (resized_images tf.Output, out_min tf.Output, out_max tf.Output)
func QueueCloseV2(scope *Scope, handle tf.Output, optional ...QueueCloseV2Attr) (o *tf.Operation)
func QueueDequeueManyV2(scope *Scope, handle tf.Output, n tf.Output, component_types []tf.DataType, ...) (components []tf.Output)
func QueueDequeueUpToV2(scope *Scope, handle tf.Output, n tf.Output, component_types []tf.DataType, ...) (components []tf.Output)
func QueueDequeueV2(scope *Scope, handle tf.Output, component_types []tf.DataType, ...) (components []tf.Output)
func QueueEnqueueManyV2(scope *Scope, handle tf.Output, components []tf.Output, ...) (o *tf.Operation)
func QueueEnqueueV2(scope *Scope, handle tf.Output, components []tf.Output, ...) (o *tf.Operation)
func QueueIsClosedV2(scope *Scope, handle tf.Output) (is_closed tf.Output)
func QueueSizeV2(scope *Scope, handle tf.Output) (size tf.Output)
func RFFT(scope *Scope, input tf.Output, fft_length tf.Output, optional ...RFFTAttr) (output tf.Output)
func RFFT2D(scope *Scope, input tf.Output, fft_length tf.Output, optional ...RFFT2DAttr) (output tf.Output)
func RFFT3D(scope *Scope, input tf.Output, fft_length tf.Output, optional ...RFFT3DAttr) (output tf.Output)
func RFFTND(scope *Scope, input tf.Output, fft_length tf.Output, axes tf.Output, ...) (output tf.Output)
func RGBToHSV(scope *Scope, images tf.Output) (output tf.Output)
func RaggedBincount(scope *Scope, splits tf.Output, values tf.Output, size tf.Output, ...) (output tf.Output)
func RaggedCountSparseOutput(scope *Scope, splits tf.Output, values tf.Output, weights tf.Output, ...) (output_indices tf.Output, output_values tf.Output, ...)
func RaggedCross(scope *Scope, ragged_values []tf.Output, ragged_row_splits []tf.Output, ...) (output_values tf.Output, output_row_splits tf.Output)
func RaggedGather(scope *Scope, params_nested_splits []tf.Output, params_dense_values tf.Output, ...) (output_nested_splits []tf.Output, output_dense_values tf.Output)
func RaggedRange(scope *Scope, starts tf.Output, limits tf.Output, deltas tf.Output, ...) (rt_nested_splits tf.Output, rt_dense_values tf.Output)
func RaggedTensorFromVariant(scope *Scope, encoded_ragged tf.Output, input_ragged_rank int64, ...) (output_nested_splits []tf.Output, output_dense_values tf.Output)
func RaggedTensorToSparse(scope *Scope, rt_nested_splits []tf.Output, rt_dense_values tf.Output) (sparse_indices tf.Output, sparse_values tf.Output, ...)
func RaggedTensorToTensor(scope *Scope, shape tf.Output, values tf.Output, default_value tf.Output, ...) (result tf.Output)
func RaggedTensorToVariant(scope *Scope, rt_nested_splits []tf.Output, rt_dense_values tf.Output, ...) (encoded_ragged tf.Output)
func RaggedTensorToVariantGradient(scope *Scope, encoded_ragged_grad tf.Output, row_splits tf.Output, ...) (dense_values_grad tf.Output)
func RandomCrop(scope *Scope, image tf.Output, size tf.Output, optional ...RandomCropAttr) (output tf.Output)
func RandomDataset(scope *Scope, seed tf.Output, seed2 tf.Output, output_types []tf.DataType, ...) (handle tf.Output)
func RandomDatasetV2(scope *Scope, seed tf.Output, seed2 tf.Output, seed_generator tf.Output, ...) (handle tf.Output)
func RandomGamma(scope *Scope, shape tf.Output, alpha tf.Output, optional ...RandomGammaAttr) (output tf.Output)
func RandomGammaGrad(scope *Scope, alpha tf.Output, sample tf.Output) (output tf.Output)
func RandomIndexShuffle(scope *Scope, index tf.Output, seed tf.Output, max_index tf.Output, ...) (output tf.Output)
func RandomPoisson(scope *Scope, shape tf.Output, rate tf.Output, optional ...RandomPoissonAttr) (output tf.Output)
func RandomPoissonV2(scope *Scope, shape tf.Output, rate tf.Output, optional ...RandomPoissonV2Attr) (output tf.Output)
func RandomShuffle(scope *Scope, value tf.Output, optional ...RandomShuffleAttr) (output tf.Output)
func RandomShuffleQueueV2(scope *Scope, component_types []tf.DataType, ...) (handle tf.Output)
func RandomStandardNormal(scope *Scope, shape tf.Output, dtype tf.DataType, ...) (output tf.Output)
func RandomUniform(scope *Scope, shape tf.Output, dtype tf.DataType, ...) (output tf.Output)
func RandomUniformInt(scope *Scope, shape tf.Output, minval tf.Output, maxval tf.Output, ...) (output tf.Output)
func Range(scope *Scope, start tf.Output, limit tf.Output, delta tf.Output) (output tf.Output)
func RangeDataset(scope *Scope, start tf.Output, stop tf.Output, step tf.Output, ...) (handle tf.Output)
func Rank(scope *Scope, input tf.Output) (output tf.Output)
func ReadFile(scope *Scope, filename tf.Output) (contents tf.Output)
func ReadVariableOp(scope *Scope, resource tf.Output, dtype tf.DataType) (value tf.Output)
func ReadVariableXlaSplitND(scope *Scope, resource tf.Output, T tf.DataType, N int64, num_splits []int64, ...) (outputs []tf.Output)
func ReaderNumRecordsProducedV2(scope *Scope, reader_handle tf.Output) (records_produced tf.Output)
func ReaderNumWorkUnitsCompletedV2(scope *Scope, reader_handle tf.Output) (units_completed tf.Output)
func ReaderReadUpToV2(scope *Scope, reader_handle tf.Output, queue_handle tf.Output, ...) (keys tf.Output, values tf.Output)
func ReaderReadV2(scope *Scope, reader_handle tf.Output, queue_handle tf.Output) (key tf.Output, value tf.Output)
func ReaderResetV2(scope *Scope, reader_handle tf.Output) (o *tf.Operation)
func ReaderRestoreStateV2(scope *Scope, reader_handle tf.Output, state tf.Output) (o *tf.Operation)
func ReaderSerializeStateV2(scope *Scope, reader_handle tf.Output) (state tf.Output)
func Real(scope *Scope, input tf.Output, optional ...RealAttr) (output tf.Output)
func RealDiv(scope *Scope, x tf.Output, y tf.Output) (z tf.Output)
func RebatchDataset(scope *Scope, input_dataset tf.Output, num_replicas tf.Output, ...) (handle tf.Output)
func RebatchDatasetV2(scope *Scope, input_dataset tf.Output, batch_sizes tf.Output, ...) (handle tf.Output)
func Reciprocal(scope *Scope, x tf.Output) (y tf.Output)
func ReciprocalGrad(scope *Scope, y tf.Output, dy tf.Output) (z tf.Output)
func RecordInput(scope *Scope, file_pattern string, optional ...RecordInputAttr) (records tf.Output)
func Recv(scope *Scope, tensor_type tf.DataType, tensor_name string, send_device string, ...) (tensor tf.Output)
func RecvTPUEmbeddingActivations(scope *Scope, num_outputs int64, config string) (outputs []tf.Output)
func ReduceJoin(scope *Scope, inputs tf.Output, reduction_indices tf.Output, ...) (output tf.Output)
func RegexFullMatch(scope *Scope, input tf.Output, pattern tf.Output) (output tf.Output)
func RegexReplace(scope *Scope, input tf.Output, pattern tf.Output, rewrite tf.Output, ...) (output tf.Output)
func RegisterDataset(scope *Scope, dataset tf.Output, address tf.Output, protocol tf.Output, ...) (dataset_id tf.Output)
func RegisterDatasetV2(scope *Scope, dataset tf.Output, address tf.Output, protocol tf.Output, ...) (dataset_id tf.Output)
func Relu(scope *Scope, features tf.Output) (activations tf.Output)
func Relu6(scope *Scope, features tf.Output) (activations tf.Output)
func Relu6Grad(scope *Scope, gradients tf.Output, features tf.Output) (backprops tf.Output)
func ReluGrad(scope *Scope, gradients tf.Output, features tf.Output) (backprops tf.Output)
func RepeatDataset(scope *Scope, input_dataset tf.Output, count tf.Output, ...) (handle tf.Output)
func RequantizationRange(scope *Scope, input tf.Output, input_min tf.Output, input_max tf.Output) (output_min tf.Output, output_max tf.Output)
func RequantizationRangePerChannel(scope *Scope, input tf.Output, input_min tf.Output, input_max tf.Output, ...) (output_min tf.Output, output_max tf.Output)
func Requantize(scope *Scope, input tf.Output, input_min tf.Output, input_max tf.Output, ...) (output tf.Output, output_min tf.Output, output_max tf.Output)
func RequantizePerChannel(scope *Scope, input tf.Output, input_min tf.Output, input_max tf.Output, ...) (output tf.Output, output_min tf.Output, output_max tf.Output)
func Reshape(scope *Scope, tensor tf.Output, shape tf.Output) (output tf.Output)
func ResizeArea(scope *Scope, images tf.Output, size tf.Output, optional ...ResizeAreaAttr) (resized_images tf.Output)
func ResizeBicubic(scope *Scope, images tf.Output, size tf.Output, optional ...ResizeBicubicAttr) (resized_images tf.Output)
func ResizeBicubicGrad(scope *Scope, grads tf.Output, original_image tf.Output, ...) (output tf.Output)
func ResizeBilinear(scope *Scope, images tf.Output, size tf.Output, optional ...ResizeBilinearAttr) (resized_images tf.Output)
func ResizeBilinearGrad(scope *Scope, grads tf.Output, original_image tf.Output, ...) (output tf.Output)
func ResizeNearestNeighbor(scope *Scope, images tf.Output, size tf.Output, ...) (resized_images tf.Output)
func ResizeNearestNeighborGrad(scope *Scope, grads tf.Output, size tf.Output, ...) (output tf.Output)
func ResourceAccumulatorApplyGradient(scope *Scope, handle tf.Output, local_step tf.Output, gradient tf.Output) (o *tf.Operation)
func ResourceAccumulatorNumAccumulated(scope *Scope, handle tf.Output) (num_accumulated tf.Output)
func ResourceAccumulatorSetGlobalStep(scope *Scope, handle tf.Output, new_global_step tf.Output) (o *tf.Operation)
func ResourceAccumulatorTakeGradient(scope *Scope, handle tf.Output, num_required tf.Output, dtype tf.DataType) (average tf.Output)
func ResourceApplyAdaMax(scope *Scope, var_ tf.Output, m tf.Output, v tf.Output, beta1_power tf.Output, ...) (o *tf.Operation)
func ResourceApplyAdadelta(scope *Scope, var_ tf.Output, accum tf.Output, accum_update tf.Output, ...) (o *tf.Operation)
func ResourceApplyAdagrad(scope *Scope, var_ tf.Output, accum tf.Output, lr tf.Output, grad tf.Output, ...) (o *tf.Operation)
func ResourceApplyAdagradDA(scope *Scope, var_ tf.Output, gradient_accumulator tf.Output, ...) (o *tf.Operation)
func ResourceApplyAdagradV2(scope *Scope, var_ tf.Output, accum tf.Output, lr tf.Output, epsilon tf.Output, ...) (o *tf.Operation)
func ResourceApplyAdam(scope *Scope, var_ tf.Output, m tf.Output, v tf.Output, beta1_power tf.Output, ...) (o *tf.Operation)
func ResourceApplyAdamWithAmsgrad(scope *Scope, var_ tf.Output, m tf.Output, v tf.Output, vhat tf.Output, ...) (o *tf.Operation)
func ResourceApplyAddSign(scope *Scope, var_ tf.Output, m tf.Output, lr tf.Output, alpha tf.Output, ...) (o *tf.Operation)
func ResourceApplyCenteredRMSProp(scope *Scope, var_ tf.Output, mg tf.Output, ms tf.Output, mom tf.Output, ...) (o *tf.Operation)
func ResourceApplyFtrl(scope *Scope, var_ tf.Output, accum tf.Output, linear tf.Output, ...) (o *tf.Operation)
func ResourceApplyFtrlV2(scope *Scope, var_ tf.Output, accum tf.Output, linear tf.Output, ...) (o *tf.Operation)
func ResourceApplyGradientDescent(scope *Scope, var_ tf.Output, alpha tf.Output, delta tf.Output, ...) (o *tf.Operation)
func ResourceApplyKerasMomentum(scope *Scope, var_ tf.Output, accum tf.Output, lr tf.Output, grad tf.Output, ...) (o *tf.Operation)
func ResourceApplyMomentum(scope *Scope, var_ tf.Output, accum tf.Output, lr tf.Output, grad tf.Output, ...) (o *tf.Operation)
func ResourceApplyPowerSign(scope *Scope, var_ tf.Output, m tf.Output, lr tf.Output, logbase tf.Output, ...) (o *tf.Operation)
func ResourceApplyProximalAdagrad(scope *Scope, var_ tf.Output, accum tf.Output, lr tf.Output, l1 tf.Output, ...) (o *tf.Operation)
func ResourceApplyProximalGradientDescent(scope *Scope, var_ tf.Output, alpha tf.Output, l1 tf.Output, l2 tf.Output, ...) (o *tf.Operation)
func ResourceApplyRMSProp(scope *Scope, var_ tf.Output, ms tf.Output, mom tf.Output, lr tf.Output, ...) (o *tf.Operation)
func ResourceConditionalAccumulator(scope *Scope, dtype tf.DataType, shape tf.Shape, ...) (handle tf.Output)
func ResourceCountUpTo(scope *Scope, resource tf.Output, limit int64, T tf.DataType) (output tf.Output)
func ResourceGather(scope *Scope, resource tf.Output, indices tf.Output, dtype tf.DataType, ...) (output tf.Output)
func ResourceScatterAdd(scope *Scope, resource tf.Output, indices tf.Output, updates tf.Output) (o *tf.Operation)
func ResourceScatterDiv(scope *Scope, resource tf.Output, indices tf.Output, updates tf.Output) (o *tf.Operation)
func ResourceScatterMax(scope *Scope, resource tf.Output, indices tf.Output, updates tf.Output) (o *tf.Operation)
func ResourceScatterMin(scope *Scope, resource tf.Output, indices tf.Output, updates tf.Output) (o *tf.Operation)
func ResourceScatterMul(scope *Scope, resource tf.Output, indices tf.Output, updates tf.Output) (o *tf.Operation)
func ResourceScatterNdAdd(scope *Scope, ref tf.Output, indices tf.Output, updates tf.Output, ...) (o *tf.Operation)
func ResourceScatterNdSub(scope *Scope, ref tf.Output, indices tf.Output, updates tf.Output, ...) (o *tf.Operation)
func ResourceScatterNdUpdate(scope *Scope, ref tf.Output, indices tf.Output, updates tf.Output, ...) (o *tf.Operation)
func ResourceScatterSub(scope *Scope, resource tf.Output, indices tf.Output, updates tf.Output) (o *tf.Operation)
func ResourceScatterUpdate(scope *Scope, resource tf.Output, indices tf.Output, updates tf.Output) (o *tf.Operation)
func ResourceSparseApplyAdadelta(scope *Scope, var_ tf.Output, accum tf.Output, accum_update tf.Output, ...) (o *tf.Operation)
func ResourceSparseApplyAdagrad(scope *Scope, var_ tf.Output, accum tf.Output, lr tf.Output, grad tf.Output, ...) (o *tf.Operation)
func ResourceSparseApplyAdagradDA(scope *Scope, var_ tf.Output, gradient_accumulator tf.Output, ...) (o *tf.Operation)
func ResourceSparseApplyAdagradV2(scope *Scope, var_ tf.Output, accum tf.Output, lr tf.Output, epsilon tf.Output, ...) (o *tf.Operation)
func ResourceSparseApplyCenteredRMSProp(scope *Scope, var_ tf.Output, mg tf.Output, ms tf.Output, mom tf.Output, ...) (o *tf.Operation)
func ResourceSparseApplyFtrl(scope *Scope, var_ tf.Output, accum tf.Output, linear tf.Output, ...) (o *tf.Operation)
func ResourceSparseApplyFtrlV2(scope *Scope, var_ tf.Output, accum tf.Output, linear tf.Output, ...) (o *tf.Operation)
func ResourceSparseApplyKerasMomentum(scope *Scope, var_ tf.Output, accum tf.Output, lr tf.Output, grad tf.Output, ...) (o *tf.Operation)
func ResourceSparseApplyMomentum(scope *Scope, var_ tf.Output, accum tf.Output, lr tf.Output, grad tf.Output, ...) (o *tf.Operation)
func ResourceSparseApplyProximalAdagrad(scope *Scope, var_ tf.Output, accum tf.Output, lr tf.Output, l1 tf.Output, ...) (o *tf.Operation)
func ResourceSparseApplyProximalGradientDescent(scope *Scope, var_ tf.Output, alpha tf.Output, l1 tf.Output, l2 tf.Output, ...) (o *tf.Operation)
func ResourceSparseApplyRMSProp(scope *Scope, var_ tf.Output, ms tf.Output, mom tf.Output, lr tf.Output, ...) (o *tf.Operation)
func ResourceStridedSliceAssign(scope *Scope, ref tf.Output, begin tf.Output, end tf.Output, strides tf.Output, ...) (o *tf.Operation)
func Restore(scope *Scope, file_pattern tf.Output, tensor_name tf.Output, dt tf.DataType, ...) (tensor tf.Output)
func RestoreSlice(scope *Scope, file_pattern tf.Output, tensor_name tf.Output, ...) (tensor tf.Output)
func RestoreV2(scope *Scope, prefix tf.Output, tensor_names tf.Output, ...) (tensors []tf.Output)
func RetrieveAllTPUEmbeddingParameters(scope *Scope, NumTables int64, config string, num_shards int64, shard_id int64) (parameters []tf.Output, auxiliary1 []tf.Output, auxiliary2 []tf.Output, ...)
func RetrieveTPUEmbeddingADAMParameters(scope *Scope, num_shards int64, shard_id int64, ...) (parameters tf.Output, momenta tf.Output, velocities tf.Output)
func RetrieveTPUEmbeddingAdadeltaParameters(scope *Scope, num_shards int64, shard_id int64, ...) (parameters tf.Output, accumulators tf.Output, updates tf.Output)
func RetrieveTPUEmbeddingAdagradMomentumParameters(scope *Scope, num_shards int64, shard_id int64, ...) (parameters tf.Output, accumulators tf.Output, momenta tf.Output)
func RetrieveTPUEmbeddingAdagradParameters(scope *Scope, num_shards int64, shard_id int64, ...) (parameters tf.Output, accumulators tf.Output)
func RetrieveTPUEmbeddingCenteredRMSPropParameters(scope *Scope, num_shards int64, shard_id int64, ...) (parameters tf.Output, ms tf.Output, mom tf.Output, mg tf.Output)
func RetrieveTPUEmbeddingFTRLParameters(scope *Scope, num_shards int64, shard_id int64, ...) (parameters tf.Output, accumulators tf.Output, linears tf.Output)
func RetrieveTPUEmbeddingFrequencyEstimatorParameters(scope *Scope, num_shards int64, shard_id int64, ...) (parameters tf.Output, last_hit_step tf.Output)
func RetrieveTPUEmbeddingMDLAdagradLightParameters(scope *Scope, num_shards int64, shard_id int64, ...) (parameters tf.Output, accumulators tf.Output, weights tf.Output, ...)
func RetrieveTPUEmbeddingMomentumParameters(scope *Scope, num_shards int64, shard_id int64, ...) (parameters tf.Output, momenta tf.Output)
func RetrieveTPUEmbeddingProximalAdagradParameters(scope *Scope, num_shards int64, shard_id int64, ...) (parameters tf.Output, accumulators tf.Output)
func RetrieveTPUEmbeddingRMSPropParameters(scope *Scope, num_shards int64, shard_id int64, ...) (parameters tf.Output, ms tf.Output, mom tf.Output)
func RetrieveTPUEmbeddingStochasticGradientDescentParameters(scope *Scope, num_shards int64, shard_id int64, ...) (parameters tf.Output)
func Reverse(scope *Scope, tensor tf.Output, dims tf.Output) (output tf.Output)
func ReverseSequence(scope *Scope, input tf.Output, seq_lengths tf.Output, seq_dim int64, ...) (output tf.Output)
func ReverseV2(scope *Scope, tensor tf.Output, axis tf.Output) (output tf.Output)
func RightShift(scope *Scope, x tf.Output, y tf.Output) (z tf.Output)
func Rint(scope *Scope, x tf.Output) (y tf.Output)
func RiscAdd(scope *Scope, x tf.Output, y tf.Output) (z tf.Output)
func RiscMax(scope *Scope, x tf.Output, y tf.Output) (max tf.Output)
func RngReadAndSkip(scope *Scope, resource tf.Output, alg tf.Output, delta tf.Output) (value tf.Output)
func RngSkip(scope *Scope, resource tf.Output, algorithm tf.Output, delta tf.Output) (o *tf.Operation)
func Roll(scope *Scope, input tf.Output, shift tf.Output, axis tf.Output) (output tf.Output)
func Round(scope *Scope, x tf.Output) (y tf.Output)
func Rsqrt(scope *Scope, x tf.Output) (y tf.Output)
func RsqrtGrad(scope *Scope, y tf.Output, dy tf.Output) (z tf.Output)
func SampleDistortedBoundingBox(scope *Scope, image_size tf.Output, bounding_boxes tf.Output, ...) (begin tf.Output, size tf.Output, bboxes tf.Output)
func SampleDistortedBoundingBoxV2(scope *Scope, image_size tf.Output, bounding_boxes tf.Output, ...) (begin tf.Output, size tf.Output, bboxes tf.Output)
func SamplingDataset(scope *Scope, input_dataset tf.Output, rate tf.Output, seed tf.Output, ...) (handle tf.Output)
func Save(scope *Scope, filename tf.Output, tensor_names tf.Output, data []tf.Output) (o *tf.Operation)
func SaveSlices(scope *Scope, filename tf.Output, tensor_names tf.Output, ...) (o *tf.Operation)
func SaveV2(scope *Scope, prefix tf.Output, tensor_names tf.Output, ...) (o *tf.Operation)
func ScalarSummary(scope *Scope, tags tf.Output, values tf.Output) (summary tf.Output)
func ScatterNd(scope *Scope, indices tf.Output, updates tf.Output, shape tf.Output) (output tf.Output)
func ScatterNdNonAliasingAdd(scope *Scope, input tf.Output, indices tf.Output, updates tf.Output) (output tf.Output)
func SdcaFprint(scope *Scope, input tf.Output) (output tf.Output)
func SdcaOptimizer(scope *Scope, sparse_example_indices []tf.Output, ...) (out_example_state_data tf.Output, out_delta_sparse_weights []tf.Output, ...)
func SdcaOptimizerV2(scope *Scope, sparse_example_indices []tf.Output, ...) (out_example_state_data tf.Output, out_delta_sparse_weights []tf.Output, ...)
func SegmentMax(scope *Scope, data tf.Output, segment_ids tf.Output) (output tf.Output)
func SegmentMaxV2(scope *Scope, data tf.Output, segment_ids tf.Output, num_segments tf.Output) (output tf.Output)
func SegmentMean(scope *Scope, data tf.Output, segment_ids tf.Output) (output tf.Output)
func SegmentMin(scope *Scope, data tf.Output, segment_ids tf.Output) (output tf.Output)
func SegmentMinV2(scope *Scope, data tf.Output, segment_ids tf.Output, num_segments tf.Output) (output tf.Output)
func SegmentProd(scope *Scope, data tf.Output, segment_ids tf.Output) (output tf.Output)
func SegmentProdV2(scope *Scope, data tf.Output, segment_ids tf.Output, num_segments tf.Output) (output tf.Output)
func SegmentSum(scope *Scope, data tf.Output, segment_ids tf.Output) (output tf.Output)
func SegmentSumV2(scope *Scope, data tf.Output, segment_ids tf.Output, num_segments tf.Output) (output tf.Output)
func Select(scope *Scope, condition tf.Output, x tf.Output, y tf.Output) (output tf.Output)
func SelfAdjointEig(scope *Scope, input tf.Output) (output tf.Output)
func SelfAdjointEigV2(scope *Scope, input tf.Output, optional ...SelfAdjointEigV2Attr) (e tf.Output, v tf.Output)
func Selu(scope *Scope, features tf.Output) (activations tf.Output)
func SeluGrad(scope *Scope, gradients tf.Output, outputs tf.Output) (backprops tf.Output)
func Send(scope *Scope, tensor tf.Output, tensor_name string, send_device string, ...) (o *tf.Operation)
func SendTPUEmbeddingGradients(scope *Scope, inputs []tf.Output, learning_rates []tf.Output, config string) (o *tf.Operation)
func SerializeIterator(scope *Scope, resource_handle tf.Output, optional ...SerializeIteratorAttr) (serialized tf.Output)
func SerializeManySparse(scope *Scope, sparse_indices tf.Output, sparse_values tf.Output, ...) (serialized_sparse tf.Output)
func SerializeSparse(scope *Scope, sparse_indices tf.Output, sparse_values tf.Output, ...) (serialized_sparse tf.Output)
func SerializeTensor(scope *Scope, tensor tf.Output) (serialized tf.Output)
func SetSize(scope *Scope, set_indices tf.Output, set_values tf.Output, set_shape tf.Output, ...) (size tf.Output)
func Shape(scope *Scope, input tf.Output, optional ...ShapeAttr) (output tf.Output)
func ShapeN(scope *Scope, input []tf.Output, optional ...ShapeNAttr) (output []tf.Output)
func ShardDataset(scope *Scope, input_dataset tf.Output, num_shards tf.Output, index tf.Output, ...) (handle tf.Output)
func ShardedFilename(scope *Scope, basename tf.Output, shard tf.Output, num_shards tf.Output) (filename tf.Output)
func ShardedFilespec(scope *Scope, basename tf.Output, num_shards tf.Output) (filename tf.Output)
func ShuffleAndRepeatDataset(scope *Scope, input_dataset tf.Output, buffer_size tf.Output, seed tf.Output, ...) (handle tf.Output)
func ShuffleDataset(scope *Scope, input_dataset tf.Output, buffer_size tf.Output, seed tf.Output, ...) (handle tf.Output)
func ShutdownDistributedTPU(scope *Scope) (o *tf.Operation)
func ShutdownTPUSystem(scope *Scope) (success tf.Output)
func Sigmoid(scope *Scope, x tf.Output) (y tf.Output)
func SigmoidGrad(scope *Scope, y tf.Output, dy tf.Output) (z tf.Output)
func Sign(scope *Scope, x tf.Output) (y tf.Output)
func Sin(scope *Scope, x tf.Output) (y tf.Output)
func Sinh(scope *Scope, x tf.Output) (y tf.Output)
func Size(scope *Scope, input tf.Output, optional ...SizeAttr) (output tf.Output)
func SkipDataset(scope *Scope, input_dataset tf.Output, count tf.Output, ...) (handle tf.Output)
func Skipgram(scope *Scope, filename string, batch_size int64, optional ...SkipgramAttr) (vocab_word tf.Output, vocab_freq tf.Output, words_per_epoch tf.Output, ...)
func Slice(scope *Scope, input tf.Output, begin tf.Output, size tf.Output) (output tf.Output)
func SlidingWindowDataset(scope *Scope, input_dataset tf.Output, window_size tf.Output, ...) (handle tf.Output)
func Snapshot(scope *Scope, input tf.Output) (output tf.Output)
func SnapshotDataset(scope *Scope, input_dataset tf.Output, path tf.Output, ...) (handle tf.Output)
func SobolSample(scope *Scope, dim tf.Output, num_results tf.Output, skip tf.Output, ...) (samples tf.Output)
func Softmax(scope *Scope, logits tf.Output) (softmax tf.Output)
func SoftmaxCrossEntropyWithLogits(scope *Scope, features tf.Output, labels tf.Output) (loss tf.Output, backprop tf.Output)
func SoftplusGrad(scope *Scope, gradients tf.Output, features tf.Output) (backprops tf.Output)
func Softsign(scope *Scope, features tf.Output) (activations tf.Output)
func SoftsignGrad(scope *Scope, gradients tf.Output, features tf.Output) (backprops tf.Output)
func SpaceToBatch(scope *Scope, input tf.Output, paddings tf.Output, block_size int64) (output tf.Output)
func SpaceToBatchND(scope *Scope, input tf.Output, block_shape tf.Output, paddings tf.Output) (output tf.Output)
func SpaceToDepth(scope *Scope, input tf.Output, block_size int64, optional ...SpaceToDepthAttr) (output tf.Output)
func SparseAdd(scope *Scope, a_indices tf.Output, a_values tf.Output, a_shape tf.Output, ...) (sum_indices tf.Output, sum_values tf.Output, sum_shape tf.Output)
func SparseAddGrad(scope *Scope, backprop_val_grad tf.Output, a_indices tf.Output, ...) (a_val_grad tf.Output, b_val_grad tf.Output)
func SparseBincount(scope *Scope, indices tf.Output, values tf.Output, dense_shape tf.Output, ...) (output tf.Output)
func SparseConcat(scope *Scope, indices []tf.Output, values []tf.Output, shapes []tf.Output, ...) (output_indices tf.Output, output_values tf.Output, output_shape tf.Output)
func SparseCountSparseOutput(scope *Scope, indices tf.Output, values tf.Output, dense_shape tf.Output, ...) (output_indices tf.Output, output_values tf.Output, ...)
func SparseCross(scope *Scope, indices []tf.Output, values []tf.Output, shapes []tf.Output, ...) (output_indices tf.Output, output_values tf.Output, output_shape tf.Output)
func SparseCrossHashed(scope *Scope, indices []tf.Output, values []tf.Output, shapes []tf.Output, ...) (output_indices tf.Output, output_values tf.Output, output_shape tf.Output)
func SparseCrossV2(scope *Scope, indices []tf.Output, values []tf.Output, shapes []tf.Output, ...) (output_indices tf.Output, output_values tf.Output, output_shape tf.Output)
func SparseDenseCwiseAdd(scope *Scope, sp_indices tf.Output, sp_values tf.Output, sp_shape tf.Output, ...) (output tf.Output)
func SparseDenseCwiseDiv(scope *Scope, sp_indices tf.Output, sp_values tf.Output, sp_shape tf.Output, ...) (output tf.Output)
func SparseDenseCwiseMul(scope *Scope, sp_indices tf.Output, sp_values tf.Output, sp_shape tf.Output, ...) (output tf.Output)
func SparseFillEmptyRows(scope *Scope, indices tf.Output, values tf.Output, dense_shape tf.Output, ...) (output_indices tf.Output, output_values tf.Output, ...)
func SparseFillEmptyRowsGrad(scope *Scope, reverse_index_map tf.Output, grad_values tf.Output) (d_values tf.Output, d_default_value tf.Output)
func SparseMatMul(scope *Scope, a tf.Output, b tf.Output, optional ...SparseMatMulAttr) (product tf.Output)
func SparseMatrixAdd(scope *Scope, a tf.Output, b tf.Output, alpha tf.Output, beta tf.Output) (c tf.Output)
func SparseMatrixMatMul(scope *Scope, a tf.Output, b tf.Output, optional ...SparseMatrixMatMulAttr) (output tf.Output)
func SparseMatrixMul(scope *Scope, a tf.Output, b tf.Output) (output tf.Output)
func SparseMatrixNNZ(scope *Scope, sparse_matrix tf.Output) (nnz tf.Output)
func SparseMatrixOrderingAMD(scope *Scope, input tf.Output) (output tf.Output)
func SparseMatrixSoftmax(scope *Scope, logits tf.Output, type_ tf.DataType) (softmax tf.Output)
func SparseMatrixSoftmaxGrad(scope *Scope, softmax tf.Output, grad_softmax tf.Output, type_ tf.DataType) (gradient tf.Output)
func SparseMatrixSparseCholesky(scope *Scope, input tf.Output, permutation tf.Output, type_ tf.DataType) (output tf.Output)
func SparseMatrixSparseMatMul(scope *Scope, a tf.Output, b tf.Output, type_ tf.DataType, ...) (c tf.Output)
func SparseMatrixTranspose(scope *Scope, input tf.Output, type_ tf.DataType, ...) (output tf.Output)
func SparseMatrixZeros(scope *Scope, dense_shape tf.Output, type_ tf.DataType) (sparse_matrix tf.Output)
func SparseReduceMax(scope *Scope, input_indices tf.Output, input_values tf.Output, ...) (output tf.Output)
func SparseReduceMaxSparse(scope *Scope, input_indices tf.Output, input_values tf.Output, ...) (output_indices tf.Output, output_values tf.Output, output_shape tf.Output)
func SparseReduceSum(scope *Scope, input_indices tf.Output, input_values tf.Output, ...) (output tf.Output)
func SparseReduceSumSparse(scope *Scope, input_indices tf.Output, input_values tf.Output, ...) (output_indices tf.Output, output_values tf.Output, output_shape tf.Output)
func SparseReorder(scope *Scope, input_indices tf.Output, input_values tf.Output, ...) (output_indices tf.Output, output_values tf.Output)
func SparseReshape(scope *Scope, input_indices tf.Output, input_shape tf.Output, ...) (output_indices tf.Output, output_shape tf.Output)
func SparseSegmentMean(scope *Scope, data tf.Output, indices tf.Output, segment_ids tf.Output, ...) (output tf.Output)
func SparseSegmentMeanGrad(scope *Scope, grad tf.Output, indices tf.Output, segment_ids tf.Output, ...) (output tf.Output)
func SparseSegmentMeanGradV2(scope *Scope, grad tf.Output, indices tf.Output, segment_ids tf.Output, ...) (output tf.Output, sorted_unique_indices tf.Output)
func SparseSegmentMeanWithNumSegments(scope *Scope, data tf.Output, indices tf.Output, segment_ids tf.Output, ...) (output tf.Output)
func SparseSegmentSqrtN(scope *Scope, data tf.Output, indices tf.Output, segment_ids tf.Output, ...) (output tf.Output)
func SparseSegmentSqrtNGrad(scope *Scope, grad tf.Output, indices tf.Output, segment_ids tf.Output, ...) (output tf.Output)
func SparseSegmentSqrtNGradV2(scope *Scope, grad tf.Output, indices tf.Output, segment_ids tf.Output, ...) (output tf.Output, sorted_unique_indices tf.Output)
func SparseSegmentSqrtNWithNumSegments(scope *Scope, data tf.Output, indices tf.Output, segment_ids tf.Output, ...) (output tf.Output)
func SparseSegmentSum(scope *Scope, data tf.Output, indices tf.Output, segment_ids tf.Output, ...) (output tf.Output)
func SparseSegmentSumGrad(scope *Scope, grad tf.Output, indices tf.Output, segment_ids tf.Output, ...) (output tf.Output)
func SparseSegmentSumGradV2(scope *Scope, grad tf.Output, indices tf.Output, segment_ids tf.Output, ...) (output tf.Output, sorted_unique_indices tf.Output)
func SparseSegmentSumWithNumSegments(scope *Scope, data tf.Output, indices tf.Output, segment_ids tf.Output, ...) (output tf.Output)
func SparseSlice(scope *Scope, indices tf.Output, values tf.Output, shape tf.Output, ...) (output_indices tf.Output, output_values tf.Output, output_shape tf.Output)
func SparseSliceGrad(scope *Scope, backprop_val_grad tf.Output, input_indices tf.Output, ...) (val_grad tf.Output)
func SparseSoftmax(scope *Scope, sp_indices tf.Output, sp_values tf.Output, sp_shape tf.Output) (output tf.Output)
func SparseSoftmaxCrossEntropyWithLogits(scope *Scope, features tf.Output, labels tf.Output) (loss tf.Output, backprop tf.Output)
func SparseSparseMaximum(scope *Scope, a_indices tf.Output, a_values tf.Output, a_shape tf.Output, ...) (output_indices tf.Output, output_values tf.Output)
func SparseSparseMinimum(scope *Scope, a_indices tf.Output, a_values tf.Output, a_shape tf.Output, ...) (output_indices tf.Output, output_values tf.Output)
func SparseSplit(scope *Scope, split_dim tf.Output, indices tf.Output, values tf.Output, ...) (output_indices []tf.Output, output_values []tf.Output, ...)
func SparseTensorDenseAdd(scope *Scope, a_indices tf.Output, a_values tf.Output, a_shape tf.Output, ...) (output tf.Output)
func SparseTensorDenseMatMul(scope *Scope, a_indices tf.Output, a_values tf.Output, a_shape tf.Output, ...) (product tf.Output)
func SparseTensorSliceDataset(scope *Scope, indices tf.Output, values tf.Output, dense_shape tf.Output) (handle tf.Output)
func SparseTensorToCSRSparseMatrix(scope *Scope, indices tf.Output, values tf.Output, dense_shape tf.Output) (sparse_matrix tf.Output)
func SparseToDense(scope *Scope, sparse_indices tf.Output, output_shape tf.Output, ...) (dense tf.Output)
func SparseToSparseSetOperation(scope *Scope, set1_indices tf.Output, set1_values tf.Output, ...) (result_indices tf.Output, result_values tf.Output, result_shape tf.Output)
func Split(scope *Scope, axis tf.Output, value tf.Output, num_split int64) (output []tf.Output)
func SplitDedupData(scope *Scope, input tf.Output, integer_type tf.DataType, ...) (integer_tensor tf.Output, float_tensor tf.Output)
func SplitV(scope *Scope, value tf.Output, size_splits tf.Output, axis tf.Output, ...) (output []tf.Output)
func SqlDataset(scope *Scope, driver_name tf.Output, data_source_name tf.Output, ...) (handle tf.Output)
func Sqrt(scope *Scope, x tf.Output) (y tf.Output)
func SqrtGrad(scope *Scope, y tf.Output, dy tf.Output) (z tf.Output)
func Square(scope *Scope, x tf.Output) (y tf.Output)
func SquaredDifference(scope *Scope, x tf.Output, y tf.Output) (z tf.Output)
func Squeeze(scope *Scope, input tf.Output, optional ...SqueezeAttr) (output tf.Output)
func StackCloseV2(scope *Scope, handle tf.Output) (o *tf.Operation)
func StackPopV2(scope *Scope, handle tf.Output, elem_type tf.DataType) (elem tf.Output)
func StackPushV2(scope *Scope, handle tf.Output, elem tf.Output, optional ...StackPushV2Attr) (output tf.Output)
func StackV2(scope *Scope, max_size tf.Output, elem_type tf.DataType, ...) (handle tf.Output)
func Stage(scope *Scope, values []tf.Output, optional ...StageAttr) (o *tf.Operation)
func StageClear(scope *Scope, dtypes []tf.DataType, optional ...StageClearAttr) (o *tf.Operation)
func StagePeek(scope *Scope, index tf.Output, dtypes []tf.DataType, optional ...StagePeekAttr) (values []tf.Output)
func StageSize(scope *Scope, dtypes []tf.DataType, optional ...StageSizeAttr) (size tf.Output)
func StatefulStandardNormal(scope *Scope, resource tf.Output, shape tf.Output, ...) (output tf.Output)
func StatefulStandardNormalV2(scope *Scope, resource tf.Output, algorithm tf.Output, shape tf.Output, ...) (output tf.Output)
func StatefulTruncatedNormal(scope *Scope, resource tf.Output, algorithm tf.Output, shape tf.Output, ...) (output tf.Output)
func StatefulUniform(scope *Scope, resource tf.Output, algorithm tf.Output, shape tf.Output, ...) (output tf.Output)
func StatefulUniformFullInt(scope *Scope, resource tf.Output, algorithm tf.Output, shape tf.Output, ...) (output tf.Output)
func StatefulUniformInt(scope *Scope, resource tf.Output, algorithm tf.Output, shape tf.Output, ...) (output tf.Output)
func StatelessMultinomial(scope *Scope, logits tf.Output, num_samples tf.Output, seed tf.Output, ...) (output tf.Output)
func StatelessRandomBinomial(scope *Scope, shape tf.Output, seed tf.Output, counts tf.Output, ...) (output tf.Output)
func StatelessRandomGammaV2(scope *Scope, shape tf.Output, seed tf.Output, alpha tf.Output) (output tf.Output)
func StatelessRandomGammaV3(scope *Scope, shape tf.Output, key tf.Output, counter tf.Output, alg tf.Output, ...) (output tf.Output)
func StatelessRandomGetAlg(scope *Scope) (alg tf.Output)
func StatelessRandomGetKeyCounter(scope *Scope, seed tf.Output) (key tf.Output, counter tf.Output)
func StatelessRandomGetKeyCounterAlg(scope *Scope, seed tf.Output) (key tf.Output, counter tf.Output, alg tf.Output)
func StatelessRandomNormal(scope *Scope, shape tf.Output, seed tf.Output, ...) (output tf.Output)
func StatelessRandomNormalV2(scope *Scope, shape tf.Output, key tf.Output, counter tf.Output, alg tf.Output, ...) (output tf.Output)
func StatelessRandomPoisson(scope *Scope, shape tf.Output, seed tf.Output, lam tf.Output, ...) (output tf.Output)
func StatelessRandomUniform(scope *Scope, shape tf.Output, seed tf.Output, ...) (output tf.Output)
func StatelessRandomUniformFullInt(scope *Scope, shape tf.Output, seed tf.Output, ...) (output tf.Output)
func StatelessRandomUniformFullIntV2(scope *Scope, shape tf.Output, key tf.Output, counter tf.Output, alg tf.Output, ...) (output tf.Output)
func StatelessRandomUniformInt(scope *Scope, shape tf.Output, seed tf.Output, minval tf.Output, ...) (output tf.Output)
func StatelessRandomUniformIntV2(scope *Scope, shape tf.Output, key tf.Output, counter tf.Output, alg tf.Output, ...) (output tf.Output)
func StatelessRandomUniformV2(scope *Scope, shape tf.Output, key tf.Output, counter tf.Output, alg tf.Output, ...) (output tf.Output)
func StatelessSampleDistortedBoundingBox(scope *Scope, image_size tf.Output, bounding_boxes tf.Output, ...) (begin tf.Output, size tf.Output, bboxes tf.Output)
func StatelessShuffle(scope *Scope, value tf.Output, key tf.Output, counter tf.Output, alg tf.Output) (output tf.Output)
func StatelessTruncatedNormal(scope *Scope, shape tf.Output, seed tf.Output, ...) (output tf.Output)
func StatelessTruncatedNormalV2(scope *Scope, shape tf.Output, key tf.Output, counter tf.Output, alg tf.Output, ...) (output tf.Output)
func StaticRegexFullMatch(scope *Scope, input tf.Output, pattern string) (output tf.Output)
func StaticRegexReplace(scope *Scope, input tf.Output, pattern string, rewrite string, ...) (output tf.Output)
func StatsAggregatorHandle(scope *Scope, optional ...StatsAggregatorHandleAttr) (handle tf.Output)
func StatsAggregatorSetSummaryWriter(scope *Scope, stats_aggregator tf.Output, summary tf.Output) (o *tf.Operation)
func StatsAggregatorSummary(scope *Scope, iterator tf.Output) (summary tf.Output)
func StochasticCastToInt(scope *Scope, input tf.Output, key tf.Output, counter tf.Output, alg tf.Output, ...) (output tf.Output)
func StopGradient(scope *Scope, input tf.Output) (output tf.Output)
func StridedSlice(scope *Scope, input tf.Output, begin tf.Output, end tf.Output, ...) (output tf.Output)
func StridedSliceGrad(scope *Scope, shape tf.Output, begin tf.Output, end tf.Output, ...) (output tf.Output)
func StringFormat(scope *Scope, inputs []tf.Output, optional ...StringFormatAttr) (output tf.Output)
func StringJoin(scope *Scope, inputs []tf.Output, optional ...StringJoinAttr) (output tf.Output)
func StringLength(scope *Scope, input tf.Output, optional ...StringLengthAttr) (output tf.Output)
func StringLower(scope *Scope, input tf.Output, optional ...StringLowerAttr) (output tf.Output)
func StringNGrams(scope *Scope, data tf.Output, data_splits tf.Output, separator string, ...) (ngrams tf.Output, ngrams_splits tf.Output)
func StringSplit(scope *Scope, input tf.Output, delimiter tf.Output, ...) (indices tf.Output, values tf.Output, shape tf.Output)
func StringSplitV2(scope *Scope, input tf.Output, sep tf.Output, optional ...StringSplitV2Attr) (indices tf.Output, values tf.Output, shape tf.Output)
func StringStrip(scope *Scope, input tf.Output) (output tf.Output)
func StringToHashBucket(scope *Scope, string_tensor tf.Output, num_buckets int64) (output tf.Output)
func StringToHashBucketFast(scope *Scope, input tf.Output, num_buckets int64) (output tf.Output)
func StringToHashBucketStrong(scope *Scope, input tf.Output, num_buckets int64, key []int64) (output tf.Output)
func StringToNumber(scope *Scope, string_tensor tf.Output, optional ...StringToNumberAttr) (output tf.Output)
func StringUpper(scope *Scope, input tf.Output, optional ...StringUpperAttr) (output tf.Output)
func Sub(scope *Scope, x tf.Output, y tf.Output) (z tf.Output)
func Substr(scope *Scope, input tf.Output, pos tf.Output, len tf.Output, ...) (output tf.Output)
func Sum(scope *Scope, input tf.Output, axis tf.Output, optional ...SumAttr) (output tf.Output)
func Svd(scope *Scope, input tf.Output, optional ...SvdAttr) (s tf.Output, u tf.Output, v tf.Output)
func Switch(scope *Scope, data tf.Output, pred tf.Output) (output_false tf.Output, output_true tf.Output)
func SyncDevice(scope *Scope) (o *tf.Operation)
func TFRecordDataset(scope *Scope, filenames tf.Output, compression_type tf.Output, ...) (handle tf.Output)
func TFRecordDatasetV2(scope *Scope, filenames tf.Output, compression_type tf.Output, ...) (handle tf.Output)
func TFRecordReaderV2(scope *Scope, optional ...TFRecordReaderV2Attr) (reader_handle tf.Output)
func TPUCompilationResult(scope *Scope) (output tf.Output)
func TPUCompileSucceededAssert(scope *Scope, compilation_status tf.Output) (o *tf.Operation)
func TPUCopyWithDynamicShape(scope *Scope, tensors []tf.Output, unpadded_sizes []tf.Output) (tpu_tensors []tf.Output)
func TPUEmbeddingActivations(scope *Scope, embedding_variable tf.Output, sliced_activations tf.Output, ...) (output tf.Output)
func TPUExecute(scope *Scope, args []tf.Output, key tf.Output, Tresults []tf.DataType) (results []tf.Output)
func TPUExecuteAndUpdateVariables(scope *Scope, args []tf.Output, key tf.Output, Tresults []tf.DataType, ...) (results []tf.Output)
func TPUOrdinalSelector(scope *Scope) (device_ordinals tf.Output)
func TPUPartitionedInput(scope *Scope, inputs []tf.Output, optional ...TPUPartitionedInputAttr) (output tf.Output)
func TPUPartitionedInputV2(scope *Scope, inputs []tf.Output, partition_dims []int64, ...) (output tf.Output)
func TPUPartitionedOutput(scope *Scope, inputs tf.Output, num_splits int64, ...) (output []tf.Output)
func TPUPartitionedOutputV2(scope *Scope, inputs tf.Output, num_splits int64, partition_dims []int64) (output []tf.Output)
func TPUReplicateMetadata(scope *Scope, num_replicas int64, optional ...TPUReplicateMetadataAttr) (o *tf.Operation)
func TPUReplicatedInput(scope *Scope, inputs []tf.Output, optional ...TPUReplicatedInputAttr) (output tf.Output)
func TPUReplicatedOutput(scope *Scope, input tf.Output, num_replicas int64) (outputs []tf.Output)
func TPUReshardVariables(scope *Scope, vars []tf.Output, new_format_key tf.Output, ...) (o *tf.Operation)
func TPURoundRobin(scope *Scope) (device_ordinal tf.Output)
func TakeDataset(scope *Scope, input_dataset tf.Output, count tf.Output, ...) (handle tf.Output)
func TakeManySparseFromTensorsMap(scope *Scope, sparse_handles tf.Output, dtype tf.DataType, ...) (sparse_indices tf.Output, sparse_values tf.Output, sparse_shape tf.Output)
func Tan(scope *Scope, x tf.Output) (y tf.Output)
func Tanh(scope *Scope, x tf.Output) (y tf.Output)
func TanhGrad(scope *Scope, y tf.Output, dy tf.Output) (z tf.Output)
func TensorArrayCloseV2(scope *Scope, handle tf.Output) (o *tf.Operation)
func TensorArrayCloseV3(scope *Scope, handle tf.Output) (o *tf.Operation)
func TensorArrayConcatV2(scope *Scope, handle tf.Output, flow_in tf.Output, dtype tf.DataType, ...) (value tf.Output, lengths tf.Output)
func TensorArrayConcatV3(scope *Scope, handle tf.Output, flow_in tf.Output, dtype tf.DataType, ...) (value tf.Output, lengths tf.Output)
func TensorArrayGatherV2(scope *Scope, handle tf.Output, indices tf.Output, flow_in tf.Output, ...) (value tf.Output)
func TensorArrayGatherV3(scope *Scope, handle tf.Output, indices tf.Output, flow_in tf.Output, ...) (value tf.Output)
func TensorArrayGradV2(scope *Scope, handle tf.Output, flow_in tf.Output, source string) (grad_handle tf.Output)
func TensorArrayGradV3(scope *Scope, handle tf.Output, flow_in tf.Output, source string) (grad_handle tf.Output, flow_out tf.Output)
func TensorArrayGradWithShape(scope *Scope, handle tf.Output, flow_in tf.Output, shape_to_prepend tf.Output, ...) (grad_handle tf.Output, flow_out tf.Output)
func TensorArrayReadV2(scope *Scope, handle tf.Output, index tf.Output, flow_in tf.Output, ...) (value tf.Output)
func TensorArrayReadV3(scope *Scope, handle tf.Output, index tf.Output, flow_in tf.Output, ...) (value tf.Output)
func TensorArrayScatterV2(scope *Scope, handle tf.Output, indices tf.Output, value tf.Output, ...) (flow_out tf.Output)
func TensorArrayScatterV3(scope *Scope, handle tf.Output, indices tf.Output, value tf.Output, ...) (flow_out tf.Output)
func TensorArraySizeV2(scope *Scope, handle tf.Output, flow_in tf.Output) (size tf.Output)
func TensorArraySizeV3(scope *Scope, handle tf.Output, flow_in tf.Output) (size tf.Output)
func TensorArraySplitV2(scope *Scope, handle tf.Output, value tf.Output, lengths tf.Output, ...) (flow_out tf.Output)
func TensorArraySplitV3(scope *Scope, handle tf.Output, value tf.Output, lengths tf.Output, ...) (flow_out tf.Output)
func TensorArrayV2(scope *Scope, size tf.Output, dtype tf.DataType, optional ...TensorArrayV2Attr) (handle tf.Output)
func TensorArrayV3(scope *Scope, size tf.Output, dtype tf.DataType, optional ...TensorArrayV3Attr) (handle tf.Output, flow tf.Output)
func TensorArrayWriteV2(scope *Scope, handle tf.Output, index tf.Output, value tf.Output, ...) (flow_out tf.Output)
func TensorArrayWriteV3(scope *Scope, handle tf.Output, index tf.Output, value tf.Output, ...) (flow_out tf.Output)
func TensorDataset(scope *Scope, components []tf.Output, output_shapes []tf.Shape, ...) (handle tf.Output)
func TensorListConcat(scope *Scope, input_handle tf.Output, element_dtype tf.DataType, ...) (tensor tf.Output, lengths tf.Output)
func TensorListConcatV2(scope *Scope, input_handle tf.Output, element_shape tf.Output, ...) (tensor tf.Output, lengths tf.Output)
func TensorListElementShape(scope *Scope, input_handle tf.Output, shape_type tf.DataType) (element_shape tf.Output)
func TensorListFromTensor(scope *Scope, tensor tf.Output, element_shape tf.Output) (output_handle tf.Output)
func TensorListGather(scope *Scope, input_handle tf.Output, indices tf.Output, ...) (values tf.Output)
func TensorListGetItem(scope *Scope, input_handle tf.Output, index tf.Output, element_shape tf.Output, ...) (item tf.Output)
func TensorListLength(scope *Scope, input_handle tf.Output) (length tf.Output)
func TensorListPopBack(scope *Scope, input_handle tf.Output, element_shape tf.Output, ...) (output_handle tf.Output, tensor tf.Output)
func TensorListPushBack(scope *Scope, input_handle tf.Output, tensor tf.Output) (output_handle tf.Output)
func TensorListReserve(scope *Scope, element_shape tf.Output, num_elements tf.Output, ...) (handle tf.Output)
func TensorListResize(scope *Scope, input_handle tf.Output, size tf.Output) (output_handle tf.Output)
func TensorListScatter(scope *Scope, tensor tf.Output, indices tf.Output, element_shape tf.Output) (output_handle tf.Output)
func TensorListScatterIntoExistingList(scope *Scope, input_handle tf.Output, tensor tf.Output, indices tf.Output) (output_handle tf.Output)
func TensorListScatterV2(scope *Scope, tensor tf.Output, indices tf.Output, element_shape tf.Output, ...) (output_handle tf.Output)
func TensorListSetItem(scope *Scope, input_handle tf.Output, index tf.Output, item tf.Output, ...) (output_handle tf.Output)
func TensorListSplit(scope *Scope, tensor tf.Output, element_shape tf.Output, lengths tf.Output) (output_handle tf.Output)
func TensorListStack(scope *Scope, input_handle tf.Output, element_shape tf.Output, ...) (tensor tf.Output)
func TensorMapErase(scope *Scope, input_handle tf.Output, key tf.Output, value_dtype tf.DataType) (output_handle tf.Output)
func TensorMapHasKey(scope *Scope, input_handle tf.Output, key tf.Output) (has_key tf.Output)
func TensorMapInsert(scope *Scope, input_handle tf.Output, key tf.Output, value tf.Output) (output_handle tf.Output)
func TensorMapLookup(scope *Scope, input_handle tf.Output, key tf.Output, value_dtype tf.DataType) (value tf.Output)
func TensorMapSize(scope *Scope, input_handle tf.Output) (size tf.Output)
func TensorMapStackKeys(scope *Scope, input_handle tf.Output, key_dtype tf.DataType) (keys tf.Output)
func TensorScatterAdd(scope *Scope, tensor tf.Output, indices tf.Output, updates tf.Output) (output tf.Output)
func TensorScatterMax(scope *Scope, tensor tf.Output, indices tf.Output, updates tf.Output) (output tf.Output)
func TensorScatterSub(scope *Scope, tensor tf.Output, indices tf.Output, updates tf.Output) (output tf.Output)
func TensorScatterUpdate(scope *Scope, tensor tf.Output, indices tf.Output, updates tf.Output) (output tf.Output)
func TensorSliceDataset(scope *Scope, components []tf.Output, output_shapes []tf.Shape, ...) (handle tf.Output)
func TensorStridedSliceUpdate(scope *Scope, input tf.Output, begin tf.Output, end tf.Output, ...) (output tf.Output)
func TensorSummary(scope *Scope, tensor tf.Output, optional ...TensorSummaryAttr) (summary tf.Output)
func TensorSummaryV2(scope *Scope, tag tf.Output, tensor tf.Output, ...) (summary tf.Output)
func TextLineDataset(scope *Scope, filenames tf.Output, compression_type tf.Output, ...) (handle tf.Output)
func TextLineReaderV2(scope *Scope, optional ...TextLineReaderV2Attr) (reader_handle tf.Output)
func ThreadPoolDataset(scope *Scope, input_dataset tf.Output, thread_pool tf.Output, ...) (handle tf.Output)
func ThreadPoolHandle(scope *Scope, num_threads int64, display_name string, ...) (handle tf.Output)
func ThreadUnsafeUnigramCandidateSampler(scope *Scope, true_classes tf.Output, num_true int64, num_sampled int64, ...) (sampled_candidates tf.Output, true_expected_count tf.Output, ...)
func Tile(scope *Scope, input tf.Output, multiples tf.Output) (output tf.Output)
func TileGrad(scope *Scope, input tf.Output, multiples tf.Output) (output tf.Output)
func Timestamp(scope *Scope) (ts tf.Output)
func ToBool(scope *Scope, input tf.Output) (output tf.Output)
func TopK(scope *Scope, input tf.Output, k int64, optional ...TopKAttr) (values tf.Output, indices tf.Output)
func TopKUnique(scope *Scope, input tf.Output, k int64) (topk tf.Output, topk_indices tf.Output)
func TopKV2(scope *Scope, input tf.Output, k tf.Output, optional ...TopKV2Attr) (values tf.Output, indices tf.Output)
func TopKWithUnique(scope *Scope, input tf.Output, k int64) (topk tf.Output, topk_indices tf.Output)
func TpuHandleToProtoKey(scope *Scope, uid tf.Output) (proto_keys tf.Output)
func Transpose(scope *Scope, x tf.Output, perm tf.Output) (y tf.Output)
func TridiagonalMatMul(scope *Scope, superdiag tf.Output, maindiag tf.Output, subdiag tf.Output, ...) (output tf.Output)
func TridiagonalSolve(scope *Scope, diagonals tf.Output, rhs tf.Output, ...) (output tf.Output)
func TruncateDiv(scope *Scope, x tf.Output, y tf.Output) (z tf.Output)
func TruncateMod(scope *Scope, x tf.Output, y tf.Output) (z tf.Output)
func TruncatedNormal(scope *Scope, shape tf.Output, dtype tf.DataType, ...) (output tf.Output)
func Unbatch(scope *Scope, batched_tensor tf.Output, batch_index tf.Output, id tf.Output, ...) (unbatched_tensor tf.Output)
func UnbatchDataset(scope *Scope, input_dataset tf.Output, output_types []tf.DataType, ...) (handle tf.Output)
func UnbatchGrad(scope *Scope, original_input tf.Output, batch_index tf.Output, grad tf.Output, ...) (batched_grad tf.Output)
func UncompressElement(scope *Scope, compressed tf.Output, output_types []tf.DataType, ...) (components []tf.Output)
func UnicodeDecode(scope *Scope, input tf.Output, input_encoding string, ...) (row_splits tf.Output, char_values tf.Output)
func UnicodeDecodeWithOffsets(scope *Scope, input tf.Output, input_encoding string, ...) (row_splits tf.Output, char_values tf.Output, char_to_byte_starts tf.Output)
func UnicodeEncode(scope *Scope, input_values tf.Output, input_splits tf.Output, ...) (output tf.Output)
func UnicodeScript(scope *Scope, input tf.Output) (output tf.Output)
func UnicodeTranscode(scope *Scope, input tf.Output, input_encoding string, output_encoding string, ...) (output tf.Output)
func UniformCandidateSampler(scope *Scope, true_classes tf.Output, num_true int64, num_sampled int64, ...) (sampled_candidates tf.Output, true_expected_count tf.Output, ...)
func UniformDequantize(scope *Scope, input tf.Output, scales tf.Output, zero_points tf.Output, ...) (output tf.Output)
func UniformQuantize(scope *Scope, input tf.Output, scales tf.Output, zero_points tf.Output, ...) (output tf.Output)
func UniformQuantizedAdd(scope *Scope, lhs tf.Output, rhs tf.Output, lhs_scales tf.Output, ...) (output tf.Output)
func UniformQuantizedClipByValue(scope *Scope, operand tf.Output, min tf.Output, max tf.Output, ...) (output tf.Output)
func UniformQuantizedConvolution(scope *Scope, lhs tf.Output, rhs tf.Output, lhs_scales tf.Output, ...) (output tf.Output)
func UniformQuantizedConvolutionHybrid(scope *Scope, lhs tf.Output, rhs tf.Output, rhs_scales tf.Output, ...) (output tf.Output)
func UniformQuantizedDot(scope *Scope, lhs tf.Output, rhs tf.Output, lhs_scales tf.Output, ...) (output tf.Output)
func UniformQuantizedDotHybrid(scope *Scope, lhs tf.Output, rhs tf.Output, rhs_scales tf.Output, ...) (output tf.Output)
func UniformRequantize(scope *Scope, input tf.Output, input_scales tf.Output, ...) (output tf.Output)
func Unique(scope *Scope, x tf.Output, optional ...UniqueAttr) (y tf.Output, idx tf.Output)
func UniqueDataset(scope *Scope, input_dataset tf.Output, output_types []tf.DataType, ...) (handle tf.Output)
func UniqueV2(scope *Scope, x tf.Output, axis tf.Output, optional ...UniqueV2Attr) (y tf.Output, idx tf.Output)
func UniqueWithCounts(scope *Scope, x tf.Output, optional ...UniqueWithCountsAttr) (y tf.Output, idx tf.Output, count tf.Output)
func UniqueWithCountsV2(scope *Scope, x tf.Output, axis tf.Output, optional ...UniqueWithCountsV2Attr) (y tf.Output, idx tf.Output, count tf.Output)
func Unpack(scope *Scope, value tf.Output, num int64, optional ...UnpackAttr) (output []tf.Output)
func UnravelIndex(scope *Scope, indices tf.Output, dims tf.Output) (output tf.Output)
func UnsortedSegmentMax(scope *Scope, data tf.Output, segment_ids tf.Output, num_segments tf.Output) (output tf.Output)
func UnsortedSegmentMin(scope *Scope, data tf.Output, segment_ids tf.Output, num_segments tf.Output) (output tf.Output)
func UnsortedSegmentProd(scope *Scope, data tf.Output, segment_ids tf.Output, num_segments tf.Output) (output tf.Output)
func UnsortedSegmentSum(scope *Scope, data tf.Output, segment_ids tf.Output, num_segments tf.Output) (output tf.Output)
func Unstage(scope *Scope, dtypes []tf.DataType, optional ...UnstageAttr) (values []tf.Output)
func UpperBound(scope *Scope, sorted_inputs tf.Output, values tf.Output, ...) (output tf.Output)
func VarHandleOp(scope *Scope, dtype tf.DataType, shape tf.Shape, optional ...VarHandleOpAttr) (resource tf.Output)
func VarIsInitializedOp(scope *Scope, resource tf.Output) (is_initialized tf.Output)
func VariableShape(scope *Scope, input tf.Output, optional ...VariableShapeAttr) (output tf.Output)
func Where(scope *Scope, condition tf.Output) (index tf.Output)
func WholeFileReaderV2(scope *Scope, optional ...WholeFileReaderV2Attr) (reader_handle tf.Output)
func WindowDataset(scope *Scope, input_dataset tf.Output, size tf.Output, shift tf.Output, ...) (handle tf.Output)
func WorkerHeartbeat(scope *Scope, request tf.Output) (response tf.Output)
func WriteAudioSummary(scope *Scope, writer tf.Output, step tf.Output, tag tf.Output, ...) (o *tf.Operation)
func WriteFile(scope *Scope, filename tf.Output, contents tf.Output) (o *tf.Operation)
func WriteGraphSummary(scope *Scope, writer tf.Output, step tf.Output, tensor tf.Output) (o *tf.Operation)
func WriteHistogramSummary(scope *Scope, writer tf.Output, step tf.Output, tag tf.Output, ...) (o *tf.Operation)
func WriteImageSummary(scope *Scope, writer tf.Output, step tf.Output, tag tf.Output, ...) (o *tf.Operation)
func WriteRawProtoSummary(scope *Scope, writer tf.Output, step tf.Output, tensor tf.Output) (o *tf.Operation)
func WriteScalarSummary(scope *Scope, writer tf.Output, step tf.Output, tag tf.Output, value tf.Output) (o *tf.Operation)
func WriteSummary(scope *Scope, writer tf.Output, step tf.Output, tensor tf.Output, ...) (o *tf.Operation)
func Xdivy(scope *Scope, x tf.Output, y tf.Output) (z tf.Output)
func XlaAllReduce(scope *Scope, input tf.Output, group_assignment tf.Output, reduce_op string, ...) (output tf.Output)
func XlaBroadcastHelper(scope *Scope, lhs tf.Output, rhs tf.Output, broadcast_dims tf.Output) (lhs_output tf.Output, rhs_output tf.Output)
func XlaConcatND(scope *Scope, inputs []tf.Output, num_concats []int64, ...) (output tf.Output)
func XlaConv(scope *Scope, lhs tf.Output, rhs tf.Output, window_strides tf.Output, ...) (output tf.Output)
func XlaConvV2(scope *Scope, lhs tf.Output, rhs tf.Output, window_strides tf.Output, ...) (output tf.Output)
func XlaCustomCall(scope *Scope, args []tf.Output, target_name string, backend_config string, ...) (output tf.Output)
func XlaCustomCallV2(scope *Scope, operands []tf.Output, call_target_name string, ...) (results []tf.Output)
func XlaDequantize(scope *Scope, input tf.Output, min_range float32, max_range float32, ...) (output tf.Output)
func XlaDot(scope *Scope, lhs tf.Output, rhs tf.Output, dimension_numbers string, ...) (output tf.Output)
func XlaDotV2(scope *Scope, lhs tf.Output, rhs tf.Output, dimension_numbers string, ...) (output tf.Output)
func XlaDynamicSlice(scope *Scope, input tf.Output, start_indices tf.Output, size_indices tf.Output) (output tf.Output)
func XlaDynamicUpdateSlice(scope *Scope, input tf.Output, update tf.Output, indices tf.Output) (output tf.Output)
func XlaEinsum(scope *Scope, a tf.Output, b tf.Output, equation string) (product tf.Output)
func XlaGather(scope *Scope, operand tf.Output, start_indices tf.Output, ...) (output tf.Output)
func XlaKeyValueSort(scope *Scope, keys tf.Output, values tf.Output) (sorted_keys tf.Output, sorted_values tf.Output)
func XlaOptimizationBarrier(scope *Scope, input []tf.Output) (output []tf.Output)
func XlaPad(scope *Scope, input tf.Output, padding_value tf.Output, padding_low tf.Output, ...) (output tf.Output)
func XlaRecv(scope *Scope, dtype tf.DataType, tensor_name string, shape tf.Shape) (tensor tf.Output)
func XlaRecvFromHost(scope *Scope, Toutput tf.DataType, shape tf.Shape, key string) (output tf.Output)
func XlaRecvTPUEmbeddingActivations(scope *Scope, deduplication_data tf.Output, num_tables int64, config string) (outputs []tf.Output)
func XlaRecvTPUEmbeddingDeduplicationData(scope *Scope, config string) (output tf.Output)
func XlaReducePrecision(scope *Scope, operand tf.Output, exponent_bits int64, mantissa_bits int64) (output tf.Output)
func XlaReduceScatter(scope *Scope, input tf.Output, group_assignment tf.Output, ...) (output tf.Output)
func XlaRemoveDynamicDimensionSize(scope *Scope, input tf.Output, dim_index tf.Output) (output tf.Output)
func XlaReplicaId(scope *Scope) (id tf.Output)
func XlaRngBitGenerator(scope *Scope, algorithm tf.Output, initial_state tf.Output, shape tf.Output, ...) (output_key tf.Output, output tf.Output)
func XlaSelfAdjointEig(scope *Scope, a tf.Output, lower bool, max_iter int64, epsilon float32) (w tf.Output, v tf.Output)
func XlaSend(scope *Scope, tensor tf.Output, tensor_name string) (o *tf.Operation)
func XlaSendTPUEmbeddingGradients(scope *Scope, gradients []tf.Output, learning_rates []tf.Output, ...) (o *tf.Operation)
func XlaSendToHost(scope *Scope, input tf.Output, key string) (o *tf.Operation)
func XlaSetBound(scope *Scope, input tf.Output, bound tf.Output) (output tf.Output)
func XlaSetDynamicDimensionSize(scope *Scope, input tf.Output, dim_index tf.Output, size tf.Output) (output tf.Output)
func XlaSharding(scope *Scope, input tf.Output, optional ...XlaShardingAttr) (output tf.Output)
func XlaSort(scope *Scope, input tf.Output) (output tf.Output)
func XlaSplitND(scope *Scope, input tf.Output, N int64, num_splits []int64, ...) (outputs []tf.Output)
func XlaSpmdFullToShardShape(scope *Scope, input tf.Output, manual_sharding string, ...) (output tf.Output)
func XlaSpmdShardToFullShape(scope *Scope, input tf.Output, manual_sharding string, full_shape tf.Shape, ...) (output tf.Output)
func XlaSvd(scope *Scope, a tf.Output, max_iter int64, epsilon float32, ...) (s tf.Output, u tf.Output, v tf.Output)
func Xlog1py(scope *Scope, x tf.Output, y tf.Output) (z tf.Output)
func Xlogy(scope *Scope, x tf.Output, y tf.Output) (z tf.Output)
func ZerosLike(scope *Scope, x tf.Output) (y tf.Output)
func Zeta(scope *Scope, x tf.Output, q tf.Output) (z tf.Output)
func ZipDataset(scope *Scope, input_datasets []tf.Output, output_types []tf.DataType, ...) (handle tf.Output)
type AbortAttr
- func AbortErrorMsg(value string) AbortAttr
- func AbortExitWithoutError(value bool) AbortAttr
type AddManySparseToTensorsMapAttr
- func AddManySparseToTensorsMapContainer(value string) AddManySparseToTensorsMapAttr
- func AddManySparseToTensorsMapSharedName(value string) AddManySparseToTensorsMapAttr
type AddSparseToTensorsMapAttr
- func AddSparseToTensorsMapContainer(value string) AddSparseToTensorsMapAttr
- func AddSparseToTensorsMapSharedName(value string) AddSparseToTensorsMapAttr
type AllAttr
- func AllKeepDims(value bool) AllAttr
type AllCandidateSamplerAttr
- func AllCandidateSamplerSeed(value int64) AllCandidateSamplerAttr
- func AllCandidateSamplerSeed2(value int64) AllCandidateSamplerAttr
type AngleAttr
- func AngleTout(value tf.DataType) AngleAttr
type AnonymousMutableDenseHashTableAttr
- func AnonymousMutableDenseHashTableInitialNumBuckets(value int64) AnonymousMutableDenseHashTableAttr
- func AnonymousMutableDenseHashTableMaxLoadFactor(value float32) AnonymousMutableDenseHashTableAttr
- func AnonymousMutableDenseHashTableValueShape(value tf.Shape) AnonymousMutableDenseHashTableAttr
type AnonymousMutableHashTableOfTensorsAttr
- func AnonymousMutableHashTableOfTensorsValueShape(value tf.Shape) AnonymousMutableHashTableOfTensorsAttr
type AnyAttr
- func AnyKeepDims(value bool) AnyAttr
type ApproxTopKAttr
- func ApproxTopKAggregateToTopk(value bool) ApproxTopKAttr
- func ApproxTopKIsMaxK(value bool) ApproxTopKAttr
- func ApproxTopKRecallTarget(value float32) ApproxTopKAttr
- func ApproxTopKReductionDimension(value int64) ApproxTopKAttr
- func ApproxTopKReductionInputSizeOverride(value int64) ApproxTopKAttr
type ApproximateEqualAttr
- func ApproximateEqualTolerance(value float32) ApproximateEqualAttr
type ArgMaxAttr
- func ArgMaxOutputType(value tf.DataType) ArgMaxAttr
type ArgMinAttr
- func ArgMinOutputType(value tf.DataType) ArgMinAttr
type AsStringAttr
- func AsStringFill(value string) AsStringAttr
- func AsStringPrecision(value int64) AsStringAttr
- func AsStringScientific(value bool) AsStringAttr
- func AsStringShortest(value bool) AsStringAttr
- func AsStringWidth(value int64) AsStringAttr
type AssertAttr
- func AssertSummarize(value int64) AssertAttr
type AssignVariableOpAttr
- func AssignVariableOpValidateShape(value bool) AssignVariableOpAttr
type AssignVariableXlaConcatNDAttr
- func AssignVariableXlaConcatNDPaddings(value []int64) AssignVariableXlaConcatNDAttr
type AudioSpectrogramAttr
- func AudioSpectrogramMagnitudeSquared(value bool) AudioSpectrogramAttr
type AudioSummaryAttr
- func AudioSummaryMaxOutputs(value int64) AudioSummaryAttr
type AudioSummaryV2Attr
- func AudioSummaryV2MaxOutputs(value int64) AudioSummaryV2Attr
type AutoShardDatasetAttr
- func AutoShardDatasetAutoShardPolicy(value int64) AutoShardDatasetAttr
- func AutoShardDatasetNumReplicas(value int64) AutoShardDatasetAttr
type AvgPool3DAttr
- func AvgPool3DDataFormat(value string) AvgPool3DAttr
type AvgPool3DGradAttr
- func AvgPool3DGradDataFormat(value string) AvgPool3DGradAttr
type AvgPoolAttr
- func AvgPoolDataFormat(value string) AvgPoolAttr
type AvgPoolGradAttr
- func AvgPoolGradDataFormat(value string) AvgPoolGradAttr
type BatchAttr
- func BatchAllowedBatchSizes(value []int64) BatchAttr
- func BatchBatchingQueue(value string) BatchAttr
- func BatchContainer(value string) BatchAttr
- func BatchMaxEnqueuedBatches(value int64) BatchAttr
- func BatchSharedName(value string) BatchAttr
type BatchDatasetAttr
- func BatchDatasetMetadata(value string) BatchDatasetAttr
type BatchDatasetV2Attr
- func BatchDatasetV2Metadata(value string) BatchDatasetV2Attr
- func BatchDatasetV2ParallelCopy(value bool) BatchDatasetV2Attr
type BatchMatMulAttr
- func BatchMatMulAdjX(value bool) BatchMatMulAttr
- func BatchMatMulAdjY(value bool) BatchMatMulAttr
- func BatchMatMulGradX(value bool) BatchMatMulAttr
- func BatchMatMulGradY(value bool) BatchMatMulAttr
type BatchMatMulV2Attr
- func BatchMatMulV2AdjX(value bool) BatchMatMulV2Attr
- func BatchMatMulV2AdjY(value bool) BatchMatMulV2Attr
- func BatchMatMulV2GradX(value bool) BatchMatMulV2Attr
- func BatchMatMulV2GradY(value bool) BatchMatMulV2Attr
type BatchMatMulV3Attr
- func BatchMatMulV3AdjX(value bool) BatchMatMulV3Attr
- func BatchMatMulV3AdjY(value bool) BatchMatMulV3Attr
- func BatchMatMulV3GradX(value bool) BatchMatMulV3Attr
- func BatchMatMulV3GradY(value bool) BatchMatMulV3Attr
type BiasAddAttr
- func BiasAddDataFormat(value string) BiasAddAttr
type BiasAddGradAttr
- func BiasAddGradDataFormat(value string) BiasAddGradAttr
type BlockLSTMAttr
- func BlockLSTMCellClip(value float32) BlockLSTMAttr
- func BlockLSTMForgetBias(value float32) BlockLSTMAttr
- func BlockLSTMUsePeephole(value bool) BlockLSTMAttr
type BlockLSTMV2Attr
- func BlockLSTMV2CellClip(value float32) BlockLSTMV2Attr
- func BlockLSTMV2UsePeephole(value bool) BlockLSTMV2Attr
type BoostedTreesCalculateBestFeatureSplitAttr
- func BoostedTreesCalculateBestFeatureSplitSplitType(value string) BoostedTreesCalculateBestFeatureSplitAttr
type BoostedTreesCreateQuantileStreamResourceAttr
- func BoostedTreesCreateQuantileStreamResourceMaxElements(value int64) BoostedTreesCreateQuantileStreamResourceAttr
type BoostedTreesEnsembleResourceHandleOpAttr
- func BoostedTreesEnsembleResourceHandleOpContainer(value string) BoostedTreesEnsembleResourceHandleOpAttr
- func BoostedTreesEnsembleResourceHandleOpSharedName(value string) BoostedTreesEnsembleResourceHandleOpAttr
type BoostedTreesQuantileStreamResourceFlushAttr
- func BoostedTreesQuantileStreamResourceFlushGenerateQuantiles(value bool) BoostedTreesQuantileStreamResourceFlushAttr
type BoostedTreesQuantileStreamResourceHandleOpAttr
- func BoostedTreesQuantileStreamResourceHandleOpContainer(value string) BoostedTreesQuantileStreamResourceHandleOpAttr
- func BoostedTreesQuantileStreamResourceHandleOpSharedName(value string) BoostedTreesQuantileStreamResourceHandleOpAttr
type BoostedTreesSparseCalculateBestFeatureSplitAttr
- func BoostedTreesSparseCalculateBestFeatureSplitSplitType(value string) BoostedTreesSparseCalculateBestFeatureSplitAttr
type BoostedTreesUpdateEnsembleV2Attr
- func BoostedTreesUpdateEnsembleV2LogitsDimension(value int64) BoostedTreesUpdateEnsembleV2Attr
type CTCBeamSearchDecoderAttr
- func CTCBeamSearchDecoderMergeRepeated(value bool) CTCBeamSearchDecoderAttr
type CTCGreedyDecoderAttr
- func CTCGreedyDecoderBlankIndex(value int64) CTCGreedyDecoderAttr
- func CTCGreedyDecoderMergeRepeated(value bool) CTCGreedyDecoderAttr
type CTCLossAttr
- func CTCLossCtcMergeRepeated(value bool) CTCLossAttr
- func CTCLossIgnoreLongerOutputsThanInputs(value bool) CTCLossAttr
- func CTCLossPreprocessCollapseRepeated(value bool) CTCLossAttr
type CTCLossV2Attr
- func CTCLossV2CtcMergeRepeated(value bool) CTCLossV2Attr
- func CTCLossV2IgnoreLongerOutputsThanInputs(value bool) CTCLossV2Attr
- func CTCLossV2PreprocessCollapseRepeated(value bool) CTCLossV2Attr
type CacheDatasetAttr
- func CacheDatasetMetadata(value string) CacheDatasetAttr
type CastAttr
- func CastTruncate(value bool) CastAttr
type CollectiveAllToAllV2Attr
- func CollectiveAllToAllV2CommunicationHint(value string) CollectiveAllToAllV2Attr
- func CollectiveAllToAllV2IsStateless(value bool) CollectiveAllToAllV2Attr
- func CollectiveAllToAllV2TimeoutSeconds(value float32) CollectiveAllToAllV2Attr
type CollectiveAllToAllV3Attr
- func CollectiveAllToAllV3TimeoutSeconds(value float32) CollectiveAllToAllV3Attr
type CollectiveBcastRecvAttr
- func CollectiveBcastRecvCommunicationHint(value string) CollectiveBcastRecvAttr
- func CollectiveBcastRecvTimeoutSeconds(value float32) CollectiveBcastRecvAttr
type CollectiveBcastRecvV2Attr
- func CollectiveBcastRecvV2CommunicationHint(value string) CollectiveBcastRecvV2Attr
- func CollectiveBcastRecvV2TimeoutSeconds(value float32) CollectiveBcastRecvV2Attr
type CollectiveBcastSendAttr
- func CollectiveBcastSendCommunicationHint(value string) CollectiveBcastSendAttr
- func CollectiveBcastSendTimeoutSeconds(value float32) CollectiveBcastSendAttr
type CollectiveBcastSendV2Attr
- func CollectiveBcastSendV2CommunicationHint(value string) CollectiveBcastSendV2Attr
- func CollectiveBcastSendV2TimeoutSeconds(value float32) CollectiveBcastSendV2Attr
type CollectiveGatherAttr
- func CollectiveGatherCommunicationHint(value string) CollectiveGatherAttr
- func CollectiveGatherTimeoutSeconds(value float32) CollectiveGatherAttr
type CollectiveGatherV2Attr
- func CollectiveGatherV2CommunicationHint(value string) CollectiveGatherV2Attr
- func CollectiveGatherV2IsStateless(value bool) CollectiveGatherV2Attr
- func CollectiveGatherV2TimeoutSeconds(value float32) CollectiveGatherV2Attr
type CollectiveInitializeCommunicatorAttr
- func CollectiveInitializeCommunicatorCommunicationHint(value string) CollectiveInitializeCommunicatorAttr
- func CollectiveInitializeCommunicatorTimeoutSeconds(value float32) CollectiveInitializeCommunicatorAttr
type CollectiveReduceAttr
- func CollectiveReduceCommunicationHint(value string) CollectiveReduceAttr
- func CollectiveReduceTimeoutSeconds(value float32) CollectiveReduceAttr
- func CollectiveReduceWaitFor(value []int64) CollectiveReduceAttr
type CollectiveReduceScatterV2Attr
- func CollectiveReduceScatterV2CommunicationHint(value string) CollectiveReduceScatterV2Attr
- func CollectiveReduceScatterV2IsStateless(value bool) CollectiveReduceScatterV2Attr
- func CollectiveReduceScatterV2MaxSubdivsPerDevice(value int64) CollectiveReduceScatterV2Attr
- func CollectiveReduceScatterV2TimeoutSeconds(value float32) CollectiveReduceScatterV2Attr
type CollectiveReduceV2Attr
- func CollectiveReduceV2CommunicationHint(value string) CollectiveReduceV2Attr
- func CollectiveReduceV2IsStateless(value bool) CollectiveReduceV2Attr
- func CollectiveReduceV2MaxSubdivsPerDevice(value int64) CollectiveReduceV2Attr
- func CollectiveReduceV2TimeoutSeconds(value float32) CollectiveReduceV2Attr
type CollectiveReduceV3Attr
- func CollectiveReduceV3TimeoutSeconds(value float32) CollectiveReduceV3Attr
type CombinedNonMaxSuppressionAttr
- func CombinedNonMaxSuppressionClipBoxes(value bool) CombinedNonMaxSuppressionAttr
- func CombinedNonMaxSuppressionPadPerClass(value bool) CombinedNonMaxSuppressionAttr
type ComplexAbsAttr
- func ComplexAbsTout(value tf.DataType) ComplexAbsAttr
type ComplexAttr
- func ComplexTout(value tf.DataType) ComplexAttr
type ComputeAccidentalHitsAttr
- func ComputeAccidentalHitsSeed(value int64) ComputeAccidentalHitsAttr
- func ComputeAccidentalHitsSeed2(value int64) ComputeAccidentalHitsAttr
type ConcatenateDatasetAttr
- func ConcatenateDatasetMetadata(value string) ConcatenateDatasetAttr
type ConfigureAndInitializeGlobalTPUAttr
- func ConfigureAndInitializeGlobalTPUUseTfrtHostRuntime(value bool) ConfigureAndInitializeGlobalTPUAttr
type ConfigureDistributedTPUAttr
- func ConfigureDistributedTPUCompilationFailureClosesChips(value bool) ConfigureDistributedTPUAttr
- func ConfigureDistributedTPUEmbeddingConfig(value string) ConfigureDistributedTPUAttr
- func ConfigureDistributedTPUEnableWholeMeshCompilations(value bool) ConfigureDistributedTPUAttr
- func ConfigureDistributedTPUIsGlobalInit(value bool) ConfigureDistributedTPUAttr
- func ConfigureDistributedTPUTpuCancellationClosesChips(value int64) ConfigureDistributedTPUAttr
- func ConfigureDistributedTPUTpuEmbeddingConfig(value string) ConfigureDistributedTPUAttr
type Conv2DAttr
- func Conv2DDataFormat(value string) Conv2DAttr
- func Conv2DDilations(value []int64) Conv2DAttr
- func Conv2DExplicitPaddings(value []int64) Conv2DAttr
- func Conv2DUseCudnnOnGpu(value bool) Conv2DAttr
type Conv2DBackpropFilterAttr
- func Conv2DBackpropFilterDataFormat(value string) Conv2DBackpropFilterAttr
- func Conv2DBackpropFilterDilations(value []int64) Conv2DBackpropFilterAttr
- func Conv2DBackpropFilterExplicitPaddings(value []int64) Conv2DBackpropFilterAttr
- func Conv2DBackpropFilterUseCudnnOnGpu(value bool) Conv2DBackpropFilterAttr
type Conv2DBackpropFilterV2Attr
- func Conv2DBackpropFilterV2DataFormat(value string) Conv2DBackpropFilterV2Attr
- func Conv2DBackpropFilterV2Dilations(value []int64) Conv2DBackpropFilterV2Attr
- func Conv2DBackpropFilterV2ExplicitPaddings(value []int64) Conv2DBackpropFilterV2Attr
- func Conv2DBackpropFilterV2UseCudnnOnGpu(value bool) Conv2DBackpropFilterV2Attr
type Conv2DBackpropInputAttr
- func Conv2DBackpropInputDataFormat(value string) Conv2DBackpropInputAttr
- func Conv2DBackpropInputDilations(value []int64) Conv2DBackpropInputAttr
- func Conv2DBackpropInputExplicitPaddings(value []int64) Conv2DBackpropInputAttr
- func Conv2DBackpropInputUseCudnnOnGpu(value bool) Conv2DBackpropInputAttr
type Conv2DBackpropInputV2Attr
- func Conv2DBackpropInputV2DataFormat(value string) Conv2DBackpropInputV2Attr
- func Conv2DBackpropInputV2Dilations(value []int64) Conv2DBackpropInputV2Attr
- func Conv2DBackpropInputV2ExplicitPaddings(value []int64) Conv2DBackpropInputV2Attr
- func Conv2DBackpropInputV2UseCudnnOnGpu(value bool) Conv2DBackpropInputV2Attr
type Conv3DAttr
- func Conv3DDataFormat(value string) Conv3DAttr
- func Conv3DDilations(value []int64) Conv3DAttr
type Conv3DBackpropFilterAttr
- func Conv3DBackpropFilterDilations(value []int64) Conv3DBackpropFilterAttr
type Conv3DBackpropFilterV2Attr
- func Conv3DBackpropFilterV2DataFormat(value string) Conv3DBackpropFilterV2Attr
- func Conv3DBackpropFilterV2Dilations(value []int64) Conv3DBackpropFilterV2Attr
type Conv3DBackpropInputAttr
- func Conv3DBackpropInputDilations(value []int64) Conv3DBackpropInputAttr
type Conv3DBackpropInputV2Attr
- func Conv3DBackpropInputV2DataFormat(value string) Conv3DBackpropInputV2Attr
- func Conv3DBackpropInputV2Dilations(value []int64) Conv3DBackpropInputV2Attr
type ConvAttr
- func ConvBatchDims(value int64) ConvAttr
- func ConvDataFormat(value string) ConvAttr
- func ConvDilations(value []int64) ConvAttr
- func ConvExplicitPaddings(value []int64) ConvAttr
- func ConvGroups(value int64) ConvAttr
type CopyAttr
- func CopyDebugOpsSpec(value []string) CopyAttr
- func CopyTensorName(value string) CopyAttr
type CopyHostAttr
- func CopyHostDebugOpsSpec(value []string) CopyHostAttr
- func CopyHostTensorName(value string) CopyHostAttr
type CropAndResizeAttr
- func CropAndResizeExtrapolationValue(value float32) CropAndResizeAttr
- func CropAndResizeMethod(value string) CropAndResizeAttr
type CropAndResizeGradBoxesAttr
- func CropAndResizeGradBoxesMethod(value string) CropAndResizeGradBoxesAttr
type CropAndResizeGradImageAttr
- func CropAndResizeGradImageMethod(value string) CropAndResizeGradImageAttr
type CudnnRNNAttr
- func CudnnRNNDirection(value string) CudnnRNNAttr
- func CudnnRNNDropout(value float32) CudnnRNNAttr
- func CudnnRNNInputMode(value string) CudnnRNNAttr
- func CudnnRNNIsTraining(value bool) CudnnRNNAttr
- func CudnnRNNRnnMode(value string) CudnnRNNAttr
- func CudnnRNNSeed(value int64) CudnnRNNAttr
- func CudnnRNNSeed2(value int64) CudnnRNNAttr
type CudnnRNNBackpropAttr
- func CudnnRNNBackpropDirection(value string) CudnnRNNBackpropAttr
- func CudnnRNNBackpropDropout(value float32) CudnnRNNBackpropAttr
- func CudnnRNNBackpropInputMode(value string) CudnnRNNBackpropAttr
- func CudnnRNNBackpropRnnMode(value string) CudnnRNNBackpropAttr
- func CudnnRNNBackpropSeed(value int64) CudnnRNNBackpropAttr
- func CudnnRNNBackpropSeed2(value int64) CudnnRNNBackpropAttr
type CudnnRNNBackpropV2Attr
- func CudnnRNNBackpropV2Direction(value string) CudnnRNNBackpropV2Attr
- func CudnnRNNBackpropV2Dropout(value float32) CudnnRNNBackpropV2Attr
- func CudnnRNNBackpropV2InputMode(value string) CudnnRNNBackpropV2Attr
- func CudnnRNNBackpropV2RnnMode(value string) CudnnRNNBackpropV2Attr
- func CudnnRNNBackpropV2Seed(value int64) CudnnRNNBackpropV2Attr
- func CudnnRNNBackpropV2Seed2(value int64) CudnnRNNBackpropV2Attr
type CudnnRNNBackpropV3Attr
- func CudnnRNNBackpropV3Direction(value string) CudnnRNNBackpropV3Attr
- func CudnnRNNBackpropV3Dropout(value float32) CudnnRNNBackpropV3Attr
- func CudnnRNNBackpropV3InputMode(value string) CudnnRNNBackpropV3Attr
- func CudnnRNNBackpropV3NumProj(value int64) CudnnRNNBackpropV3Attr
- func CudnnRNNBackpropV3RnnMode(value string) CudnnRNNBackpropV3Attr
- func CudnnRNNBackpropV3Seed(value int64) CudnnRNNBackpropV3Attr
- func CudnnRNNBackpropV3Seed2(value int64) CudnnRNNBackpropV3Attr
- func CudnnRNNBackpropV3TimeMajor(value bool) CudnnRNNBackpropV3Attr
type CudnnRNNCanonicalToParamsAttr
- func CudnnRNNCanonicalToParamsDirection(value string) CudnnRNNCanonicalToParamsAttr
- func CudnnRNNCanonicalToParamsDropout(value float32) CudnnRNNCanonicalToParamsAttr
- func CudnnRNNCanonicalToParamsInputMode(value string) CudnnRNNCanonicalToParamsAttr
- func CudnnRNNCanonicalToParamsRnnMode(value string) CudnnRNNCanonicalToParamsAttr
- func CudnnRNNCanonicalToParamsSeed(value int64) CudnnRNNCanonicalToParamsAttr
- func CudnnRNNCanonicalToParamsSeed2(value int64) CudnnRNNCanonicalToParamsAttr
type CudnnRNNCanonicalToParamsV2Attr
- func CudnnRNNCanonicalToParamsV2Direction(value string) CudnnRNNCanonicalToParamsV2Attr
- func CudnnRNNCanonicalToParamsV2Dropout(value float32) CudnnRNNCanonicalToParamsV2Attr
- func CudnnRNNCanonicalToParamsV2InputMode(value string) CudnnRNNCanonicalToParamsV2Attr
- func CudnnRNNCanonicalToParamsV2NumProj(value int64) CudnnRNNCanonicalToParamsV2Attr
- func CudnnRNNCanonicalToParamsV2RnnMode(value string) CudnnRNNCanonicalToParamsV2Attr
- func CudnnRNNCanonicalToParamsV2Seed(value int64) CudnnRNNCanonicalToParamsV2Attr
- func CudnnRNNCanonicalToParamsV2Seed2(value int64) CudnnRNNCanonicalToParamsV2Attr
type CudnnRNNParamsSizeAttr
- func CudnnRNNParamsSizeDirection(value string) CudnnRNNParamsSizeAttr
- func CudnnRNNParamsSizeDropout(value float32) CudnnRNNParamsSizeAttr
- func CudnnRNNParamsSizeInputMode(value string) CudnnRNNParamsSizeAttr
- func CudnnRNNParamsSizeNumProj(value int64) CudnnRNNParamsSizeAttr
- func CudnnRNNParamsSizeRnnMode(value string) CudnnRNNParamsSizeAttr
- func CudnnRNNParamsSizeSeed(value int64) CudnnRNNParamsSizeAttr
- func CudnnRNNParamsSizeSeed2(value int64) CudnnRNNParamsSizeAttr
type CudnnRNNParamsToCanonicalAttr
- func CudnnRNNParamsToCanonicalDirection(value string) CudnnRNNParamsToCanonicalAttr
- func CudnnRNNParamsToCanonicalDropout(value float32) CudnnRNNParamsToCanonicalAttr
- func CudnnRNNParamsToCanonicalInputMode(value string) CudnnRNNParamsToCanonicalAttr
- func CudnnRNNParamsToCanonicalRnnMode(value string) CudnnRNNParamsToCanonicalAttr
- func CudnnRNNParamsToCanonicalSeed(value int64) CudnnRNNParamsToCanonicalAttr
- func CudnnRNNParamsToCanonicalSeed2(value int64) CudnnRNNParamsToCanonicalAttr
type CudnnRNNParamsToCanonicalV2Attr
- func CudnnRNNParamsToCanonicalV2Direction(value string) CudnnRNNParamsToCanonicalV2Attr
- func CudnnRNNParamsToCanonicalV2Dropout(value float32) CudnnRNNParamsToCanonicalV2Attr
- func CudnnRNNParamsToCanonicalV2InputMode(value string) CudnnRNNParamsToCanonicalV2Attr
- func CudnnRNNParamsToCanonicalV2NumProj(value int64) CudnnRNNParamsToCanonicalV2Attr
- func CudnnRNNParamsToCanonicalV2RnnMode(value string) CudnnRNNParamsToCanonicalV2Attr
- func CudnnRNNParamsToCanonicalV2Seed(value int64) CudnnRNNParamsToCanonicalV2Attr
- func CudnnRNNParamsToCanonicalV2Seed2(value int64) CudnnRNNParamsToCanonicalV2Attr
type CudnnRNNV2Attr
- func CudnnRNNV2Direction(value string) CudnnRNNV2Attr
- func CudnnRNNV2Dropout(value float32) CudnnRNNV2Attr
- func CudnnRNNV2InputMode(value string) CudnnRNNV2Attr
- func CudnnRNNV2IsTraining(value bool) CudnnRNNV2Attr
- func CudnnRNNV2RnnMode(value string) CudnnRNNV2Attr
- func CudnnRNNV2Seed(value int64) CudnnRNNV2Attr
- func CudnnRNNV2Seed2(value int64) CudnnRNNV2Attr
type CudnnRNNV3Attr
- func CudnnRNNV3Direction(value string) CudnnRNNV3Attr
- func CudnnRNNV3Dropout(value float32) CudnnRNNV3Attr
- func CudnnRNNV3InputMode(value string) CudnnRNNV3Attr
- func CudnnRNNV3IsTraining(value bool) CudnnRNNV3Attr
- func CudnnRNNV3NumProj(value int64) CudnnRNNV3Attr
- func CudnnRNNV3RnnMode(value string) CudnnRNNV3Attr
- func CudnnRNNV3Seed(value int64) CudnnRNNV3Attr
- func CudnnRNNV3Seed2(value int64) CudnnRNNV3Attr
- func CudnnRNNV3TimeMajor(value bool) CudnnRNNV3Attr
type CumprodAttr
- func CumprodExclusive(value bool) CumprodAttr
- func CumprodReverse(value bool) CumprodAttr
type CumsumAttr
- func CumsumExclusive(value bool) CumsumAttr
- func CumsumReverse(value bool) CumsumAttr
type CumulativeLogsumexpAttr
- func CumulativeLogsumexpExclusive(value bool) CumulativeLogsumexpAttr
- func CumulativeLogsumexpReverse(value bool) CumulativeLogsumexpAttr
type DataFormatDimMapAttr
- func DataFormatDimMapDstFormat(value string) DataFormatDimMapAttr
- func DataFormatDimMapSrcFormat(value string) DataFormatDimMapAttr
type DataFormatVecPermuteAttr
- func DataFormatVecPermuteDstFormat(value string) DataFormatVecPermuteAttr
- func DataFormatVecPermuteSrcFormat(value string) DataFormatVecPermuteAttr
type DataServiceDatasetAttr
- func DataServiceDatasetCrossTrainerCacheOptions(value string) DataServiceDatasetAttr
- func DataServiceDatasetDataTransferProtocol(value string) DataServiceDatasetAttr
- func DataServiceDatasetTargetWorkers(value string) DataServiceDatasetAttr
- func DataServiceDatasetTaskRefreshIntervalHintMs(value int64) DataServiceDatasetAttr
type DataServiceDatasetV2Attr
- func DataServiceDatasetV2CrossTrainerCacheOptions(value string) DataServiceDatasetV2Attr
- func DataServiceDatasetV2DataTransferProtocol(value string) DataServiceDatasetV2Attr
- func DataServiceDatasetV2TargetWorkers(value string) DataServiceDatasetV2Attr
- func DataServiceDatasetV2TaskRefreshIntervalHintMs(value int64) DataServiceDatasetV2Attr
type DatasetCardinalityAttr
- func DatasetCardinalityCardinalityOptions(value string) DatasetCardinalityAttr
type DatasetToGraphAttr
- func DatasetToGraphAllowStateful(value bool) DatasetToGraphAttr
- func DatasetToGraphStatefulWhitelist(value []string) DatasetToGraphAttr
- func DatasetToGraphStripDeviceAssignment(value bool) DatasetToGraphAttr
type DatasetToGraphV2Attr
- func DatasetToGraphV2ExternalStatePolicy(value int64) DatasetToGraphV2Attr
- func DatasetToGraphV2StripDeviceAssignment(value bool) DatasetToGraphV2Attr
type DatasetToSingleElementAttr
- func DatasetToSingleElementMetadata(value string) DatasetToSingleElementAttr
type DebugIdentityAttr
- func DebugIdentityDebugUrls(value []string) DebugIdentityAttr
- func DebugIdentityDeviceName(value string) DebugIdentityAttr
- func DebugIdentityGatedGrpc(value bool) DebugIdentityAttr
- func DebugIdentityTensorName(value string) DebugIdentityAttr
type DebugIdentityV2Attr
- func DebugIdentityV2CircularBufferSize(value int64) DebugIdentityV2Attr
- func DebugIdentityV2DebugUrls(value []string) DebugIdentityV2Attr
- func DebugIdentityV2OpName(value string) DebugIdentityV2Attr
- func DebugIdentityV2OutputSlot(value int64) DebugIdentityV2Attr
- func DebugIdentityV2TensorDebugMode(value int64) DebugIdentityV2Attr
- func DebugIdentityV2TfdbgContextId(value string) DebugIdentityV2Attr
- func DebugIdentityV2TfdbgRunId(value string) DebugIdentityV2Attr
type DebugIdentityV3Attr
- func DebugIdentityV3DebugUrls(value []string) DebugIdentityV3Attr
- func DebugIdentityV3DeviceName(value string) DebugIdentityV3Attr
- func DebugIdentityV3GatedGrpc(value bool) DebugIdentityV3Attr
- func DebugIdentityV3IoIndex(value int64) DebugIdentityV3Attr
- func DebugIdentityV3IoOfNode(value string) DebugIdentityV3Attr
- func DebugIdentityV3IsInput(value bool) DebugIdentityV3Attr
- func DebugIdentityV3TensorName(value string) DebugIdentityV3Attr
type DebugNanCountAttr
- func DebugNanCountDebugUrls(value []string) DebugNanCountAttr
- func DebugNanCountDeviceName(value string) DebugNanCountAttr
- func DebugNanCountGatedGrpc(value bool) DebugNanCountAttr
- func DebugNanCountTensorName(value string) DebugNanCountAttr
type DebugNumericSummaryAttr
- func DebugNumericSummaryDebugUrls(value []string) DebugNumericSummaryAttr
- func DebugNumericSummaryDeviceName(value string) DebugNumericSummaryAttr
- func DebugNumericSummaryGatedGrpc(value bool) DebugNumericSummaryAttr
- func DebugNumericSummaryLowerBound(value float32) DebugNumericSummaryAttr
- func DebugNumericSummaryMuteIfHealthy(value bool) DebugNumericSummaryAttr
- func DebugNumericSummaryTensorName(value string) DebugNumericSummaryAttr
- func DebugNumericSummaryUpperBound(value float32) DebugNumericSummaryAttr
type DebugNumericSummaryV2Attr
- func DebugNumericSummaryV2OutputDtype(value tf.DataType) DebugNumericSummaryV2Attr
- func DebugNumericSummaryV2TensorDebugMode(value int64) DebugNumericSummaryV2Attr
- func DebugNumericSummaryV2TensorId(value int64) DebugNumericSummaryV2Attr
type DecodeAndCropJpegAttr
- func DecodeAndCropJpegAcceptableFraction(value float32) DecodeAndCropJpegAttr
- func DecodeAndCropJpegChannels(value int64) DecodeAndCropJpegAttr
- func DecodeAndCropJpegDctMethod(value string) DecodeAndCropJpegAttr
- func DecodeAndCropJpegFancyUpscaling(value bool) DecodeAndCropJpegAttr
- func DecodeAndCropJpegRatio(value int64) DecodeAndCropJpegAttr
- func DecodeAndCropJpegTryRecoverTruncated(value bool) DecodeAndCropJpegAttr
type DecodeBmpAttr
- func DecodeBmpChannels(value int64) DecodeBmpAttr
type DecodeCSVAttr
- func DecodeCSVFieldDelim(value string) DecodeCSVAttr
- func DecodeCSVNaValue(value string) DecodeCSVAttr
- func DecodeCSVSelectCols(value []int64) DecodeCSVAttr
- func DecodeCSVUseQuoteDelim(value bool) DecodeCSVAttr
type DecodeCompressedAttr
- func DecodeCompressedCompressionType(value string) DecodeCompressedAttr
type DecodeImageAttr
- func DecodeImageChannels(value int64) DecodeImageAttr
- func DecodeImageDtype(value tf.DataType) DecodeImageAttr
- func DecodeImageExpandAnimations(value bool) DecodeImageAttr
type DecodeJpegAttr
- func DecodeJpegAcceptableFraction(value float32) DecodeJpegAttr
- func DecodeJpegChannels(value int64) DecodeJpegAttr
- func DecodeJpegDctMethod(value string) DecodeJpegAttr
- func DecodeJpegFancyUpscaling(value bool) DecodeJpegAttr
- func DecodeJpegRatio(value int64) DecodeJpegAttr
- func DecodeJpegTryRecoverTruncated(value bool) DecodeJpegAttr
type DecodePaddedRawAttr
- func DecodePaddedRawLittleEndian(value bool) DecodePaddedRawAttr
type DecodePngAttr
- func DecodePngChannels(value int64) DecodePngAttr
- func DecodePngDtype(value tf.DataType) DecodePngAttr
type DecodeProtoV2Attr
- func DecodeProtoV2DescriptorSource(value string) DecodeProtoV2Attr
- func DecodeProtoV2MessageFormat(value string) DecodeProtoV2Attr
- func DecodeProtoV2Sanitize(value bool) DecodeProtoV2Attr
type DecodeRawAttr
- func DecodeRawLittleEndian(value bool) DecodeRawAttr
type DecodeWavAttr
- func DecodeWavDesiredChannels(value int64) DecodeWavAttr
- func DecodeWavDesiredSamples(value int64) DecodeWavAttr
type DenseBincountAttr
- func DenseBincountBinaryOutput(value bool) DenseBincountAttr
type DenseCountSparseOutputAttr
- func DenseCountSparseOutputMaxlength(value int64) DenseCountSparseOutputAttr
- func DenseCountSparseOutputMinlength(value int64) DenseCountSparseOutputAttr
type DenseToDenseSetOperationAttr
- func DenseToDenseSetOperationValidateIndices(value bool) DenseToDenseSetOperationAttr
type DenseToSparseSetOperationAttr
- func DenseToSparseSetOperationValidateIndices(value bool) DenseToSparseSetOperationAttr
type DepthToSpaceAttr
- func DepthToSpaceDataFormat(value string) DepthToSpaceAttr
type DepthwiseConv2dNativeAttr
- func DepthwiseConv2dNativeDataFormat(value string) DepthwiseConv2dNativeAttr
- func DepthwiseConv2dNativeDilations(value []int64) DepthwiseConv2dNativeAttr
- func DepthwiseConv2dNativeExplicitPaddings(value []int64) DepthwiseConv2dNativeAttr
type DepthwiseConv2dNativeBackpropFilterAttr
- func DepthwiseConv2dNativeBackpropFilterDataFormat(value string) DepthwiseConv2dNativeBackpropFilterAttr
- func DepthwiseConv2dNativeBackpropFilterDilations(value []int64) DepthwiseConv2dNativeBackpropFilterAttr
- func DepthwiseConv2dNativeBackpropFilterExplicitPaddings(value []int64) DepthwiseConv2dNativeBackpropFilterAttr
type DepthwiseConv2dNativeBackpropInputAttr
- func DepthwiseConv2dNativeBackpropInputDataFormat(value string) DepthwiseConv2dNativeBackpropInputAttr
- func DepthwiseConv2dNativeBackpropInputDilations(value []int64) DepthwiseConv2dNativeBackpropInputAttr
- func DepthwiseConv2dNativeBackpropInputExplicitPaddings(value []int64) DepthwiseConv2dNativeBackpropInputAttr
type DequantizeAttr
- func DequantizeAxis(value int64) DequantizeAttr
- func DequantizeDtype(value tf.DataType) DequantizeAttr
- func DequantizeMode(value string) DequantizeAttr
- func DequantizeNarrowRange(value bool) DequantizeAttr
type DestroyResourceOpAttr
- func DestroyResourceOpIgnoreLookupError(value bool) DestroyResourceOpAttr
type DirectedInterleaveDatasetAttr
- func DirectedInterleaveDatasetStopOnEmptyDataset(value bool) DirectedInterleaveDatasetAttr
type DynamicEnqueueTPUEmbeddingArbitraryTensorBatchAttr
- func DynamicEnqueueTPUEmbeddingArbitraryTensorBatchCombiners(value []string) DynamicEnqueueTPUEmbeddingArbitraryTensorBatchAttr
type EagerPyFuncAttr
- func EagerPyFuncIsAsync(value bool) EagerPyFuncAttr
type EditDistanceAttr
- func EditDistanceNormalize(value bool) EditDistanceAttr
type EigAttr
- func EigComputeV(value bool) EigAttr
type EmptyAttr
- func EmptyInit(value bool) EmptyAttr
type EncodeBase64Attr
- func EncodeBase64Pad(value bool) EncodeBase64Attr
type EncodeJpegAttr
- func EncodeJpegChromaDownsampling(value bool) EncodeJpegAttr
- func EncodeJpegDensityUnit(value string) EncodeJpegAttr
- func EncodeJpegFormat(value string) EncodeJpegAttr
- func EncodeJpegOptimizeSize(value bool) EncodeJpegAttr
- func EncodeJpegProgressive(value bool) EncodeJpegAttr
- func EncodeJpegQuality(value int64) EncodeJpegAttr
- func EncodeJpegXDensity(value int64) EncodeJpegAttr
- func EncodeJpegXmpMetadata(value string) EncodeJpegAttr
- func EncodeJpegYDensity(value int64) EncodeJpegAttr
type EncodePngAttr
- func EncodePngCompression(value int64) EncodePngAttr
type EncodeProtoAttr
- func EncodeProtoDescriptorSource(value string) EncodeProtoAttr
type EnqueueTPUEmbeddingArbitraryTensorBatchAttr
- func EnqueueTPUEmbeddingArbitraryTensorBatchCombiners(value []string) EnqueueTPUEmbeddingArbitraryTensorBatchAttr
- func EnqueueTPUEmbeddingArbitraryTensorBatchDeviceOrdinal(value int64) EnqueueTPUEmbeddingArbitraryTensorBatchAttr
type EnqueueTPUEmbeddingBatchAttr
- func EnqueueTPUEmbeddingBatchCombiners(value []string) EnqueueTPUEmbeddingBatchAttr
- func EnqueueTPUEmbeddingBatchDeviceOrdinal(value int64) EnqueueTPUEmbeddingBatchAttr
type EnqueueTPUEmbeddingIntegerBatchAttr
- func EnqueueTPUEmbeddingIntegerBatchDeviceOrdinal(value int64) EnqueueTPUEmbeddingIntegerBatchAttr
type EnqueueTPUEmbeddingRaggedTensorBatchAttr
- func EnqueueTPUEmbeddingRaggedTensorBatchCombiners(value []string) EnqueueTPUEmbeddingRaggedTensorBatchAttr
- func EnqueueTPUEmbeddingRaggedTensorBatchDeviceOrdinal(value int64) EnqueueTPUEmbeddingRaggedTensorBatchAttr
- func EnqueueTPUEmbeddingRaggedTensorBatchMaxSequenceLengths(value []int64) EnqueueTPUEmbeddingRaggedTensorBatchAttr
- func EnqueueTPUEmbeddingRaggedTensorBatchNumFeatures(value []int64) EnqueueTPUEmbeddingRaggedTensorBatchAttr
type EnqueueTPUEmbeddingSparseBatchAttr
- func EnqueueTPUEmbeddingSparseBatchCombiners(value []string) EnqueueTPUEmbeddingSparseBatchAttr
- func EnqueueTPUEmbeddingSparseBatchDeviceOrdinal(value int64) EnqueueTPUEmbeddingSparseBatchAttr
type EnqueueTPUEmbeddingSparseTensorBatchAttr
- func EnqueueTPUEmbeddingSparseTensorBatchCombiners(value []string) EnqueueTPUEmbeddingSparseTensorBatchAttr
- func EnqueueTPUEmbeddingSparseTensorBatchDeviceOrdinal(value int64) EnqueueTPUEmbeddingSparseTensorBatchAttr
- func EnqueueTPUEmbeddingSparseTensorBatchMaxSequenceLengths(value []int64) EnqueueTPUEmbeddingSparseTensorBatchAttr
- func EnqueueTPUEmbeddingSparseTensorBatchNumFeatures(value []int64) EnqueueTPUEmbeddingSparseTensorBatchAttr
type EnterAttr
- func EnterIsConstant(value bool) EnterAttr
- func EnterParallelIterations(value int64) EnterAttr
type EqualAttr
- func EqualIncompatibleShapeError(value bool) EqualAttr
type EuclideanNormAttr
- func EuclideanNormKeepDims(value bool) EuclideanNormAttr
type ExperimentalAutoShardDatasetAttr
- func ExperimentalAutoShardDatasetAutoShardPolicy(value int64) ExperimentalAutoShardDatasetAttr
type ExperimentalIgnoreErrorsDatasetAttr
- func ExperimentalIgnoreErrorsDatasetLogWarning(value bool) ExperimentalIgnoreErrorsDatasetAttr
type ExperimentalParseExampleDatasetAttr
- func ExperimentalParseExampleDatasetSloppy(value bool) ExperimentalParseExampleDatasetAttr
type ExperimentalRebatchDatasetAttr
- func ExperimentalRebatchDatasetUseFallback(value bool) ExperimentalRebatchDatasetAttr
type ExperimentalStatsAggregatorHandleAttr
- func ExperimentalStatsAggregatorHandleContainer(value string) ExperimentalStatsAggregatorHandleAttr
- func ExperimentalStatsAggregatorHandleSharedName(value string) ExperimentalStatsAggregatorHandleAttr
type ExperimentalThreadPoolHandleAttr
- func ExperimentalThreadPoolHandleContainer(value string) ExperimentalThreadPoolHandleAttr
- func ExperimentalThreadPoolHandleMaxIntraOpParallelism(value int64) ExperimentalThreadPoolHandleAttr
- func ExperimentalThreadPoolHandleSharedName(value string) ExperimentalThreadPoolHandleAttr
type ExtractGlimpseAttr
- func ExtractGlimpseCentered(value bool) ExtractGlimpseAttr
- func ExtractGlimpseNoise(value string) ExtractGlimpseAttr
- func ExtractGlimpseNormalized(value bool) ExtractGlimpseAttr
- func ExtractGlimpseUniformNoise(value bool) ExtractGlimpseAttr
type ExtractGlimpseV2Attr
- func ExtractGlimpseV2Centered(value bool) ExtractGlimpseV2Attr
- func ExtractGlimpseV2Noise(value string) ExtractGlimpseV2Attr
- func ExtractGlimpseV2Normalized(value bool) ExtractGlimpseV2Attr
- func ExtractGlimpseV2UniformNoise(value bool) ExtractGlimpseV2Attr
type ExtractJpegShapeAttr
- func ExtractJpegShapeOutputType(value tf.DataType) ExtractJpegShapeAttr
type FIFOQueueV2Attr
- func FIFOQueueV2Capacity(value int64) FIFOQueueV2Attr
- func FIFOQueueV2Container(value string) FIFOQueueV2Attr
- func FIFOQueueV2Shapes(value []tf.Shape) FIFOQueueV2Attr
- func FIFOQueueV2SharedName(value string) FIFOQueueV2Attr
type FakeQuantWithMinMaxArgsAttr
- func FakeQuantWithMinMaxArgsMax(value float32) FakeQuantWithMinMaxArgsAttr
- func FakeQuantWithMinMaxArgsMin(value float32) FakeQuantWithMinMaxArgsAttr
- func FakeQuantWithMinMaxArgsNarrowRange(value bool) FakeQuantWithMinMaxArgsAttr
- func FakeQuantWithMinMaxArgsNumBits(value int64) FakeQuantWithMinMaxArgsAttr
type FakeQuantWithMinMaxArgsGradientAttr
- func FakeQuantWithMinMaxArgsGradientMax(value float32) FakeQuantWithMinMaxArgsGradientAttr
- func FakeQuantWithMinMaxArgsGradientMin(value float32) FakeQuantWithMinMaxArgsGradientAttr
- func FakeQuantWithMinMaxArgsGradientNarrowRange(value bool) FakeQuantWithMinMaxArgsGradientAttr
- func FakeQuantWithMinMaxArgsGradientNumBits(value int64) FakeQuantWithMinMaxArgsGradientAttr
type FakeQuantWithMinMaxVarsAttr
- func FakeQuantWithMinMaxVarsNarrowRange(value bool) FakeQuantWithMinMaxVarsAttr
- func FakeQuantWithMinMaxVarsNumBits(value int64) FakeQuantWithMinMaxVarsAttr
type FakeQuantWithMinMaxVarsGradientAttr
- func FakeQuantWithMinMaxVarsGradientNarrowRange(value bool) FakeQuantWithMinMaxVarsGradientAttr
- func FakeQuantWithMinMaxVarsGradientNumBits(value int64) FakeQuantWithMinMaxVarsGradientAttr
type FakeQuantWithMinMaxVarsPerChannelAttr
- func FakeQuantWithMinMaxVarsPerChannelNarrowRange(value bool) FakeQuantWithMinMaxVarsPerChannelAttr
- func FakeQuantWithMinMaxVarsPerChannelNumBits(value int64) FakeQuantWithMinMaxVarsPerChannelAttr
type FakeQuantWithMinMaxVarsPerChannelGradientAttr
- func FakeQuantWithMinMaxVarsPerChannelGradientNarrowRange(value bool) FakeQuantWithMinMaxVarsPerChannelGradientAttr
- func FakeQuantWithMinMaxVarsPerChannelGradientNumBits(value int64) FakeQuantWithMinMaxVarsPerChannelGradientAttr
type FinalizeDatasetAttr
- func FinalizeDatasetHasCapturedRef(value bool) FinalizeDatasetAttr
type FixedLengthRecordDatasetAttr
- func FixedLengthRecordDatasetMetadata(value string) FixedLengthRecordDatasetAttr
type FixedLengthRecordReaderV2Attr
- func FixedLengthRecordReaderV2Container(value string) FixedLengthRecordReaderV2Attr
- func FixedLengthRecordReaderV2Encoding(value string) FixedLengthRecordReaderV2Attr
- func FixedLengthRecordReaderV2FooterBytes(value int64) FixedLengthRecordReaderV2Attr
- func FixedLengthRecordReaderV2HeaderBytes(value int64) FixedLengthRecordReaderV2Attr
- func FixedLengthRecordReaderV2HopBytes(value int64) FixedLengthRecordReaderV2Attr
- func FixedLengthRecordReaderV2SharedName(value string) FixedLengthRecordReaderV2Attr
type FixedUnigramCandidateSamplerAttr
- func FixedUnigramCandidateSamplerDistortion(value float32) FixedUnigramCandidateSamplerAttr
- func FixedUnigramCandidateSamplerNumReservedIds(value int64) FixedUnigramCandidateSamplerAttr
- func FixedUnigramCandidateSamplerNumShards(value int64) FixedUnigramCandidateSamplerAttr
- func FixedUnigramCandidateSamplerSeed(value int64) FixedUnigramCandidateSamplerAttr
- func FixedUnigramCandidateSamplerSeed2(value int64) FixedUnigramCandidateSamplerAttr
- func FixedUnigramCandidateSamplerShard(value int64) FixedUnigramCandidateSamplerAttr
- func FixedUnigramCandidateSamplerUnigrams(value []float32) FixedUnigramCandidateSamplerAttr
- func FixedUnigramCandidateSamplerVocabFile(value string) FixedUnigramCandidateSamplerAttr
type FractionalAvgPoolAttr
- func FractionalAvgPoolDeterministic(value bool) FractionalAvgPoolAttr
- func FractionalAvgPoolOverlapping(value bool) FractionalAvgPoolAttr
- func FractionalAvgPoolPseudoRandom(value bool) FractionalAvgPoolAttr
- func FractionalAvgPoolSeed(value int64) FractionalAvgPoolAttr
- func FractionalAvgPoolSeed2(value int64) FractionalAvgPoolAttr
type FractionalAvgPoolGradAttr
- func FractionalAvgPoolGradOverlapping(value bool) FractionalAvgPoolGradAttr
type FractionalMaxPoolAttr
- func FractionalMaxPoolDeterministic(value bool) FractionalMaxPoolAttr
- func FractionalMaxPoolOverlapping(value bool) FractionalMaxPoolAttr
- func FractionalMaxPoolPseudoRandom(value bool) FractionalMaxPoolAttr
- func FractionalMaxPoolSeed(value int64) FractionalMaxPoolAttr
- func FractionalMaxPoolSeed2(value int64) FractionalMaxPoolAttr
type FractionalMaxPoolGradAttr
- func FractionalMaxPoolGradOverlapping(value bool) FractionalMaxPoolGradAttr
type FusedBatchNormAttr
- func FusedBatchNormDataFormat(value string) FusedBatchNormAttr
- func FusedBatchNormEpsilon(value float32) FusedBatchNormAttr
- func FusedBatchNormExponentialAvgFactor(value float32) FusedBatchNormAttr
- func FusedBatchNormIsTraining(value bool) FusedBatchNormAttr
type FusedBatchNormGradAttr
- func FusedBatchNormGradDataFormat(value string) FusedBatchNormGradAttr
- func FusedBatchNormGradEpsilon(value float32) FusedBatchNormGradAttr
- func FusedBatchNormGradIsTraining(value bool) FusedBatchNormGradAttr
type FusedBatchNormGradV2Attr
- func FusedBatchNormGradV2DataFormat(value string) FusedBatchNormGradV2Attr
- func FusedBatchNormGradV2Epsilon(value float32) FusedBatchNormGradV2Attr
- func FusedBatchNormGradV2IsTraining(value bool) FusedBatchNormGradV2Attr
type FusedBatchNormGradV3Attr
- func FusedBatchNormGradV3DataFormat(value string) FusedBatchNormGradV3Attr
- func FusedBatchNormGradV3Epsilon(value float32) FusedBatchNormGradV3Attr
- func FusedBatchNormGradV3IsTraining(value bool) FusedBatchNormGradV3Attr
type FusedBatchNormV2Attr
- func FusedBatchNormV2DataFormat(value string) FusedBatchNormV2Attr
- func FusedBatchNormV2Epsilon(value float32) FusedBatchNormV2Attr
- func FusedBatchNormV2ExponentialAvgFactor(value float32) FusedBatchNormV2Attr
- func FusedBatchNormV2IsTraining(value bool) FusedBatchNormV2Attr
type FusedBatchNormV3Attr
- func FusedBatchNormV3DataFormat(value string) FusedBatchNormV3Attr
- func FusedBatchNormV3Epsilon(value float32) FusedBatchNormV3Attr
- func FusedBatchNormV3ExponentialAvgFactor(value float32) FusedBatchNormV3Attr
- func FusedBatchNormV3IsTraining(value bool) FusedBatchNormV3Attr
type FusedResizeAndPadConv2DAttr
- func FusedResizeAndPadConv2DResizeAlignCorners(value bool) FusedResizeAndPadConv2DAttr
type GatherAttr
- func GatherValidateIndices(value bool) GatherAttr
type GatherV2Attr
- func GatherV2BatchDims(value int64) GatherV2Attr
type GenerateBoundingBoxProposalsAttr
- func GenerateBoundingBoxProposalsPostNmsTopn(value int64) GenerateBoundingBoxProposalsAttr
type GenerateVocabRemappingAttr
- func GenerateVocabRemappingOldVocabSize(value int64) GenerateVocabRemappingAttr
type HashTableV2Attr
- func HashTableV2Container(value string) HashTableV2Attr
- func HashTableV2SharedName(value string) HashTableV2Attr
- func HashTableV2UseNodeNameSharing(value bool) HashTableV2Attr
type HistogramFixedWidthAttr
- func HistogramFixedWidthDtype(value tf.DataType) HistogramFixedWidthAttr
type IRFFT2DAttr
- func IRFFT2DTreal(value tf.DataType) IRFFT2DAttr
type IRFFT3DAttr
- func IRFFT3DTreal(value tf.DataType) IRFFT3DAttr
type IRFFTAttr
- func IRFFTTreal(value tf.DataType) IRFFTAttr
type IRFFTNDAttr
- func IRFFTNDTreal(value tf.DataType) IRFFTNDAttr
type IdentityReaderV2Attr
- func IdentityReaderV2Container(value string) IdentityReaderV2Attr
- func IdentityReaderV2SharedName(value string) IdentityReaderV2Attr
type IgnoreErrorsDatasetAttr
- func IgnoreErrorsDatasetLogWarning(value bool) IgnoreErrorsDatasetAttr
type ImagAttr
- func ImagTout(value tf.DataType) ImagAttr
type ImageProjectiveTransformV2Attr
- func ImageProjectiveTransformV2FillMode(value string) ImageProjectiveTransformV2Attr
type ImageProjectiveTransformV3Attr
- func ImageProjectiveTransformV3FillMode(value string) ImageProjectiveTransformV3Attr
type ImageSummaryAttr
- func ImageSummaryBadColor(value tf.Tensor) ImageSummaryAttr
- func ImageSummaryMaxImages(value int64) ImageSummaryAttr
type InfeedEnqueueAttr
- func InfeedEnqueueDeviceOrdinal(value int64) InfeedEnqueueAttr
- func InfeedEnqueueLayout(value []int64) InfeedEnqueueAttr
- func InfeedEnqueueShape(value tf.Shape) InfeedEnqueueAttr
type InfeedEnqueuePrelinearizedBufferAttr
- func InfeedEnqueuePrelinearizedBufferDeviceOrdinal(value int64) InfeedEnqueuePrelinearizedBufferAttr
type InfeedEnqueueTupleAttr
- func InfeedEnqueueTupleDeviceOrdinal(value int64) InfeedEnqueueTupleAttr
- func InfeedEnqueueTupleLayouts(value []int64) InfeedEnqueueTupleAttr
type InitializeTableFromTextFileV2Attr
- func InitializeTableFromTextFileV2Delimiter(value string) InitializeTableFromTextFileV2Attr
- func InitializeTableFromTextFileV2Offset(value int64) InitializeTableFromTextFileV2Attr
- func InitializeTableFromTextFileV2VocabSize(value int64) InitializeTableFromTextFileV2Attr
type IsTPUEmbeddingInitializedAttr
- func IsTPUEmbeddingInitializedConfig(value string) IsTPUEmbeddingInitializedAttr
type IsotonicRegressionAttr
- func IsotonicRegressionOutputDtype(value tf.DataType) IsotonicRegressionAttr
type IteratorFromStringHandleAttr
- func IteratorFromStringHandleOutputShapes(value []tf.Shape) IteratorFromStringHandleAttr
- func IteratorFromStringHandleOutputTypes(value []tf.DataType) IteratorFromStringHandleAttr
type LRNAttr
- func LRNAlpha(value float32) LRNAttr
- func LRNBeta(value float32) LRNAttr
- func LRNBias(value float32) LRNAttr
- func LRNDepthRadius(value int64) LRNAttr
type LRNGradAttr
- func LRNGradAlpha(value float32) LRNGradAttr
- func LRNGradBeta(value float32) LRNGradAttr
- func LRNGradBias(value float32) LRNGradAttr
- func LRNGradDepthRadius(value int64) LRNGradAttr
type LSTMBlockCellAttr
- func LSTMBlockCellCellClip(value float32) LSTMBlockCellAttr
- func LSTMBlockCellForgetBias(value float32) LSTMBlockCellAttr
- func LSTMBlockCellUsePeephole(value bool) LSTMBlockCellAttr
type LeakyReluAttr
- func LeakyReluAlpha(value float32) LeakyReluAttr
type LeakyReluGradAttr
- func LeakyReluGradAlpha(value float32) LeakyReluGradAttr
type LearnedUnigramCandidateSamplerAttr
- func LearnedUnigramCandidateSamplerSeed(value int64) LearnedUnigramCandidateSamplerAttr
- func LearnedUnigramCandidateSamplerSeed2(value int64) LearnedUnigramCandidateSamplerAttr
type ListDatasetAttr
- func ListDatasetMetadata(value string) ListDatasetAttr
type ListDiffAttr
- func ListDiffOutIdx(value tf.DataType) ListDiffAttr
type LoadAndRemapMatrixAttr
- func LoadAndRemapMatrixMaxRowsInMemory(value int64) LoadAndRemapMatrixAttr
type LoadTPUEmbeddingADAMParametersAttr
- func LoadTPUEmbeddingADAMParametersConfig(value string) LoadTPUEmbeddingADAMParametersAttr
- func LoadTPUEmbeddingADAMParametersTableId(value int64) LoadTPUEmbeddingADAMParametersAttr
- func LoadTPUEmbeddingADAMParametersTableName(value string) LoadTPUEmbeddingADAMParametersAttr
type LoadTPUEmbeddingAdadeltaParametersAttr
- func LoadTPUEmbeddingAdadeltaParametersConfig(value string) LoadTPUEmbeddingAdadeltaParametersAttr
- func LoadTPUEmbeddingAdadeltaParametersTableId(value int64) LoadTPUEmbeddingAdadeltaParametersAttr
- func LoadTPUEmbeddingAdadeltaParametersTableName(value string) LoadTPUEmbeddingAdadeltaParametersAttr
type LoadTPUEmbeddingAdagradMomentumParametersAttr
- func LoadTPUEmbeddingAdagradMomentumParametersConfig(value string) LoadTPUEmbeddingAdagradMomentumParametersAttr
- func LoadTPUEmbeddingAdagradMomentumParametersTableId(value int64) LoadTPUEmbeddingAdagradMomentumParametersAttr
- func LoadTPUEmbeddingAdagradMomentumParametersTableName(value string) LoadTPUEmbeddingAdagradMomentumParametersAttr
type LoadTPUEmbeddingAdagradParametersAttr
- func LoadTPUEmbeddingAdagradParametersConfig(value string) LoadTPUEmbeddingAdagradParametersAttr
- func LoadTPUEmbeddingAdagradParametersTableId(value int64) LoadTPUEmbeddingAdagradParametersAttr
- func LoadTPUEmbeddingAdagradParametersTableName(value string) LoadTPUEmbeddingAdagradParametersAttr
type LoadTPUEmbeddingCenteredRMSPropParametersAttr
- func LoadTPUEmbeddingCenteredRMSPropParametersConfig(value string) LoadTPUEmbeddingCenteredRMSPropParametersAttr
- func LoadTPUEmbeddingCenteredRMSPropParametersTableId(value int64) LoadTPUEmbeddingCenteredRMSPropParametersAttr
- func LoadTPUEmbeddingCenteredRMSPropParametersTableName(value string) LoadTPUEmbeddingCenteredRMSPropParametersAttr
type LoadTPUEmbeddingFTRLParametersAttr
- func LoadTPUEmbeddingFTRLParametersConfig(value string) LoadTPUEmbeddingFTRLParametersAttr
- func LoadTPUEmbeddingFTRLParametersTableId(value int64) LoadTPUEmbeddingFTRLParametersAttr
- func LoadTPUEmbeddingFTRLParametersTableName(value string) LoadTPUEmbeddingFTRLParametersAttr
type LoadTPUEmbeddingFrequencyEstimatorParametersAttr
- func LoadTPUEmbeddingFrequencyEstimatorParametersConfig(value string) LoadTPUEmbeddingFrequencyEstimatorParametersAttr
- func LoadTPUEmbeddingFrequencyEstimatorParametersTableId(value int64) LoadTPUEmbeddingFrequencyEstimatorParametersAttr
- func LoadTPUEmbeddingFrequencyEstimatorParametersTableName(value string) LoadTPUEmbeddingFrequencyEstimatorParametersAttr
type LoadTPUEmbeddingMDLAdagradLightParametersAttr
- func LoadTPUEmbeddingMDLAdagradLightParametersConfig(value string) LoadTPUEmbeddingMDLAdagradLightParametersAttr
- func LoadTPUEmbeddingMDLAdagradLightParametersTableId(value int64) LoadTPUEmbeddingMDLAdagradLightParametersAttr
- func LoadTPUEmbeddingMDLAdagradLightParametersTableName(value string) LoadTPUEmbeddingMDLAdagradLightParametersAttr
type LoadTPUEmbeddingMomentumParametersAttr
- func LoadTPUEmbeddingMomentumParametersConfig(value string) LoadTPUEmbeddingMomentumParametersAttr
- func LoadTPUEmbeddingMomentumParametersTableId(value int64) LoadTPUEmbeddingMomentumParametersAttr
- func LoadTPUEmbeddingMomentumParametersTableName(value string) LoadTPUEmbeddingMomentumParametersAttr
type LoadTPUEmbeddingProximalAdagradParametersAttr
- func LoadTPUEmbeddingProximalAdagradParametersConfig(value string) LoadTPUEmbeddingProximalAdagradParametersAttr
- func LoadTPUEmbeddingProximalAdagradParametersTableId(value int64) LoadTPUEmbeddingProximalAdagradParametersAttr
- func LoadTPUEmbeddingProximalAdagradParametersTableName(value string) LoadTPUEmbeddingProximalAdagradParametersAttr
type LoadTPUEmbeddingRMSPropParametersAttr
- func LoadTPUEmbeddingRMSPropParametersConfig(value string) LoadTPUEmbeddingRMSPropParametersAttr
- func LoadTPUEmbeddingRMSPropParametersTableId(value int64) LoadTPUEmbeddingRMSPropParametersAttr
- func LoadTPUEmbeddingRMSPropParametersTableName(value string) LoadTPUEmbeddingRMSPropParametersAttr
type LoadTPUEmbeddingStochasticGradientDescentParametersAttr
- func LoadTPUEmbeddingStochasticGradientDescentParametersConfig(value string) LoadTPUEmbeddingStochasticGradientDescentParametersAttr
- func LoadTPUEmbeddingStochasticGradientDescentParametersTableId(value int64) LoadTPUEmbeddingStochasticGradientDescentParametersAttr
- func LoadTPUEmbeddingStochasticGradientDescentParametersTableName(value string) LoadTPUEmbeddingStochasticGradientDescentParametersAttr
type LogUniformCandidateSamplerAttr
- func LogUniformCandidateSamplerSeed(value int64) LogUniformCandidateSamplerAttr
- func LogUniformCandidateSamplerSeed2(value int64) LogUniformCandidateSamplerAttr
type LowerBoundAttr
- func LowerBoundOutType(value tf.DataType) LowerBoundAttr
type LuAttr
- func LuOutputIdxType(value tf.DataType) LuAttr
type MapClearAttr
- func MapClearCapacity(value int64) MapClearAttr
- func MapClearContainer(value string) MapClearAttr
- func MapClearMemoryLimit(value int64) MapClearAttr
- func MapClearSharedName(value string) MapClearAttr
type MapIncompleteSizeAttr
- func MapIncompleteSizeCapacity(value int64) MapIncompleteSizeAttr
- func MapIncompleteSizeContainer(value string) MapIncompleteSizeAttr
- func MapIncompleteSizeMemoryLimit(value int64) MapIncompleteSizeAttr
- func MapIncompleteSizeSharedName(value string) MapIncompleteSizeAttr
type MapPeekAttr
- func MapPeekCapacity(value int64) MapPeekAttr
- func MapPeekContainer(value string) MapPeekAttr
- func MapPeekMemoryLimit(value int64) MapPeekAttr
- func MapPeekSharedName(value string) MapPeekAttr
type MapSizeAttr
- func MapSizeCapacity(value int64) MapSizeAttr
- func MapSizeContainer(value string) MapSizeAttr
- func MapSizeMemoryLimit(value int64) MapSizeAttr
- func MapSizeSharedName(value string) MapSizeAttr
type MapStageAttr
- func MapStageCapacity(value int64) MapStageAttr
- func MapStageContainer(value string) MapStageAttr
- func MapStageMemoryLimit(value int64) MapStageAttr
- func MapStageSharedName(value string) MapStageAttr
type MapUnstageAttr
- func MapUnstageCapacity(value int64) MapUnstageAttr
- func MapUnstageContainer(value string) MapUnstageAttr
- func MapUnstageMemoryLimit(value int64) MapUnstageAttr
- func MapUnstageSharedName(value string) MapUnstageAttr
type MapUnstageNoKeyAttr
- func MapUnstageNoKeyCapacity(value int64) MapUnstageNoKeyAttr
- func MapUnstageNoKeyContainer(value string) MapUnstageNoKeyAttr
- func MapUnstageNoKeyMemoryLimit(value int64) MapUnstageNoKeyAttr
- func MapUnstageNoKeySharedName(value string) MapUnstageNoKeyAttr
type MatMulAttr
- func MatMulGradA(value bool) MatMulAttr
- func MatMulGradB(value bool) MatMulAttr
- func MatMulTransposeA(value bool) MatMulAttr
- func MatMulTransposeB(value bool) MatMulAttr
type MatrixDiagPartV3Attr
- func MatrixDiagPartV3Align(value string) MatrixDiagPartV3Attr
type MatrixDiagV3Attr
- func MatrixDiagV3Align(value string) MatrixDiagV3Attr
type MatrixInverseAttr
- func MatrixInverseAdjoint(value bool) MatrixInverseAttr
type MatrixSetDiagV3Attr
- func MatrixSetDiagV3Align(value string) MatrixSetDiagV3Attr
type MatrixSolveAttr
- func MatrixSolveAdjoint(value bool) MatrixSolveAttr
type MatrixSolveLsAttr
- func MatrixSolveLsFast(value bool) MatrixSolveLsAttr
type MatrixTriangularSolveAttr
- func MatrixTriangularSolveAdjoint(value bool) MatrixTriangularSolveAttr
- func MatrixTriangularSolveLower(value bool) MatrixTriangularSolveAttr
type MaxAttr
- func MaxKeepDims(value bool) MaxAttr
type MaxPool3DAttr
- func MaxPool3DDataFormat(value string) MaxPool3DAttr
type MaxPool3DGradAttr
- func MaxPool3DGradDataFormat(value string) MaxPool3DGradAttr
type MaxPool3DGradGradAttr
- func MaxPool3DGradGradDataFormat(value string) MaxPool3DGradGradAttr
type MaxPoolAttr
- func MaxPoolDataFormat(value string) MaxPoolAttr
- func MaxPoolExplicitPaddings(value []int64) MaxPoolAttr
type MaxPoolGradAttr
- func MaxPoolGradDataFormat(value string) MaxPoolGradAttr
- func MaxPoolGradExplicitPaddings(value []int64) MaxPoolGradAttr
type MaxPoolGradGradAttr
- func MaxPoolGradGradDataFormat(value string) MaxPoolGradGradAttr
type MaxPoolGradGradV2Attr
- func MaxPoolGradGradV2DataFormat(value string) MaxPoolGradGradV2Attr
type MaxPoolGradGradWithArgmaxAttr
- func MaxPoolGradGradWithArgmaxIncludeBatchInIndex(value bool) MaxPoolGradGradWithArgmaxAttr
type MaxPoolGradV2Attr
- func MaxPoolGradV2DataFormat(value string) MaxPoolGradV2Attr
type MaxPoolGradWithArgmaxAttr
- func MaxPoolGradWithArgmaxIncludeBatchInIndex(value bool) MaxPoolGradWithArgmaxAttr
type MaxPoolV2Attr
- func MaxPoolV2DataFormat(value string) MaxPoolV2Attr
type MaxPoolWithArgmaxAttr
- func MaxPoolWithArgmaxIncludeBatchInIndex(value bool) MaxPoolWithArgmaxAttr
- func MaxPoolWithArgmaxTargmax(value tf.DataType) MaxPoolWithArgmaxAttr
type MeanAttr
- func MeanKeepDims(value bool) MeanAttr
type MergeDedupDataAttr
- func MergeDedupDataConfig(value string) MergeDedupDataAttr
type MergeV2CheckpointsAttr
- func MergeV2CheckpointsAllowMissingFiles(value bool) MergeV2CheckpointsAttr
- func MergeV2CheckpointsDeleteOldDirs(value bool) MergeV2CheckpointsAttr
type MfccAttr
- func MfccDctCoefficientCount(value int64) MfccAttr
- func MfccFilterbankChannelCount(value int64) MfccAttr
- func MfccLowerFrequencyLimit(value float32) MfccAttr
- func MfccUpperFrequencyLimit(value float32) MfccAttr
type MinAttr
- func MinKeepDims(value bool) MinAttr
type ModelDatasetAttr
- func ModelDatasetAlgorithm(value int64) ModelDatasetAttr
- func ModelDatasetCpuBudget(value int64) ModelDatasetAttr
- func ModelDatasetRamBudget(value int64) ModelDatasetAttr
type MultiDeviceIteratorFromStringHandleAttr
- func MultiDeviceIteratorFromStringHandleOutputShapes(value []tf.Shape) MultiDeviceIteratorFromStringHandleAttr
- func MultiDeviceIteratorFromStringHandleOutputTypes(value []tf.DataType) MultiDeviceIteratorFromStringHandleAttr
type MultinomialAttr
- func MultinomialOutputDtype(value tf.DataType) MultinomialAttr
- func MultinomialSeed(value int64) MultinomialAttr
- func MultinomialSeed2(value int64) MultinomialAttr
type MutableDenseHashTableV2Attr
- func MutableDenseHashTableV2Container(value string) MutableDenseHashTableV2Attr
- func MutableDenseHashTableV2InitialNumBuckets(value int64) MutableDenseHashTableV2Attr
- func MutableDenseHashTableV2MaxLoadFactor(value float32) MutableDenseHashTableV2Attr
- func MutableDenseHashTableV2SharedName(value string) MutableDenseHashTableV2Attr
- func MutableDenseHashTableV2UseNodeNameSharing(value bool) MutableDenseHashTableV2Attr
- func MutableDenseHashTableV2ValueShape(value tf.Shape) MutableDenseHashTableV2Attr
type MutableHashTableOfTensorsV2Attr
- func MutableHashTableOfTensorsV2Container(value string) MutableHashTableOfTensorsV2Attr
- func MutableHashTableOfTensorsV2SharedName(value string) MutableHashTableOfTensorsV2Attr
- func MutableHashTableOfTensorsV2UseNodeNameSharing(value bool) MutableHashTableOfTensorsV2Attr
- func MutableHashTableOfTensorsV2ValueShape(value tf.Shape) MutableHashTableOfTensorsV2Attr
type MutableHashTableV2Attr
- func MutableHashTableV2Container(value string) MutableHashTableV2Attr
- func MutableHashTableV2SharedName(value string) MutableHashTableV2Attr
- func MutableHashTableV2UseNodeNameSharing(value bool) MutableHashTableV2Attr
type MutexV2Attr
- func MutexV2Container(value string) MutexV2Attr
- func MutexV2SharedName(value string) MutexV2Attr
type NonDeterministicIntsAttr
- func NonDeterministicIntsDtype(value tf.DataType) NonDeterministicIntsAttr
type NonMaxSuppressionAttr
- func NonMaxSuppressionIouThreshold(value float32) NonMaxSuppressionAttr
type NonMaxSuppressionV4Attr
- func NonMaxSuppressionV4PadToMaxOutputSize(value bool) NonMaxSuppressionV4Attr
type NonMaxSuppressionV5Attr
- func NonMaxSuppressionV5PadToMaxOutputSize(value bool) NonMaxSuppressionV5Attr
type NotEqualAttr
- func NotEqualIncompatibleShapeError(value bool) NotEqualAttr
type NthElementAttr
- func NthElementReverse(value bool) NthElementAttr
type OneHotAttr
- func OneHotAxis(value int64) OneHotAttr
type OptimizeDatasetAttr
- func OptimizeDatasetOptimizationConfigs(value []string) OptimizeDatasetAttr
type OptimizeDatasetV2Attr
- func OptimizeDatasetV2OptimizationConfigs(value []string) OptimizeDatasetV2Attr
type OptionsDatasetAttr
- func OptionsDatasetMetadata(value string) OptionsDatasetAttr
type OrderedMapClearAttr
- func OrderedMapClearCapacity(value int64) OrderedMapClearAttr
- func OrderedMapClearContainer(value string) OrderedMapClearAttr
- func OrderedMapClearMemoryLimit(value int64) OrderedMapClearAttr
- func OrderedMapClearSharedName(value string) OrderedMapClearAttr
type OrderedMapIncompleteSizeAttr
- func OrderedMapIncompleteSizeCapacity(value int64) OrderedMapIncompleteSizeAttr
- func OrderedMapIncompleteSizeContainer(value string) OrderedMapIncompleteSizeAttr
- func OrderedMapIncompleteSizeMemoryLimit(value int64) OrderedMapIncompleteSizeAttr
- func OrderedMapIncompleteSizeSharedName(value string) OrderedMapIncompleteSizeAttr
type OrderedMapPeekAttr
- func OrderedMapPeekCapacity(value int64) OrderedMapPeekAttr
- func OrderedMapPeekContainer(value string) OrderedMapPeekAttr
- func OrderedMapPeekMemoryLimit(value int64) OrderedMapPeekAttr
- func OrderedMapPeekSharedName(value string) OrderedMapPeekAttr
type OrderedMapSizeAttr
- func OrderedMapSizeCapacity(value int64) OrderedMapSizeAttr
- func OrderedMapSizeContainer(value string) OrderedMapSizeAttr
- func OrderedMapSizeMemoryLimit(value int64) OrderedMapSizeAttr
- func OrderedMapSizeSharedName(value string) OrderedMapSizeAttr
type OrderedMapStageAttr
- func OrderedMapStageCapacity(value int64) OrderedMapStageAttr
- func OrderedMapStageContainer(value string) OrderedMapStageAttr
- func OrderedMapStageMemoryLimit(value int64) OrderedMapStageAttr
- func OrderedMapStageSharedName(value string) OrderedMapStageAttr
type OrderedMapUnstageAttr
- func OrderedMapUnstageCapacity(value int64) OrderedMapUnstageAttr
- func OrderedMapUnstageContainer(value string) OrderedMapUnstageAttr
- func OrderedMapUnstageMemoryLimit(value int64) OrderedMapUnstageAttr
- func OrderedMapUnstageSharedName(value string) OrderedMapUnstageAttr
type OrderedMapUnstageNoKeyAttr
- func OrderedMapUnstageNoKeyCapacity(value int64) OrderedMapUnstageNoKeyAttr
- func OrderedMapUnstageNoKeyContainer(value string) OrderedMapUnstageNoKeyAttr
- func OrderedMapUnstageNoKeyMemoryLimit(value int64) OrderedMapUnstageNoKeyAttr
- func OrderedMapUnstageNoKeySharedName(value string) OrderedMapUnstageNoKeyAttr
type OutfeedDequeueAttr
- func OutfeedDequeueDeviceOrdinal(value int64) OutfeedDequeueAttr
type OutfeedDequeueTupleAttr
- func OutfeedDequeueTupleDeviceOrdinal(value int64) OutfeedDequeueTupleAttr
type PackAttr
- func PackAxis(value int64) PackAttr
type PaddedBatchDatasetAttr
- func PaddedBatchDatasetMetadata(value string) PaddedBatchDatasetAttr
type PaddedBatchDatasetV2Attr
- func PaddedBatchDatasetV2Metadata(value string) PaddedBatchDatasetV2Attr
- func PaddedBatchDatasetV2ParallelCopy(value bool) PaddedBatchDatasetV2Attr
type PaddingFIFOQueueV2Attr
- func PaddingFIFOQueueV2Capacity(value int64) PaddingFIFOQueueV2Attr
- func PaddingFIFOQueueV2Container(value string) PaddingFIFOQueueV2Attr
- func PaddingFIFOQueueV2Shapes(value []tf.Shape) PaddingFIFOQueueV2Attr
- func PaddingFIFOQueueV2SharedName(value string) PaddingFIFOQueueV2Attr
type ParameterizedTruncatedNormalAttr
- func ParameterizedTruncatedNormalSeed(value int64) ParameterizedTruncatedNormalAttr
- func ParameterizedTruncatedNormalSeed2(value int64) ParameterizedTruncatedNormalAttr
type ParseExampleDatasetAttr
- func ParseExampleDatasetRaggedKeys(value []string) ParseExampleDatasetAttr
- func ParseExampleDatasetRaggedSplitTypes(value []tf.DataType) ParseExampleDatasetAttr
- func ParseExampleDatasetRaggedValueTypes(value []tf.DataType) ParseExampleDatasetAttr
- func ParseExampleDatasetSloppy(value bool) ParseExampleDatasetAttr
type ParseExampleDatasetV2Attr
- func ParseExampleDatasetV2Deterministic(value string) ParseExampleDatasetV2Attr
- func ParseExampleDatasetV2RaggedKeys(value []string) ParseExampleDatasetV2Attr
- func ParseExampleDatasetV2RaggedSplitTypes(value []tf.DataType) ParseExampleDatasetV2Attr
- func ParseExampleDatasetV2RaggedValueTypes(value []tf.DataType) ParseExampleDatasetV2Attr
type ParseSequenceExampleAttr
- func ParseSequenceExampleContextDenseShapes(value []tf.Shape) ParseSequenceExampleAttr
- func ParseSequenceExampleContextSparseTypes(value []tf.DataType) ParseSequenceExampleAttr
- func ParseSequenceExampleFeatureListDenseShapes(value []tf.Shape) ParseSequenceExampleAttr
- func ParseSequenceExampleFeatureListDenseTypes(value []tf.DataType) ParseSequenceExampleAttr
- func ParseSequenceExampleFeatureListSparseTypes(value []tf.DataType) ParseSequenceExampleAttr
- func ParseSequenceExampleNcontextDense(value int64) ParseSequenceExampleAttr
- func ParseSequenceExampleNcontextSparse(value int64) ParseSequenceExampleAttr
- func ParseSequenceExampleNfeatureListDense(value int64) ParseSequenceExampleAttr
- func ParseSequenceExampleNfeatureListSparse(value int64) ParseSequenceExampleAttr
type ParseSequenceExampleV2Attr
- func ParseSequenceExampleV2ContextDenseShapes(value []tf.Shape) ParseSequenceExampleV2Attr
- func ParseSequenceExampleV2ContextRaggedSplitTypes(value []tf.DataType) ParseSequenceExampleV2Attr
- func ParseSequenceExampleV2ContextRaggedValueTypes(value []tf.DataType) ParseSequenceExampleV2Attr
- func ParseSequenceExampleV2ContextSparseTypes(value []tf.DataType) ParseSequenceExampleV2Attr
- func ParseSequenceExampleV2FeatureListDenseShapes(value []tf.Shape) ParseSequenceExampleV2Attr
- func ParseSequenceExampleV2FeatureListDenseTypes(value []tf.DataType) ParseSequenceExampleV2Attr
- func ParseSequenceExampleV2FeatureListRaggedSplitTypes(value []tf.DataType) ParseSequenceExampleV2Attr
- func ParseSequenceExampleV2FeatureListRaggedValueTypes(value []tf.DataType) ParseSequenceExampleV2Attr
- func ParseSequenceExampleV2FeatureListSparseTypes(value []tf.DataType) ParseSequenceExampleV2Attr
- func ParseSequenceExampleV2NcontextSparse(value int64) ParseSequenceExampleV2Attr
- func ParseSequenceExampleV2NfeatureListDense(value int64) ParseSequenceExampleV2Attr
- func ParseSequenceExampleV2NfeatureListSparse(value int64) ParseSequenceExampleV2Attr
type ParseSingleSequenceExampleAttr
- func ParseSingleSequenceExampleContextDenseShapes(value []tf.Shape) ParseSingleSequenceExampleAttr
- func ParseSingleSequenceExampleContextSparseTypes(value []tf.DataType) ParseSingleSequenceExampleAttr
- func ParseSingleSequenceExampleFeatureListDenseShapes(value []tf.Shape) ParseSingleSequenceExampleAttr
- func ParseSingleSequenceExampleFeatureListDenseTypes(value []tf.DataType) ParseSingleSequenceExampleAttr
- func ParseSingleSequenceExampleFeatureListSparseTypes(value []tf.DataType) ParseSingleSequenceExampleAttr
type PlaceholderAttr
- func PlaceholderShape(value tf.Shape) PlaceholderAttr
type PrefetchDatasetAttr
- func PrefetchDatasetBufferSizeMin(value int64) PrefetchDatasetAttr
- func PrefetchDatasetLegacyAutotune(value bool) PrefetchDatasetAttr
- func PrefetchDatasetMetadata(value string) PrefetchDatasetAttr
- func PrefetchDatasetSlackPeriod(value int64) PrefetchDatasetAttr
type PrelinearizeAttr
- func PrelinearizeLayout(value []int64) PrelinearizeAttr
- func PrelinearizeShape(value tf.Shape) PrelinearizeAttr
type PrelinearizeTupleAttr
- func PrelinearizeTupleLayouts(value []int64) PrelinearizeTupleAttr
type PreventGradientAttr
- func PreventGradientMessage(value string) PreventGradientAttr
type PrintAttr
- func PrintFirstN(value int64) PrintAttr
- func PrintMessage(value string) PrintAttr
- func PrintSummarize(value int64) PrintAttr
type PrintV2Attr
- func PrintV2End(value string) PrintV2Attr
- func PrintV2OutputStream(value string) PrintV2Attr
type PriorityQueueV2Attr
- func PriorityQueueV2Capacity(value int64) PriorityQueueV2Attr
- func PriorityQueueV2ComponentTypes(value []tf.DataType) PriorityQueueV2Attr
- func PriorityQueueV2Container(value string) PriorityQueueV2Attr
- func PriorityQueueV2SharedName(value string) PriorityQueueV2Attr
type ProdAttr
- func ProdKeepDims(value bool) ProdAttr
type QrAttr
- func QrFullMatrices(value bool) QrAttr
type QuantizeAndDequantizeAttr
- func QuantizeAndDequantizeInputMax(value float32) QuantizeAndDequantizeAttr
- func QuantizeAndDequantizeInputMin(value float32) QuantizeAndDequantizeAttr
- func QuantizeAndDequantizeNumBits(value int64) QuantizeAndDequantizeAttr
- func QuantizeAndDequantizeRangeGiven(value bool) QuantizeAndDequantizeAttr
- func QuantizeAndDequantizeSignedInput(value bool) QuantizeAndDequantizeAttr
type QuantizeAndDequantizeV2Attr
- func QuantizeAndDequantizeV2Axis(value int64) QuantizeAndDequantizeV2Attr
- func QuantizeAndDequantizeV2NarrowRange(value bool) QuantizeAndDequantizeV2Attr
- func QuantizeAndDequantizeV2NumBits(value int64) QuantizeAndDequantizeV2Attr
- func QuantizeAndDequantizeV2RangeGiven(value bool) QuantizeAndDequantizeV2Attr
- func QuantizeAndDequantizeV2RoundMode(value string) QuantizeAndDequantizeV2Attr
- func QuantizeAndDequantizeV2SignedInput(value bool) QuantizeAndDequantizeV2Attr
type QuantizeAndDequantizeV3Attr
- func QuantizeAndDequantizeV3Axis(value int64) QuantizeAndDequantizeV3Attr
- func QuantizeAndDequantizeV3NarrowRange(value bool) QuantizeAndDequantizeV3Attr
- func QuantizeAndDequantizeV3RangeGiven(value bool) QuantizeAndDequantizeV3Attr
- func QuantizeAndDequantizeV3SignedInput(value bool) QuantizeAndDequantizeV3Attr
type QuantizeAndDequantizeV4Attr
- func QuantizeAndDequantizeV4Axis(value int64) QuantizeAndDequantizeV4Attr
- func QuantizeAndDequantizeV4NarrowRange(value bool) QuantizeAndDequantizeV4Attr
- func QuantizeAndDequantizeV4NumBits(value int64) QuantizeAndDequantizeV4Attr
- func QuantizeAndDequantizeV4RangeGiven(value bool) QuantizeAndDequantizeV4Attr
- func QuantizeAndDequantizeV4RoundMode(value string) QuantizeAndDequantizeV4Attr
- func QuantizeAndDequantizeV4SignedInput(value bool) QuantizeAndDequantizeV4Attr
type QuantizeAndDequantizeV4GradAttr
- func QuantizeAndDequantizeV4GradAxis(value int64) QuantizeAndDequantizeV4GradAttr
type QuantizeV2Attr
- func QuantizeV2Axis(value int64) QuantizeV2Attr
- func QuantizeV2EnsureMinimumRange(value float32) QuantizeV2Attr
- func QuantizeV2Mode(value string) QuantizeV2Attr
- func QuantizeV2NarrowRange(value bool) QuantizeV2Attr
- func QuantizeV2RoundMode(value string) QuantizeV2Attr
type QuantizedAddAttr
- func QuantizedAddToutput(value tf.DataType) QuantizedAddAttr
type QuantizedConv2DAttr
- func QuantizedConv2DDilations(value []int64) QuantizedConv2DAttr
- func QuantizedConv2DOutType(value tf.DataType) QuantizedConv2DAttr
type QuantizedConv2DPerChannelAttr
- func QuantizedConv2DPerChannelDilations(value []int64) QuantizedConv2DPerChannelAttr
- func QuantizedConv2DPerChannelOutType(value tf.DataType) QuantizedConv2DPerChannelAttr
type QuantizedDepthwiseConv2DAttr
- func QuantizedDepthwiseConv2DDilations(value []int64) QuantizedDepthwiseConv2DAttr
- func QuantizedDepthwiseConv2DOutType(value tf.DataType) QuantizedDepthwiseConv2DAttr
type QuantizedDepthwiseConv2DWithBiasAndReluAndRequantizeAttr
- func QuantizedDepthwiseConv2DWithBiasAndReluAndRequantizeDilations(value []int64) QuantizedDepthwiseConv2DWithBiasAndReluAndRequantizeAttr
- func QuantizedDepthwiseConv2DWithBiasAndReluAndRequantizeOutType(value tf.DataType) QuantizedDepthwiseConv2DWithBiasAndReluAndRequantizeAttr
- func QuantizedDepthwiseConv2DWithBiasAndReluAndRequantizePaddingList(value []int64) QuantizedDepthwiseConv2DWithBiasAndReluAndRequantizeAttr
type QuantizedDepthwiseConv2DWithBiasAndReluAttr
- func QuantizedDepthwiseConv2DWithBiasAndReluDilations(value []int64) QuantizedDepthwiseConv2DWithBiasAndReluAttr
- func QuantizedDepthwiseConv2DWithBiasAndReluOutType(value tf.DataType) QuantizedDepthwiseConv2DWithBiasAndReluAttr
- func QuantizedDepthwiseConv2DWithBiasAndReluPaddingList(value []int64) QuantizedDepthwiseConv2DWithBiasAndReluAttr
type QuantizedDepthwiseConv2DWithBiasAttr
- func QuantizedDepthwiseConv2DWithBiasDilations(value []int64) QuantizedDepthwiseConv2DWithBiasAttr
- func QuantizedDepthwiseConv2DWithBiasOutType(value tf.DataType) QuantizedDepthwiseConv2DWithBiasAttr
type QuantizedInstanceNormAttr
- func QuantizedInstanceNormGivenYMax(value float32) QuantizedInstanceNormAttr
- func QuantizedInstanceNormGivenYMin(value float32) QuantizedInstanceNormAttr
- func QuantizedInstanceNormMinSeparation(value float32) QuantizedInstanceNormAttr
- func QuantizedInstanceNormOutputRangeGiven(value bool) QuantizedInstanceNormAttr
- func QuantizedInstanceNormVarianceEpsilon(value float32) QuantizedInstanceNormAttr
type QuantizedMatMulAttr
- func QuantizedMatMulTactivation(value tf.DataType) QuantizedMatMulAttr
- func QuantizedMatMulToutput(value tf.DataType) QuantizedMatMulAttr
- func QuantizedMatMulTransposeA(value bool) QuantizedMatMulAttr
- func QuantizedMatMulTransposeB(value bool) QuantizedMatMulAttr
type QuantizedMatMulWithBiasAndReluAndRequantizeAttr
- func QuantizedMatMulWithBiasAndReluAndRequantizeInputQuantMode(value string) QuantizedMatMulWithBiasAndReluAndRequantizeAttr
- func QuantizedMatMulWithBiasAndReluAndRequantizeToutput(value tf.DataType) QuantizedMatMulWithBiasAndReluAndRequantizeAttr
- func QuantizedMatMulWithBiasAndReluAndRequantizeTransposeA(value bool) QuantizedMatMulWithBiasAndReluAndRequantizeAttr
- func QuantizedMatMulWithBiasAndReluAndRequantizeTransposeB(value bool) QuantizedMatMulWithBiasAndReluAndRequantizeAttr
type QuantizedMatMulWithBiasAndReluAttr
- func QuantizedMatMulWithBiasAndReluInputQuantMode(value string) QuantizedMatMulWithBiasAndReluAttr
- func QuantizedMatMulWithBiasAndReluToutput(value tf.DataType) QuantizedMatMulWithBiasAndReluAttr
- func QuantizedMatMulWithBiasAndReluTransposeA(value bool) QuantizedMatMulWithBiasAndReluAttr
- func QuantizedMatMulWithBiasAndReluTransposeB(value bool) QuantizedMatMulWithBiasAndReluAttr
type QuantizedMatMulWithBiasAttr
- func QuantizedMatMulWithBiasInputQuantMode(value string) QuantizedMatMulWithBiasAttr
- func QuantizedMatMulWithBiasToutput(value tf.DataType) QuantizedMatMulWithBiasAttr
- func QuantizedMatMulWithBiasTransposeA(value bool) QuantizedMatMulWithBiasAttr
- func QuantizedMatMulWithBiasTransposeB(value bool) QuantizedMatMulWithBiasAttr
type QuantizedMulAttr
- func QuantizedMulToutput(value tf.DataType) QuantizedMulAttr
type QuantizedRelu6Attr
- func QuantizedRelu6OutType(value tf.DataType) QuantizedRelu6Attr
type QuantizedReluAttr
- func QuantizedReluOutType(value tf.DataType) QuantizedReluAttr
type QuantizedReluXAttr
- func QuantizedReluXOutType(value tf.DataType) QuantizedReluXAttr
type QuantizedResizeBilinearAttr
- func QuantizedResizeBilinearAlignCorners(value bool) QuantizedResizeBilinearAttr
- func QuantizedResizeBilinearHalfPixelCenters(value bool) QuantizedResizeBilinearAttr
type QueueCloseV2Attr
- func QueueCloseV2CancelPendingEnqueues(value bool) QueueCloseV2Attr
type QueueDequeueManyV2Attr
- func QueueDequeueManyV2TimeoutMs(value int64) QueueDequeueManyV2Attr
type QueueDequeueUpToV2Attr
- func QueueDequeueUpToV2TimeoutMs(value int64) QueueDequeueUpToV2Attr
type QueueDequeueV2Attr
- func QueueDequeueV2TimeoutMs(value int64) QueueDequeueV2Attr
type QueueEnqueueManyV2Attr
- func QueueEnqueueManyV2TimeoutMs(value int64) QueueEnqueueManyV2Attr
type QueueEnqueueV2Attr
- func QueueEnqueueV2TimeoutMs(value int64) QueueEnqueueV2Attr
type RFFT2DAttr
- func RFFT2DTcomplex(value tf.DataType) RFFT2DAttr
type RFFT3DAttr
- func RFFT3DTcomplex(value tf.DataType) RFFT3DAttr
type RFFTAttr
- func RFFTTcomplex(value tf.DataType) RFFTAttr
type RFFTNDAttr
- func RFFTNDTcomplex(value tf.DataType) RFFTNDAttr
type RaggedBincountAttr
- func RaggedBincountBinaryOutput(value bool) RaggedBincountAttr
type RaggedCountSparseOutputAttr
- func RaggedCountSparseOutputMaxlength(value int64) RaggedCountSparseOutputAttr
- func RaggedCountSparseOutputMinlength(value int64) RaggedCountSparseOutputAttr
type RaggedRangeAttr
- func RaggedRangeTsplits(value tf.DataType) RaggedRangeAttr
type RaggedTensorFromVariantAttr
- func RaggedTensorFromVariantTsplits(value tf.DataType) RaggedTensorFromVariantAttr
type RandomCropAttr
- func RandomCropSeed(value int64) RandomCropAttr
- func RandomCropSeed2(value int64) RandomCropAttr
type RandomDatasetAttr
- func RandomDatasetMetadata(value string) RandomDatasetAttr
type RandomDatasetV2Attr
- func RandomDatasetV2Metadata(value string) RandomDatasetV2Attr
- func RandomDatasetV2RerandomizeEachIteration(value bool) RandomDatasetV2Attr
type RandomGammaAttr
- func RandomGammaSeed(value int64) RandomGammaAttr
- func RandomGammaSeed2(value int64) RandomGammaAttr
type RandomIndexShuffleAttr
- func RandomIndexShuffleRounds(value int64) RandomIndexShuffleAttr
type RandomPoissonAttr
- func RandomPoissonSeed(value int64) RandomPoissonAttr
- func RandomPoissonSeed2(value int64) RandomPoissonAttr
type RandomPoissonV2Attr
- func RandomPoissonV2Dtype(value tf.DataType) RandomPoissonV2Attr
- func RandomPoissonV2Seed(value int64) RandomPoissonV2Attr
- func RandomPoissonV2Seed2(value int64) RandomPoissonV2Attr
type RandomShuffleAttr
- func RandomShuffleSeed(value int64) RandomShuffleAttr
- func RandomShuffleSeed2(value int64) RandomShuffleAttr
type RandomShuffleQueueV2Attr
- func RandomShuffleQueueV2Capacity(value int64) RandomShuffleQueueV2Attr
- func RandomShuffleQueueV2Container(value string) RandomShuffleQueueV2Attr
- func RandomShuffleQueueV2MinAfterDequeue(value int64) RandomShuffleQueueV2Attr
- func RandomShuffleQueueV2Seed(value int64) RandomShuffleQueueV2Attr
- func RandomShuffleQueueV2Seed2(value int64) RandomShuffleQueueV2Attr
- func RandomShuffleQueueV2Shapes(value []tf.Shape) RandomShuffleQueueV2Attr
- func RandomShuffleQueueV2SharedName(value string) RandomShuffleQueueV2Attr
type RandomStandardNormalAttr
- func RandomStandardNormalSeed(value int64) RandomStandardNormalAttr
- func RandomStandardNormalSeed2(value int64) RandomStandardNormalAttr
type RandomUniformAttr
- func RandomUniformSeed(value int64) RandomUniformAttr
- func RandomUniformSeed2(value int64) RandomUniformAttr
type RandomUniformIntAttr
- func RandomUniformIntSeed(value int64) RandomUniformIntAttr
- func RandomUniformIntSeed2(value int64) RandomUniformIntAttr
type RangeDatasetAttr
- func RangeDatasetMetadata(value string) RangeDatasetAttr
- func RangeDatasetReplicateOnSplit(value bool) RangeDatasetAttr
type ReadVariableXlaSplitNDAttr
- func ReadVariableXlaSplitNDPaddings(value []int64) ReadVariableXlaSplitNDAttr
type RealAttr
- func RealTout(value tf.DataType) RealAttr
type RebatchDatasetAttr
- func RebatchDatasetUseFallback(value bool) RebatchDatasetAttr
type RecordInputAttr
- func RecordInputBatchSize(value int64) RecordInputAttr
- func RecordInputCompressionType(value string) RecordInputAttr
- func RecordInputFileBufferSize(value int64) RecordInputAttr
- func RecordInputFileParallelism(value int64) RecordInputAttr
- func RecordInputFileRandomSeed(value int64) RecordInputAttr
- func RecordInputFileShuffleShiftRatio(value float32) RecordInputAttr
type RecvAttr
- func RecvClientTerminated(value bool) RecvAttr
type ReduceJoinAttr
- func ReduceJoinKeepDims(value bool) ReduceJoinAttr
- func ReduceJoinSeparator(value string) ReduceJoinAttr
type RegexReplaceAttr
- func RegexReplaceReplaceGlobal(value bool) RegexReplaceAttr
type RegisterDatasetAttr
- func RegisterDatasetElementSpec(value string) RegisterDatasetAttr
- func RegisterDatasetMetadata(value string) RegisterDatasetAttr
type RegisterDatasetV2Attr
- func RegisterDatasetV2ElementSpec(value string) RegisterDatasetV2Attr
- func RegisterDatasetV2Metadata(value string) RegisterDatasetV2Attr
- func RegisterDatasetV2RequestedDatasetId(value string) RegisterDatasetV2Attr
type RepeatDatasetAttr
- func RepeatDatasetMetadata(value string) RepeatDatasetAttr
type RequantizePerChannelAttr
- func RequantizePerChannelOutType(value tf.DataType) RequantizePerChannelAttr
type ResizeAreaAttr
- func ResizeAreaAlignCorners(value bool) ResizeAreaAttr
type ResizeBicubicAttr
- func ResizeBicubicAlignCorners(value bool) ResizeBicubicAttr
- func ResizeBicubicHalfPixelCenters(value bool) ResizeBicubicAttr
type ResizeBicubicGradAttr
- func ResizeBicubicGradAlignCorners(value bool) ResizeBicubicGradAttr
- func ResizeBicubicGradHalfPixelCenters(value bool) ResizeBicubicGradAttr
type ResizeBilinearAttr
- func ResizeBilinearAlignCorners(value bool) ResizeBilinearAttr
- func ResizeBilinearHalfPixelCenters(value bool) ResizeBilinearAttr
type ResizeBilinearGradAttr
- func ResizeBilinearGradAlignCorners(value bool) ResizeBilinearGradAttr
- func ResizeBilinearGradHalfPixelCenters(value bool) ResizeBilinearGradAttr
type ResizeNearestNeighborAttr
- func ResizeNearestNeighborAlignCorners(value bool) ResizeNearestNeighborAttr
- func ResizeNearestNeighborHalfPixelCenters(value bool) ResizeNearestNeighborAttr
type ResizeNearestNeighborGradAttr
- func ResizeNearestNeighborGradAlignCorners(value bool) ResizeNearestNeighborGradAttr
- func ResizeNearestNeighborGradHalfPixelCenters(value bool) ResizeNearestNeighborGradAttr
type ResourceApplyAdaMaxAttr
- func ResourceApplyAdaMaxUseLocking(value bool) ResourceApplyAdaMaxAttr
type ResourceApplyAdadeltaAttr
- func ResourceApplyAdadeltaUseLocking(value bool) ResourceApplyAdadeltaAttr
type ResourceApplyAdagradAttr
- func ResourceApplyAdagradUpdateSlots(value bool) ResourceApplyAdagradAttr
- func ResourceApplyAdagradUseLocking(value bool) ResourceApplyAdagradAttr
type ResourceApplyAdagradDAAttr
- func ResourceApplyAdagradDAUseLocking(value bool) ResourceApplyAdagradDAAttr
type ResourceApplyAdagradV2Attr
- func ResourceApplyAdagradV2UpdateSlots(value bool) ResourceApplyAdagradV2Attr
- func ResourceApplyAdagradV2UseLocking(value bool) ResourceApplyAdagradV2Attr
type ResourceApplyAdamAttr
- func ResourceApplyAdamUseLocking(value bool) ResourceApplyAdamAttr
- func ResourceApplyAdamUseNesterov(value bool) ResourceApplyAdamAttr
type ResourceApplyAdamWithAmsgradAttr
- func ResourceApplyAdamWithAmsgradUseLocking(value bool) ResourceApplyAdamWithAmsgradAttr
type ResourceApplyAddSignAttr
- func ResourceApplyAddSignUseLocking(value bool) ResourceApplyAddSignAttr
type ResourceApplyCenteredRMSPropAttr
- func ResourceApplyCenteredRMSPropUseLocking(value bool) ResourceApplyCenteredRMSPropAttr
type ResourceApplyFtrlAttr
- func ResourceApplyFtrlMultiplyLinearByLr(value bool) ResourceApplyFtrlAttr
- func ResourceApplyFtrlUseLocking(value bool) ResourceApplyFtrlAttr
type ResourceApplyFtrlV2Attr
- func ResourceApplyFtrlV2MultiplyLinearByLr(value bool) ResourceApplyFtrlV2Attr
- func ResourceApplyFtrlV2UseLocking(value bool) ResourceApplyFtrlV2Attr
type ResourceApplyGradientDescentAttr
- func ResourceApplyGradientDescentUseLocking(value bool) ResourceApplyGradientDescentAttr
type ResourceApplyKerasMomentumAttr
- func ResourceApplyKerasMomentumUseLocking(value bool) ResourceApplyKerasMomentumAttr
- func ResourceApplyKerasMomentumUseNesterov(value bool) ResourceApplyKerasMomentumAttr
type ResourceApplyMomentumAttr
- func ResourceApplyMomentumUseLocking(value bool) ResourceApplyMomentumAttr
- func ResourceApplyMomentumUseNesterov(value bool) ResourceApplyMomentumAttr
type ResourceApplyPowerSignAttr
- func ResourceApplyPowerSignUseLocking(value bool) ResourceApplyPowerSignAttr
type ResourceApplyProximalAdagradAttr
- func ResourceApplyProximalAdagradUseLocking(value bool) ResourceApplyProximalAdagradAttr
type ResourceApplyProximalGradientDescentAttr
- func ResourceApplyProximalGradientDescentUseLocking(value bool) ResourceApplyProximalGradientDescentAttr
type ResourceApplyRMSPropAttr
- func ResourceApplyRMSPropUseLocking(value bool) ResourceApplyRMSPropAttr
type ResourceConditionalAccumulatorAttr
- func ResourceConditionalAccumulatorContainer(value string) ResourceConditionalAccumulatorAttr
- func ResourceConditionalAccumulatorReductionType(value string) ResourceConditionalAccumulatorAttr
- func ResourceConditionalAccumulatorSharedName(value string) ResourceConditionalAccumulatorAttr
type ResourceGatherAttr
- func ResourceGatherBatchDims(value int64) ResourceGatherAttr
- func ResourceGatherValidateIndices(value bool) ResourceGatherAttr
type ResourceScatterNdAddAttr
- func ResourceScatterNdAddUseLocking(value bool) ResourceScatterNdAddAttr
type ResourceScatterNdSubAttr
- func ResourceScatterNdSubUseLocking(value bool) ResourceScatterNdSubAttr
type ResourceScatterNdUpdateAttr
- func ResourceScatterNdUpdateUseLocking(value bool) ResourceScatterNdUpdateAttr
type ResourceSparseApplyAdadeltaAttr
- func ResourceSparseApplyAdadeltaUseLocking(value bool) ResourceSparseApplyAdadeltaAttr
type ResourceSparseApplyAdagradAttr
- func ResourceSparseApplyAdagradUpdateSlots(value bool) ResourceSparseApplyAdagradAttr
- func ResourceSparseApplyAdagradUseLocking(value bool) ResourceSparseApplyAdagradAttr
type ResourceSparseApplyAdagradDAAttr
- func ResourceSparseApplyAdagradDAUseLocking(value bool) ResourceSparseApplyAdagradDAAttr
type ResourceSparseApplyAdagradV2Attr
- func ResourceSparseApplyAdagradV2UpdateSlots(value bool) ResourceSparseApplyAdagradV2Attr
- func ResourceSparseApplyAdagradV2UseLocking(value bool) ResourceSparseApplyAdagradV2Attr
type ResourceSparseApplyCenteredRMSPropAttr
- func ResourceSparseApplyCenteredRMSPropUseLocking(value bool) ResourceSparseApplyCenteredRMSPropAttr
type ResourceSparseApplyFtrlAttr
- func ResourceSparseApplyFtrlMultiplyLinearByLr(value bool) ResourceSparseApplyFtrlAttr
- func ResourceSparseApplyFtrlUseLocking(value bool) ResourceSparseApplyFtrlAttr
type ResourceSparseApplyFtrlV2Attr
- func ResourceSparseApplyFtrlV2MultiplyLinearByLr(value bool) ResourceSparseApplyFtrlV2Attr
- func ResourceSparseApplyFtrlV2UseLocking(value bool) ResourceSparseApplyFtrlV2Attr
type ResourceSparseApplyKerasMomentumAttr
- func ResourceSparseApplyKerasMomentumUseLocking(value bool) ResourceSparseApplyKerasMomentumAttr
- func ResourceSparseApplyKerasMomentumUseNesterov(value bool) ResourceSparseApplyKerasMomentumAttr
type ResourceSparseApplyMomentumAttr
- func ResourceSparseApplyMomentumUseLocking(value bool) ResourceSparseApplyMomentumAttr
- func ResourceSparseApplyMomentumUseNesterov(value bool) ResourceSparseApplyMomentumAttr
type ResourceSparseApplyProximalAdagradAttr
- func ResourceSparseApplyProximalAdagradUseLocking(value bool) ResourceSparseApplyProximalAdagradAttr
type ResourceSparseApplyProximalGradientDescentAttr
- func ResourceSparseApplyProximalGradientDescentUseLocking(value bool) ResourceSparseApplyProximalGradientDescentAttr
type ResourceSparseApplyRMSPropAttr
- func ResourceSparseApplyRMSPropUseLocking(value bool) ResourceSparseApplyRMSPropAttr
type ResourceStridedSliceAssignAttr
- func ResourceStridedSliceAssignBeginMask(value int64) ResourceStridedSliceAssignAttr
- func ResourceStridedSliceAssignEllipsisMask(value int64) ResourceStridedSliceAssignAttr
- func ResourceStridedSliceAssignEndMask(value int64) ResourceStridedSliceAssignAttr
- func ResourceStridedSliceAssignNewAxisMask(value int64) ResourceStridedSliceAssignAttr
- func ResourceStridedSliceAssignShrinkAxisMask(value int64) ResourceStridedSliceAssignAttr
type RestoreAttr
- func RestorePreferredShard(value int64) RestoreAttr
type RestoreSliceAttr
- func RestoreSlicePreferredShard(value int64) RestoreSliceAttr
type RetrieveTPUEmbeddingADAMParametersAttr
- func RetrieveTPUEmbeddingADAMParametersConfig(value string) RetrieveTPUEmbeddingADAMParametersAttr
- func RetrieveTPUEmbeddingADAMParametersTableId(value int64) RetrieveTPUEmbeddingADAMParametersAttr
- func RetrieveTPUEmbeddingADAMParametersTableName(value string) RetrieveTPUEmbeddingADAMParametersAttr
type RetrieveTPUEmbeddingAdadeltaParametersAttr
- func RetrieveTPUEmbeddingAdadeltaParametersConfig(value string) RetrieveTPUEmbeddingAdadeltaParametersAttr
- func RetrieveTPUEmbeddingAdadeltaParametersTableId(value int64) RetrieveTPUEmbeddingAdadeltaParametersAttr
- func RetrieveTPUEmbeddingAdadeltaParametersTableName(value string) RetrieveTPUEmbeddingAdadeltaParametersAttr
type RetrieveTPUEmbeddingAdagradMomentumParametersAttr
- func RetrieveTPUEmbeddingAdagradMomentumParametersConfig(value string) RetrieveTPUEmbeddingAdagradMomentumParametersAttr
- func RetrieveTPUEmbeddingAdagradMomentumParametersTableId(value int64) RetrieveTPUEmbeddingAdagradMomentumParametersAttr
- func RetrieveTPUEmbeddingAdagradMomentumParametersTableName(value string) RetrieveTPUEmbeddingAdagradMomentumParametersAttr
type RetrieveTPUEmbeddingAdagradParametersAttr
- func RetrieveTPUEmbeddingAdagradParametersConfig(value string) RetrieveTPUEmbeddingAdagradParametersAttr
- func RetrieveTPUEmbeddingAdagradParametersTableId(value int64) RetrieveTPUEmbeddingAdagradParametersAttr
- func RetrieveTPUEmbeddingAdagradParametersTableName(value string) RetrieveTPUEmbeddingAdagradParametersAttr
type RetrieveTPUEmbeddingCenteredRMSPropParametersAttr
- func RetrieveTPUEmbeddingCenteredRMSPropParametersConfig(value string) RetrieveTPUEmbeddingCenteredRMSPropParametersAttr
- func RetrieveTPUEmbeddingCenteredRMSPropParametersTableId(value int64) RetrieveTPUEmbeddingCenteredRMSPropParametersAttr
- func RetrieveTPUEmbeddingCenteredRMSPropParametersTableName(value string) RetrieveTPUEmbeddingCenteredRMSPropParametersAttr
type RetrieveTPUEmbeddingFTRLParametersAttr
- func RetrieveTPUEmbeddingFTRLParametersConfig(value string) RetrieveTPUEmbeddingFTRLParametersAttr
- func RetrieveTPUEmbeddingFTRLParametersTableId(value int64) RetrieveTPUEmbeddingFTRLParametersAttr
- func RetrieveTPUEmbeddingFTRLParametersTableName(value string) RetrieveTPUEmbeddingFTRLParametersAttr
type RetrieveTPUEmbeddingFrequencyEstimatorParametersAttr
- func RetrieveTPUEmbeddingFrequencyEstimatorParametersConfig(value string) RetrieveTPUEmbeddingFrequencyEstimatorParametersAttr
- func RetrieveTPUEmbeddingFrequencyEstimatorParametersTableId(value int64) RetrieveTPUEmbeddingFrequencyEstimatorParametersAttr
- func RetrieveTPUEmbeddingFrequencyEstimatorParametersTableName(value string) RetrieveTPUEmbeddingFrequencyEstimatorParametersAttr
type RetrieveTPUEmbeddingMDLAdagradLightParametersAttr
- func RetrieveTPUEmbeddingMDLAdagradLightParametersConfig(value string) RetrieveTPUEmbeddingMDLAdagradLightParametersAttr
- func RetrieveTPUEmbeddingMDLAdagradLightParametersTableId(value int64) RetrieveTPUEmbeddingMDLAdagradLightParametersAttr
- func RetrieveTPUEmbeddingMDLAdagradLightParametersTableName(value string) RetrieveTPUEmbeddingMDLAdagradLightParametersAttr
type RetrieveTPUEmbeddingMomentumParametersAttr
- func RetrieveTPUEmbeddingMomentumParametersConfig(value string) RetrieveTPUEmbeddingMomentumParametersAttr
- func RetrieveTPUEmbeddingMomentumParametersTableId(value int64) RetrieveTPUEmbeddingMomentumParametersAttr
- func RetrieveTPUEmbeddingMomentumParametersTableName(value string) RetrieveTPUEmbeddingMomentumParametersAttr
type RetrieveTPUEmbeddingProximalAdagradParametersAttr
- func RetrieveTPUEmbeddingProximalAdagradParametersConfig(value string) RetrieveTPUEmbeddingProximalAdagradParametersAttr
- func RetrieveTPUEmbeddingProximalAdagradParametersTableId(value int64) RetrieveTPUEmbeddingProximalAdagradParametersAttr
- func RetrieveTPUEmbeddingProximalAdagradParametersTableName(value string) RetrieveTPUEmbeddingProximalAdagradParametersAttr
type RetrieveTPUEmbeddingRMSPropParametersAttr
- func RetrieveTPUEmbeddingRMSPropParametersConfig(value string) RetrieveTPUEmbeddingRMSPropParametersAttr
- func RetrieveTPUEmbeddingRMSPropParametersTableId(value int64) RetrieveTPUEmbeddingRMSPropParametersAttr
- func RetrieveTPUEmbeddingRMSPropParametersTableName(value string) RetrieveTPUEmbeddingRMSPropParametersAttr
type RetrieveTPUEmbeddingStochasticGradientDescentParametersAttr
- func RetrieveTPUEmbeddingStochasticGradientDescentParametersConfig(value string) RetrieveTPUEmbeddingStochasticGradientDescentParametersAttr
- func RetrieveTPUEmbeddingStochasticGradientDescentParametersTableId(value int64) RetrieveTPUEmbeddingStochasticGradientDescentParametersAttr
- func RetrieveTPUEmbeddingStochasticGradientDescentParametersTableName(value string) RetrieveTPUEmbeddingStochasticGradientDescentParametersAttr
type ReverseSequenceAttr
- func ReverseSequenceBatchDim(value int64) ReverseSequenceAttr
type SampleDistortedBoundingBoxAttr
- func SampleDistortedBoundingBoxAreaRange(value []float32) SampleDistortedBoundingBoxAttr
- func SampleDistortedBoundingBoxAspectRatioRange(value []float32) SampleDistortedBoundingBoxAttr
- func SampleDistortedBoundingBoxMaxAttempts(value int64) SampleDistortedBoundingBoxAttr
- func SampleDistortedBoundingBoxMinObjectCovered(value float32) SampleDistortedBoundingBoxAttr
- func SampleDistortedBoundingBoxSeed(value int64) SampleDistortedBoundingBoxAttr
- func SampleDistortedBoundingBoxSeed2(value int64) SampleDistortedBoundingBoxAttr
- func SampleDistortedBoundingBoxUseImageIfNoBoundingBoxes(value bool) SampleDistortedBoundingBoxAttr
type SampleDistortedBoundingBoxV2Attr
- func SampleDistortedBoundingBoxV2AreaRange(value []float32) SampleDistortedBoundingBoxV2Attr
- func SampleDistortedBoundingBoxV2AspectRatioRange(value []float32) SampleDistortedBoundingBoxV2Attr
- func SampleDistortedBoundingBoxV2MaxAttempts(value int64) SampleDistortedBoundingBoxV2Attr
- func SampleDistortedBoundingBoxV2Seed(value int64) SampleDistortedBoundingBoxV2Attr
- func SampleDistortedBoundingBoxV2Seed2(value int64) SampleDistortedBoundingBoxV2Attr
- func SampleDistortedBoundingBoxV2UseImageIfNoBoundingBoxes(value bool) SampleDistortedBoundingBoxV2Attr
type Scope
- func NewScope() *Scope
- func NewScopeWithGraph(g *tf.Graph) *Scope
- func (s *Scope) AddOperation(args tf.OpSpec) *tf.Operation
- func (s *Scope) Err() error
- func (s *Scope) Finalize() (*tf.Graph, error)
- func (s *Scope) SubScope(namespace string) *Scope
- func (s *Scope) UpdateErr(op string, err error)
- func (s *Scope) WithControlDependencies(ops ...*tf.Operation) *Scope
- func (s *Scope) WithDevice(device string) *Scope
type SdcaOptimizerAttr
- func SdcaOptimizerAdaptative(value bool) SdcaOptimizerAttr
type SdcaOptimizerV2Attr
- func SdcaOptimizerV2Adaptive(value bool) SdcaOptimizerV2Attr
type SelfAdjointEigV2Attr
- func SelfAdjointEigV2ComputeV(value bool) SelfAdjointEigV2Attr
type SendAttr
- func SendClientTerminated(value bool) SendAttr
type SerializeIteratorAttr
- func SerializeIteratorExternalStatePolicy(value int64) SerializeIteratorAttr
type SerializeManySparseAttr
- func SerializeManySparseOutType(value tf.DataType) SerializeManySparseAttr
type SerializeSparseAttr
- func SerializeSparseOutType(value tf.DataType) SerializeSparseAttr
type SetSizeAttr
- func SetSizeValidateIndices(value bool) SetSizeAttr
type ShapeAttr
- func ShapeOutType(value tf.DataType) ShapeAttr
type ShapeNAttr
- func ShapeNOutType(value tf.DataType) ShapeNAttr
type ShardDatasetAttr
- func ShardDatasetMetadata(value string) ShardDatasetAttr
- func ShardDatasetRequireNonEmpty(value bool) ShardDatasetAttr
type ShuffleAndRepeatDatasetAttr
- func ShuffleAndRepeatDatasetMetadata(value string) ShuffleAndRepeatDatasetAttr
- func ShuffleAndRepeatDatasetReshuffleEachIteration(value bool) ShuffleAndRepeatDatasetAttr
type ShuffleDatasetAttr
- func ShuffleDatasetMetadata(value string) ShuffleDatasetAttr
- func ShuffleDatasetReshuffleEachIteration(value bool) ShuffleDatasetAttr
type SizeAttr
- func SizeOutType(value tf.DataType) SizeAttr
type SkipDatasetAttr
- func SkipDatasetMetadata(value string) SkipDatasetAttr
type SkipgramAttr
- func SkipgramMinCount(value int64) SkipgramAttr
- func SkipgramSubsample(value float32) SkipgramAttr
- func SkipgramWindowSize(value int64) SkipgramAttr
type SlidingWindowDatasetAttr
- func SlidingWindowDatasetDropRemainder(value bool) SlidingWindowDatasetAttr
type SnapshotDatasetAttr
- func SnapshotDatasetCompression(value string) SnapshotDatasetAttr
- func SnapshotDatasetMode(value string) SnapshotDatasetAttr
- func SnapshotDatasetNumReaderThreads(value int64) SnapshotDatasetAttr
- func SnapshotDatasetNumWriterThreads(value int64) SnapshotDatasetAttr
- func SnapshotDatasetPendingSnapshotExpirySeconds(value int64) SnapshotDatasetAttr
- func SnapshotDatasetReaderBufferSize(value int64) SnapshotDatasetAttr
- func SnapshotDatasetReaderPathPrefix(value string) SnapshotDatasetAttr
- func SnapshotDatasetSeed(value int64) SnapshotDatasetAttr
- func SnapshotDatasetSeed2(value int64) SnapshotDatasetAttr
- func SnapshotDatasetShardSizeBytes(value int64) SnapshotDatasetAttr
- func SnapshotDatasetShuffleOnRead(value bool) SnapshotDatasetAttr
- func SnapshotDatasetSnapshotName(value string) SnapshotDatasetAttr
- func SnapshotDatasetWriterBufferSize(value int64) SnapshotDatasetAttr
- func SnapshotDatasetWriterPathPrefix(value string) SnapshotDatasetAttr
type SobolSampleAttr
- func SobolSampleDtype(value tf.DataType) SobolSampleAttr
type SpaceToDepthAttr
- func SpaceToDepthDataFormat(value string) SpaceToDepthAttr
type SparseBincountAttr
- func SparseBincountBinaryOutput(value bool) SparseBincountAttr
type SparseCountSparseOutputAttr
- func SparseCountSparseOutputMaxlength(value int64) SparseCountSparseOutputAttr
- func SparseCountSparseOutputMinlength(value int64) SparseCountSparseOutputAttr
type SparseMatMulAttr
- func SparseMatMulAIsSparse(value bool) SparseMatMulAttr
- func SparseMatMulBIsSparse(value bool) SparseMatMulAttr
- func SparseMatMulTransposeA(value bool) SparseMatMulAttr
- func SparseMatMulTransposeB(value bool) SparseMatMulAttr
type SparseMatrixMatMulAttr
- func SparseMatrixMatMulAdjointA(value bool) SparseMatrixMatMulAttr
- func SparseMatrixMatMulAdjointB(value bool) SparseMatrixMatMulAttr
- func SparseMatrixMatMulConjugateOutput(value bool) SparseMatrixMatMulAttr
- func SparseMatrixMatMulTransposeA(value bool) SparseMatrixMatMulAttr
- func SparseMatrixMatMulTransposeB(value bool) SparseMatrixMatMulAttr
- func SparseMatrixMatMulTransposeOutput(value bool) SparseMatrixMatMulAttr
type SparseMatrixSparseMatMulAttr
- func SparseMatrixSparseMatMulAdjointA(value bool) SparseMatrixSparseMatMulAttr
- func SparseMatrixSparseMatMulAdjointB(value bool) SparseMatrixSparseMatMulAttr
- func SparseMatrixSparseMatMulTransposeA(value bool) SparseMatrixSparseMatMulAttr
- func SparseMatrixSparseMatMulTransposeB(value bool) SparseMatrixSparseMatMulAttr
type SparseMatrixTransposeAttr
- func SparseMatrixTransposeConjugate(value bool) SparseMatrixTransposeAttr
type SparseReduceMaxAttr
- func SparseReduceMaxKeepDims(value bool) SparseReduceMaxAttr
type SparseReduceMaxSparseAttr
- func SparseReduceMaxSparseKeepDims(value bool) SparseReduceMaxSparseAttr
type SparseReduceSumAttr
- func SparseReduceSumKeepDims(value bool) SparseReduceSumAttr
type SparseReduceSumSparseAttr
- func SparseReduceSumSparseKeepDims(value bool) SparseReduceSumSparseAttr
type SparseSegmentMeanAttr
- func SparseSegmentMeanSparseGradient(value bool) SparseSegmentMeanAttr
type SparseSegmentMeanWithNumSegmentsAttr
- func SparseSegmentMeanWithNumSegmentsSparseGradient(value bool) SparseSegmentMeanWithNumSegmentsAttr
type SparseSegmentSqrtNAttr
- func SparseSegmentSqrtNSparseGradient(value bool) SparseSegmentSqrtNAttr
type SparseSegmentSqrtNWithNumSegmentsAttr
- func SparseSegmentSqrtNWithNumSegmentsSparseGradient(value bool) SparseSegmentSqrtNWithNumSegmentsAttr
type SparseSegmentSumAttr
- func SparseSegmentSumSparseGradient(value bool) SparseSegmentSumAttr
type SparseSegmentSumWithNumSegmentsAttr
- func SparseSegmentSumWithNumSegmentsSparseGradient(value bool) SparseSegmentSumWithNumSegmentsAttr
type SparseTensorDenseMatMulAttr
- func SparseTensorDenseMatMulAdjointA(value bool) SparseTensorDenseMatMulAttr
- func SparseTensorDenseMatMulAdjointB(value bool) SparseTensorDenseMatMulAttr
type SparseToDenseAttr
- func SparseToDenseValidateIndices(value bool) SparseToDenseAttr
type SparseToSparseSetOperationAttr
- func SparseToSparseSetOperationValidateIndices(value bool) SparseToSparseSetOperationAttr
type SplitDedupDataAttr
- func SplitDedupDataConfig(value string) SplitDedupDataAttr
type SqueezeAttr
- func SqueezeAxis(value []int64) SqueezeAttr
type StackPushV2Attr
- func StackPushV2SwapMemory(value bool) StackPushV2Attr
type StackV2Attr
- func StackV2StackName(value string) StackV2Attr
type StageAttr
- func StageCapacity(value int64) StageAttr
- func StageContainer(value string) StageAttr
- func StageMemoryLimit(value int64) StageAttr
- func StageSharedName(value string) StageAttr
type StageClearAttr
- func StageClearCapacity(value int64) StageClearAttr
- func StageClearContainer(value string) StageClearAttr
- func StageClearMemoryLimit(value int64) StageClearAttr
- func StageClearSharedName(value string) StageClearAttr
type StagePeekAttr
- func StagePeekCapacity(value int64) StagePeekAttr
- func StagePeekContainer(value string) StagePeekAttr
- func StagePeekMemoryLimit(value int64) StagePeekAttr
- func StagePeekSharedName(value string) StagePeekAttr
type StageSizeAttr
- func StageSizeCapacity(value int64) StageSizeAttr
- func StageSizeContainer(value string) StageSizeAttr
- func StageSizeMemoryLimit(value int64) StageSizeAttr
- func StageSizeSharedName(value string) StageSizeAttr
type StatefulStandardNormalAttr
- func StatefulStandardNormalDtype(value tf.DataType) StatefulStandardNormalAttr
type StatefulStandardNormalV2Attr
- func StatefulStandardNormalV2Dtype(value tf.DataType) StatefulStandardNormalV2Attr
type StatefulTruncatedNormalAttr
- func StatefulTruncatedNormalDtype(value tf.DataType) StatefulTruncatedNormalAttr
type StatefulUniformAttr
- func StatefulUniformDtype(value tf.DataType) StatefulUniformAttr
type StatefulUniformFullIntAttr
- func StatefulUniformFullIntDtype(value tf.DataType) StatefulUniformFullIntAttr
type StatelessMultinomialAttr
- func StatelessMultinomialOutputDtype(value tf.DataType) StatelessMultinomialAttr
type StatelessRandomBinomialAttr
- func StatelessRandomBinomialDtype(value tf.DataType) StatelessRandomBinomialAttr
type StatelessRandomNormalAttr
- func StatelessRandomNormalDtype(value tf.DataType) StatelessRandomNormalAttr
type StatelessRandomNormalV2Attr
- func StatelessRandomNormalV2Dtype(value tf.DataType) StatelessRandomNormalV2Attr
type StatelessRandomUniformAttr
- func StatelessRandomUniformDtype(value tf.DataType) StatelessRandomUniformAttr
type StatelessRandomUniformFullIntAttr
- func StatelessRandomUniformFullIntDtype(value tf.DataType) StatelessRandomUniformFullIntAttr
type StatelessRandomUniformFullIntV2Attr
- func StatelessRandomUniformFullIntV2Dtype(value tf.DataType) StatelessRandomUniformFullIntV2Attr
type StatelessRandomUniformV2Attr
- func StatelessRandomUniformV2Dtype(value tf.DataType) StatelessRandomUniformV2Attr
type StatelessSampleDistortedBoundingBoxAttr
- func StatelessSampleDistortedBoundingBoxAreaRange(value []float32) StatelessSampleDistortedBoundingBoxAttr
- func StatelessSampleDistortedBoundingBoxAspectRatioRange(value []float32) StatelessSampleDistortedBoundingBoxAttr
- func StatelessSampleDistortedBoundingBoxMaxAttempts(value int64) StatelessSampleDistortedBoundingBoxAttr
- func StatelessSampleDistortedBoundingBoxUseImageIfNoBoundingBoxes(value bool) StatelessSampleDistortedBoundingBoxAttr
type StatelessTruncatedNormalAttr
- func StatelessTruncatedNormalDtype(value tf.DataType) StatelessTruncatedNormalAttr
type StatelessTruncatedNormalV2Attr
- func StatelessTruncatedNormalV2Dtype(value tf.DataType) StatelessTruncatedNormalV2Attr
type StaticRegexReplaceAttr
- func StaticRegexReplaceReplaceGlobal(value bool) StaticRegexReplaceAttr
type StatsAggregatorHandleAttr
- func StatsAggregatorHandleContainer(value string) StatsAggregatorHandleAttr
- func StatsAggregatorHandleSharedName(value string) StatsAggregatorHandleAttr
type StridedSliceAttr
- func StridedSliceBeginMask(value int64) StridedSliceAttr
- func StridedSliceEllipsisMask(value int64) StridedSliceAttr
- func StridedSliceEndMask(value int64) StridedSliceAttr
- func StridedSliceNewAxisMask(value int64) StridedSliceAttr
- func StridedSliceShrinkAxisMask(value int64) StridedSliceAttr
type StridedSliceGradAttr
- func StridedSliceGradBeginMask(value int64) StridedSliceGradAttr
- func StridedSliceGradEllipsisMask(value int64) StridedSliceGradAttr
- func StridedSliceGradEndMask(value int64) StridedSliceGradAttr
- func StridedSliceGradNewAxisMask(value int64) StridedSliceGradAttr
- func StridedSliceGradShrinkAxisMask(value int64) StridedSliceGradAttr
type StringFormatAttr
- func StringFormatPlaceholder(value string) StringFormatAttr
- func StringFormatSummarize(value int64) StringFormatAttr
- func StringFormatTemplate(value string) StringFormatAttr
type StringJoinAttr
- func StringJoinSeparator(value string) StringJoinAttr
type StringLengthAttr
- func StringLengthUnit(value string) StringLengthAttr
type StringLowerAttr
- func StringLowerEncoding(value string) StringLowerAttr
type StringSplitAttr
- func StringSplitSkipEmpty(value bool) StringSplitAttr
type StringSplitV2Attr
- func StringSplitV2Maxsplit(value int64) StringSplitV2Attr
type StringToNumberAttr
- func StringToNumberOutType(value tf.DataType) StringToNumberAttr
type StringUpperAttr
- func StringUpperEncoding(value string) StringUpperAttr
type SubstrAttr
- func SubstrUnit(value string) SubstrAttr
type SumAttr
- func SumKeepDims(value bool) SumAttr
type SvdAttr
- func SvdComputeUv(value bool) SvdAttr
- func SvdFullMatrices(value bool) SvdAttr
type TFRecordDatasetAttr
- func TFRecordDatasetMetadata(value string) TFRecordDatasetAttr
type TFRecordDatasetV2Attr
- func TFRecordDatasetV2Metadata(value string) TFRecordDatasetV2Attr
type TFRecordReaderV2Attr
- func TFRecordReaderV2CompressionType(value string) TFRecordReaderV2Attr
- func TFRecordReaderV2Container(value string) TFRecordReaderV2Attr
- func TFRecordReaderV2SharedName(value string) TFRecordReaderV2Attr
type TPUPartitionedInputAttr
- func TPUPartitionedInputPartitionDim(value int64) TPUPartitionedInputAttr
type TPUPartitionedInputV2Attr
- func TPUPartitionedInputV2IsPacked(value bool) TPUPartitionedInputV2Attr
type TPUPartitionedOutputAttr
- func TPUPartitionedOutputPartitionDim(value int64) TPUPartitionedOutputAttr
type TPUReplicateMetadataAttr
- func TPUReplicateMetadataAllowSoftPlacement(value bool) TPUReplicateMetadataAttr
- func TPUReplicateMetadataComputationShape(value []int64) TPUReplicateMetadataAttr
- func TPUReplicateMetadataDeviceAssignment(value []int64) TPUReplicateMetadataAttr
- func TPUReplicateMetadataHostComputeCore(value []string) TPUReplicateMetadataAttr
- func TPUReplicateMetadataNumCoresPerReplica(value int64) TPUReplicateMetadataAttr
- func TPUReplicateMetadataPaddingMap(value []string) TPUReplicateMetadataAttr
- func TPUReplicateMetadataStepMarkerLocation(value string) TPUReplicateMetadataAttr
- func TPUReplicateMetadataTopology(value string) TPUReplicateMetadataAttr
- func TPUReplicateMetadataTpuCompileOptionsProto(value string) TPUReplicateMetadataAttr
- func TPUReplicateMetadataUseSpmdForXlaPartitioning(value bool) TPUReplicateMetadataAttr
- func TPUReplicateMetadataUseTpu(value bool) TPUReplicateMetadataAttr
type TPUReplicatedInputAttr
- func TPUReplicatedInputIndex(value int64) TPUReplicatedInputAttr
- func TPUReplicatedInputIsMirroredVariable(value bool) TPUReplicatedInputAttr
- func TPUReplicatedInputIsPacked(value bool) TPUReplicatedInputAttr
type TakeDatasetAttr
- func TakeDatasetMetadata(value string) TakeDatasetAttr
type TakeManySparseFromTensorsMapAttr
- func TakeManySparseFromTensorsMapContainer(value string) TakeManySparseFromTensorsMapAttr
- func TakeManySparseFromTensorsMapSharedName(value string) TakeManySparseFromTensorsMapAttr
type TensorArrayConcatV2Attr
- func TensorArrayConcatV2ElementShapeExcept0(value tf.Shape) TensorArrayConcatV2Attr
type TensorArrayConcatV3Attr
- func TensorArrayConcatV3ElementShapeExcept0(value tf.Shape) TensorArrayConcatV3Attr
type TensorArrayGatherV2Attr
- func TensorArrayGatherV2ElementShape(value tf.Shape) TensorArrayGatherV2Attr
type TensorArrayGatherV3Attr
- func TensorArrayGatherV3ElementShape(value tf.Shape) TensorArrayGatherV3Attr
type TensorArrayV2Attr
- func TensorArrayV2ClearAfterRead(value bool) TensorArrayV2Attr
- func TensorArrayV2DynamicSize(value bool) TensorArrayV2Attr
- func TensorArrayV2ElementShape(value tf.Shape) TensorArrayV2Attr
- func TensorArrayV2TensorArrayName(value string) TensorArrayV2Attr
type TensorArrayV3Attr
- func TensorArrayV3ClearAfterRead(value bool) TensorArrayV3Attr
- func TensorArrayV3DynamicSize(value bool) TensorArrayV3Attr
- func TensorArrayV3ElementShape(value tf.Shape) TensorArrayV3Attr
- func TensorArrayV3IdenticalElementShapes(value bool) TensorArrayV3Attr
- func TensorArrayV3TensorArrayName(value string) TensorArrayV3Attr
type TensorDatasetAttr
- func TensorDatasetMetadata(value string) TensorDatasetAttr
type TensorListConcatAttr
- func TensorListConcatElementShape(value tf.Shape) TensorListConcatAttr
type TensorListSetItemAttr
- func TensorListSetItemResizeIfIndexOutOfBounds(value bool) TensorListSetItemAttr
type TensorListStackAttr
- func TensorListStackNumElements(value int64) TensorListStackAttr
type TensorSliceDatasetAttr
- func TensorSliceDatasetIsFiles(value bool) TensorSliceDatasetAttr
- func TensorSliceDatasetMetadata(value string) TensorSliceDatasetAttr
- func TensorSliceDatasetReplicateOnSplit(value bool) TensorSliceDatasetAttr
type TensorStridedSliceUpdateAttr
- func TensorStridedSliceUpdateBeginMask(value int64) TensorStridedSliceUpdateAttr
- func TensorStridedSliceUpdateEllipsisMask(value int64) TensorStridedSliceUpdateAttr
- func TensorStridedSliceUpdateEndMask(value int64) TensorStridedSliceUpdateAttr
- func TensorStridedSliceUpdateNewAxisMask(value int64) TensorStridedSliceUpdateAttr
- func TensorStridedSliceUpdateShrinkAxisMask(value int64) TensorStridedSliceUpdateAttr
type TensorSummaryAttr
- func TensorSummaryDescription(value string) TensorSummaryAttr
- func TensorSummaryDisplayName(value string) TensorSummaryAttr
- func TensorSummaryLabels(value []string) TensorSummaryAttr
type TextLineDatasetAttr
- func TextLineDatasetMetadata(value string) TextLineDatasetAttr
type TextLineReaderV2Attr
- func TextLineReaderV2Container(value string) TextLineReaderV2Attr
- func TextLineReaderV2SharedName(value string) TextLineReaderV2Attr
- func TextLineReaderV2SkipHeaderLines(value int64) TextLineReaderV2Attr
type ThreadPoolHandleAttr
- func ThreadPoolHandleContainer(value string) ThreadPoolHandleAttr
- func ThreadPoolHandleMaxIntraOpParallelism(value int64) ThreadPoolHandleAttr
- func ThreadPoolHandleSharedName(value string) ThreadPoolHandleAttr
type ThreadUnsafeUnigramCandidateSamplerAttr
- func ThreadUnsafeUnigramCandidateSamplerSeed(value int64) ThreadUnsafeUnigramCandidateSamplerAttr
- func ThreadUnsafeUnigramCandidateSamplerSeed2(value int64) ThreadUnsafeUnigramCandidateSamplerAttr
type TopKAttr
- func TopKSorted(value bool) TopKAttr
type TopKV2Attr
- func TopKV2IndexType(value tf.DataType) TopKV2Attr
- func TopKV2Sorted(value bool) TopKV2Attr
type TridiagonalSolveAttr
- func TridiagonalSolvePartialPivoting(value bool) TridiagonalSolveAttr
- func TridiagonalSolvePerturbSingular(value bool) TridiagonalSolveAttr
type TruncatedNormalAttr
- func TruncatedNormalSeed(value int64) TruncatedNormalAttr
- func TruncatedNormalSeed2(value int64) TruncatedNormalAttr
type UnbatchAttr
- func UnbatchContainer(value string) UnbatchAttr
- func UnbatchSharedName(value string) UnbatchAttr
type UnbatchDatasetAttr
- func UnbatchDatasetMetadata(value string) UnbatchDatasetAttr
type UnbatchGradAttr
- func UnbatchGradContainer(value string) UnbatchGradAttr
- func UnbatchGradSharedName(value string) UnbatchGradAttr
type UnicodeDecodeAttr
- func UnicodeDecodeErrors(value string) UnicodeDecodeAttr
- func UnicodeDecodeReplaceControlCharacters(value bool) UnicodeDecodeAttr
- func UnicodeDecodeReplacementChar(value int64) UnicodeDecodeAttr
- func UnicodeDecodeTsplits(value tf.DataType) UnicodeDecodeAttr
type UnicodeDecodeWithOffsetsAttr
- func UnicodeDecodeWithOffsetsErrors(value string) UnicodeDecodeWithOffsetsAttr
- func UnicodeDecodeWithOffsetsReplaceControlCharacters(value bool) UnicodeDecodeWithOffsetsAttr
- func UnicodeDecodeWithOffsetsReplacementChar(value int64) UnicodeDecodeWithOffsetsAttr
- func UnicodeDecodeWithOffsetsTsplits(value tf.DataType) UnicodeDecodeWithOffsetsAttr
type UnicodeEncodeAttr
- func UnicodeEncodeErrors(value string) UnicodeEncodeAttr
- func UnicodeEncodeReplacementChar(value int64) UnicodeEncodeAttr
type UnicodeTranscodeAttr
- func UnicodeTranscodeErrors(value string) UnicodeTranscodeAttr
- func UnicodeTranscodeReplaceControlCharacters(value bool) UnicodeTranscodeAttr
- func UnicodeTranscodeReplacementChar(value int64) UnicodeTranscodeAttr
type UniformCandidateSamplerAttr
- func UniformCandidateSamplerSeed(value int64) UniformCandidateSamplerAttr
- func UniformCandidateSamplerSeed2(value int64) UniformCandidateSamplerAttr
type UniformDequantizeAttr
- func UniformDequantizeQuantizationAxis(value int64) UniformDequantizeAttr
type UniformQuantizeAttr
- func UniformQuantizeQuantizationAxis(value int64) UniformQuantizeAttr
type UniformQuantizedAddAttr
- func UniformQuantizedAddLhsQuantizationAxis(value int64) UniformQuantizedAddAttr
- func UniformQuantizedAddOutputQuantizationAxis(value int64) UniformQuantizedAddAttr
- func UniformQuantizedAddRhsQuantizationAxis(value int64) UniformQuantizedAddAttr
type UniformQuantizedClipByValueAttr
- func UniformQuantizedClipByValueQuantizationAxis(value int64) UniformQuantizedClipByValueAttr
type UniformQuantizedConvolutionAttr
- func UniformQuantizedConvolutionBatchGroupCount(value int64) UniformQuantizedConvolutionAttr
- func UniformQuantizedConvolutionDimensionNumbers(value string) UniformQuantizedConvolutionAttr
- func UniformQuantizedConvolutionExplicitPadding(value []int64) UniformQuantizedConvolutionAttr
- func UniformQuantizedConvolutionFeatureGroupCount(value int64) UniformQuantizedConvolutionAttr
- func UniformQuantizedConvolutionLhsDilation(value []int64) UniformQuantizedConvolutionAttr
- func UniformQuantizedConvolutionLhsQuantizationAxis(value int64) UniformQuantizedConvolutionAttr
- func UniformQuantizedConvolutionOutputQuantizationAxis(value int64) UniformQuantizedConvolutionAttr
- func UniformQuantizedConvolutionRhsDilation(value []int64) UniformQuantizedConvolutionAttr
- func UniformQuantizedConvolutionRhsQuantizationAxis(value int64) UniformQuantizedConvolutionAttr
- func UniformQuantizedConvolutionWindowStrides(value []int64) UniformQuantizedConvolutionAttr
type UniformQuantizedConvolutionHybridAttr
- func UniformQuantizedConvolutionHybridBatchGroupCount(value int64) UniformQuantizedConvolutionHybridAttr
- func UniformQuantizedConvolutionHybridDimensionNumbers(value string) UniformQuantizedConvolutionHybridAttr
- func UniformQuantizedConvolutionHybridExplicitPadding(value []int64) UniformQuantizedConvolutionHybridAttr
- func UniformQuantizedConvolutionHybridFeatureGroupCount(value int64) UniformQuantizedConvolutionHybridAttr
- func UniformQuantizedConvolutionHybridLhsDilation(value []int64) UniformQuantizedConvolutionHybridAttr
- func UniformQuantizedConvolutionHybridRhsDilation(value []int64) UniformQuantizedConvolutionHybridAttr
- func UniformQuantizedConvolutionHybridRhsQuantizationAxis(value int64) UniformQuantizedConvolutionHybridAttr
- func UniformQuantizedConvolutionHybridWindowStrides(value []int64) UniformQuantizedConvolutionHybridAttr
type UniformQuantizedDotAttr
- func UniformQuantizedDotLhsQuantizationAxis(value int64) UniformQuantizedDotAttr
- func UniformQuantizedDotOutputQuantizationAxis(value int64) UniformQuantizedDotAttr
- func UniformQuantizedDotRhsQuantizationAxis(value int64) UniformQuantizedDotAttr
type UniformQuantizedDotHybridAttr
- func UniformQuantizedDotHybridRhsQuantizationAxis(value int64) UniformQuantizedDotHybridAttr
type UniformRequantizeAttr
- func UniformRequantizeInputQuantizationAxis(value int64) UniformRequantizeAttr
- func UniformRequantizeOutputQuantizationAxis(value int64) UniformRequantizeAttr
type UniqueAttr
- func UniqueOutIdx(value tf.DataType) UniqueAttr
type UniqueDatasetAttr
- func UniqueDatasetMetadata(value string) UniqueDatasetAttr
type UniqueV2Attr
- func UniqueV2OutIdx(value tf.DataType) UniqueV2Attr
type UniqueWithCountsAttr
- func UniqueWithCountsOutIdx(value tf.DataType) UniqueWithCountsAttr
type UniqueWithCountsV2Attr
- func UniqueWithCountsV2OutIdx(value tf.DataType) UniqueWithCountsV2Attr
type UnpackAttr
- func UnpackAxis(value int64) UnpackAttr
type UnstageAttr
- func UnstageCapacity(value int64) UnstageAttr
- func UnstageContainer(value string) UnstageAttr
- func UnstageMemoryLimit(value int64) UnstageAttr
- func UnstageSharedName(value string) UnstageAttr
type UpperBoundAttr
- func UpperBoundOutType(value tf.DataType) UpperBoundAttr
type VarHandleOpAttr
- func VarHandleOpAllowedDevices(value []string) VarHandleOpAttr
- func VarHandleOpContainer(value string) VarHandleOpAttr
- func VarHandleOpDebugName(value string) VarHandleOpAttr
- func VarHandleOpSharedName(value string) VarHandleOpAttr
type VariableShapeAttr
- func VariableShapeOutType(value tf.DataType) VariableShapeAttr
type WholeFileReaderV2Attr
- func WholeFileReaderV2Container(value string) WholeFileReaderV2Attr
- func WholeFileReaderV2SharedName(value string) WholeFileReaderV2Attr
type WindowDatasetAttr
- func WindowDatasetMetadata(value string) WindowDatasetAttr
type WriteAudioSummaryAttr
- func WriteAudioSummaryMaxOutputs(value int64) WriteAudioSummaryAttr
type WriteImageSummaryAttr
- func WriteImageSummaryMaxImages(value int64) WriteImageSummaryAttr
type XlaConcatNDAttr
- func XlaConcatNDPaddings(value []int64) XlaConcatNDAttr
type XlaConvV2Attr
- func XlaConvV2BatchGroupCount(value int64) XlaConvV2Attr
type XlaRngBitGeneratorAttr
- func XlaRngBitGeneratorDtype(value tf.DataType) XlaRngBitGeneratorAttr
type XlaShardingAttr
- func XlaShardingSharding(value string) XlaShardingAttr
- func XlaShardingUnspecifiedDims(value []int64) XlaShardingAttr
type XlaSplitNDAttr
- func XlaSplitNDPaddings(value []int64) XlaSplitNDAttr
type XlaSpmdFullToShardShapeAttr
- func XlaSpmdFullToShardShapeDim(value int64) XlaSpmdFullToShardShapeAttr
- func XlaSpmdFullToShardShapeUnspecifiedDims(value []int64) XlaSpmdFullToShardShapeAttr
type XlaSpmdShardToFullShapeAttr
- func XlaSpmdShardToFullShapeDim(value int64) XlaSpmdShardToFullShapeAttr
- func XlaSpmdShardToFullShapeUnspecifiedDims(value []int64) XlaSpmdShardToFullShapeAttr
type ZipDatasetAttr
- func ZipDatasetMetadata(value string) ZipDatasetAttr

Constants ¶

This section is empty.

Variables ¶

This section is empty.

Functions ¶

func Abort ¶ added in v1.1.0

func Abort(scope *Scope, optional ...AbortAttr) (o *tf.Operation)

Raise a exception to abort the process when called.

If exit_without_error is true, the process will exit normally, otherwise it will exit with a SIGABORT signal.

Returns nothing but an exception.

Returns the created operation.

func Abs ¶ added in v1.1.0

func Abs(scope *Scope, x tf.Output) (y tf.Output)

Computes the absolute value of a tensor.

Given a tensor `x`, this operation returns a tensor containing the absolute value of each element in `x`. For example, if x is an input element and y is an output element, this operation computes \$y = |x|\$.

func AccumulateNV2 ¶ added in v1.5.0

func AccumulateNV2(scope *Scope, inputs []tf.Output, shape tf.Shape) (sum tf.Output)

Returns the element-wise sum of a list of tensors.

`tf.accumulate_n_v2` performs the same operation as `tf.add_n`, but does not wait for all of its inputs to be ready before beginning to sum. This can save memory if inputs are ready at different times, since minimum temporary storage is proportional to the output size rather than the inputs size.

Unlike the original `accumulate_n`, `accumulate_n_v2` is differentiable.

Returns a `Tensor` of same shape and type as the elements of `inputs`.

Arguments:

inputs: A list of `Tensor` objects, each with same shape and type.
shape: Shape of elements of `inputs`.

func Acos ¶ added in v1.1.0

func Acos(scope *Scope, x tf.Output) (y tf.Output)

Computes acos of x element-wise.

Provided an input tensor, the `tf.math.acos` operation returns the inverse cosine of each element of the tensor. If `y = tf.math.cos(x)` then, `x = tf.math.acos(y)`.

Input range is `[-1, 1]` and the output has a range of `[0, pi]`.

func Acosh ¶ added in v1.3.0

func Acosh(scope *Scope, x tf.Output) (y tf.Output)

Computes inverse hyperbolic cosine of x element-wise.

Given an input tensor, the function computes inverse hyperbolic cosine of every element. Input range is `[1, inf]`. It returns `nan` if the input lies outside the range.

```python x = tf.constant([-2, -0.5, 1, 1.2, 200, 10000, float("inf")]) tf.math.acosh(x) ==> [nan nan 0. 0.62236255 5.9914584 9.903487 inf] ```

func Add ¶ added in v1.1.0

func Add(scope *Scope, x tf.Output, y tf.Output) (z tf.Output)

Returns x + y element-wise.

*NOTE*: `Add` supports broadcasting. `AddN` does not. More about broadcasting [here](http://docs.scipy.org/doc/numpy/user/basics.broadcasting.html)

Given two input tensors, the `tf.add` operation computes the sum for every element in the tensor.

Both input and output have a range `(-inf, inf)`.

func AddManySparseToTensorsMap ¶ added in v1.1.0

func AddManySparseToTensorsMap(scope *Scope, sparse_indices tf.Output, sparse_values tf.Output, sparse_shape tf.Output, optional ...AddManySparseToTensorsMapAttr) (sparse_handles tf.Output)

Add an `N`-minibatch `SparseTensor` to a `SparseTensorsMap`, return `N` handles.

A `SparseTensor` of rank `R` is represented by three tensors: `sparse_indices`, `sparse_values`, and `sparse_shape`, where

```sparse_indices.shape[1] == sparse_shape.shape[0] == R```

An `N`-minibatch of `SparseTensor` objects is represented as a `SparseTensor` having a first `sparse_indices` column taking values between `[0, N)`, where the minibatch size `N == sparse_shape[0]`.

The input `SparseTensor` must have rank `R` greater than 1, and the first dimension is treated as the minibatch dimension. Elements of the `SparseTensor` must be sorted in increasing order of this first dimension. The stored `SparseTensor` objects pointed to by each row of the output `sparse_handles` will have rank `R-1`.

The `SparseTensor` values can then be read out as part of a minibatch by passing the given keys as vector elements to `TakeManySparseFromTensorsMap`. To ensure the correct `SparseTensorsMap` is accessed, ensure that the same `container` and `shared_name` are passed to that Op. If no `shared_name` is provided here, instead use the *name* of the Operation created by calling `AddManySparseToTensorsMap` as the `shared_name` passed to `TakeManySparseFromTensorsMap`. Ensure the Operations are colocated.

Arguments:

sparse_indices: 2-D.  The `indices` of the minibatch `SparseTensor`.

`sparse_indices[:, 0]` must be ordered values in `[0, N)`.

sparse_values: 1-D.  The `values` of the minibatch `SparseTensor`.
sparse_shape: 1-D.  The `shape` of the minibatch `SparseTensor`.

The minibatch size `N == sparse_shape[0]`.

Returns 1-D. The handles of the `SparseTensor` now stored in the `SparseTensorsMap`. Shape: `[N]`.

func AddN ¶ added in v1.1.0

func AddN(scope *Scope, inputs []tf.Output) (sum tf.Output)

Add all input tensors element wise.

Inputs must be of same size and shape.

```python
x = [9, 7, 10]
tf.math.add_n(x) ==> 26
```

func AddSparseToTensorsMap ¶ added in v1.1.0

func AddSparseToTensorsMap(scope *Scope, sparse_indices tf.Output, sparse_values tf.Output, sparse_shape tf.Output, optional ...AddSparseToTensorsMapAttr) (sparse_handle tf.Output)

Add a `SparseTensor` to a `SparseTensorsMap` return its handle.

A `SparseTensor` is represented by three tensors: `sparse_indices`, `sparse_values`, and `sparse_shape`.

This operator takes the given `SparseTensor` and adds it to a container object (a `SparseTensorsMap`). A unique key within this container is generated in the form of an `int64`, and this is the value that is returned.

The `SparseTensor` can then be read out as part of a minibatch by passing the key as a vector element to `TakeManySparseFromTensorsMap`. To ensure the correct `SparseTensorsMap` is accessed, ensure that the same `container` and `shared_name` are passed to that Op. If no `shared_name` is provided here, instead use the *name* of the Operation created by calling `AddSparseToTensorsMap` as the `shared_name` passed to `TakeManySparseFromTensorsMap`. Ensure the Operations are colocated.

Arguments:

sparse_indices: 2-D.  The `indices` of the `SparseTensor`.
sparse_values: 1-D.  The `values` of the `SparseTensor`.
sparse_shape: 1-D.  The `shape` of the `SparseTensor`.

Returns 0-D. The handle of the `SparseTensor` now stored in the `SparseTensorsMap`.

func AddV2 ¶ added in v1.5.0

func AddV2(scope *Scope, x tf.Output, y tf.Output) (z tf.Output)

Returns x + y element-wise.

*NOTE*: `Add` supports broadcasting. `AddN` does not. More about broadcasting [here](http://docs.scipy.org/doc/numpy/user/basics.broadcasting.html)

func AdjustContrast ¶ added in v1.1.0

func AdjustContrast(scope *Scope, images tf.Output, contrast_factor tf.Output, min_value tf.Output, max_value tf.Output) (output tf.Output)

Deprecated. Disallowed in GraphDef version >= 2.

DEPRECATED at GraphDef version 2: Use AdjustContrastv2 instead

func AdjustContrastv2 ¶ added in v1.1.0

func AdjustContrastv2(scope *Scope, images tf.Output, contrast_factor tf.Output) (output tf.Output)

Adjust the contrast of one or more images.

`images` is a tensor of at least 3 dimensions. The last 3 dimensions are interpreted as `[height, width, channels]`. The other dimensions only represent a collection of images, such as `[batch, height, width, channels].`

Contrast is adjusted independently for each channel of each image.

For each channel, the Op first computes the mean of the image pixels in the channel and then adjusts each component of each pixel to `(x - mean) * contrast_factor + mean`.

Arguments:

images: Images to adjust.  At least 3-D.
contrast_factor: A float multiplier for adjusting contrast.

Returns The contrast-adjusted image or images.

func AdjustHue ¶ added in v1.1.0

func AdjustHue(scope *Scope, images tf.Output, delta tf.Output) (output tf.Output)

Adjust the hue of one or more images.

`images` is a tensor of at least 3 dimensions. The last dimension is interpreted as channels, and must be three.

The input image is considered in the RGB colorspace. Conceptually, the RGB colors are first mapped into HSV. A delta is then applied all the hue values, and then remapped back to RGB colorspace.

Arguments:

images: Images to adjust.  At least 3-D.
delta: A float delta to add to the hue.

Returns The hue-adjusted image or images.

func AdjustSaturation ¶ added in v1.1.0

func AdjustSaturation(scope *Scope, images tf.Output, scale tf.Output) (output tf.Output)

Adjust the saturation of one or more images.

`images` is a tensor of at least 3 dimensions. The last dimension is interpreted as channels, and must be three.

The input image is considered in the RGB colorspace. Conceptually, the RGB colors are first mapped into HSV. A scale is then applied all the saturation values, and then remapped back to RGB colorspace.

Arguments:

images: Images to adjust.  At least 3-D.
scale: A float scale to add to the saturation.

Returns The hue-adjusted image or images.

func All ¶ added in v1.1.0

func All(scope *Scope, input tf.Output, axis tf.Output, optional ...AllAttr) (output tf.Output)

Computes the "logical and" of elements across dimensions of a tensor.

Reduces `input` along the dimensions given in `axis`. Unless `keep_dims` is true, the rank of the tensor is reduced by 1 for each entry in `axis`. If `keep_dims` is true, the reduced dimensions are retained with length 1.

Arguments:

input: The tensor to reduce.
axis: The dimensions to reduce. Must be in the range

`[-rank(input), rank(input))`.

Returns The reduced tensor.

func AllCandidateSampler ¶ added in v1.1.0

func AllCandidateSampler(scope *Scope, true_classes tf.Output, num_true int64, num_sampled int64, unique bool, optional ...AllCandidateSamplerAttr) (sampled_candidates tf.Output, true_expected_count tf.Output, sampled_expected_count tf.Output)

Generates labels for candidate sampling with a learned unigram distribution.

See explanations of candidate sampling and the data formats at go/candidate-sampling.

For each batch, this op picks a single set of sampled candidate labels.

The advantages of sampling candidates per-batch are simplicity and the possibility of efficient dense matrix multiplication. The disadvantage is that the sampled candidates must be chosen independently of the context and of the true labels.

Arguments:

true_classes: A batch_size * num_true matrix, in which each row contains the

IDs of the num_true target_classes in the corresponding original label.

num_true: Number of true labels per context.
num_sampled: Number of candidates to produce.
unique: If unique is true, we sample with rejection, so that all sampled

candidates in a batch are unique. This requires some approximation to estimate the post-rejection sampling probabilities.

Returns:

sampled_candidates: A vector of length num_sampled, in which each element is

the ID of a sampled candidate.

true_expected_count: A batch_size * num_true matrix, representing

the number of times each candidate is expected to occur in a batch of sampled candidates. If unique=true, then this is a probability.

sampled_expected_count: A vector of length num_sampled, for each sampled

candidate representing the number of times the candidate is expected to occur in a batch of sampled candidates. If unique=true, then this is a probability.

func AllToAll ¶ added in v1.12.1

func AllToAll(scope *Scope, input tf.Output, group_assignment tf.Output, concat_dimension int64, split_dimension int64, split_count int64) (output tf.Output)

An Op to exchange data across TPU replicas.

On each replica, the input is split into `split_count` blocks along `split_dimension` and send to the other replicas given group_assignment. After receiving `split_count` - 1 blocks from other replicas, we concatenate the blocks along `concat_dimension` as the output.

For example, suppose there are 2 TPU replicas: replica 0 receives input: `[[A, B]]` replica 1 receives input: `[[C, D]]`

group_assignment=`[[0, 1]]` concat_dimension=0 split_dimension=1 split_count=2

replica 0's output: `[[A], [C]]` replica 1's output: `[[B], [D]]`

Arguments:

input: The local input to the sum.
group_assignment: An int32 tensor with shape

[num_groups, num_replicas_per_group]. `group_assignment[i]` represents the replica ids in the ith subgroup.

concat_dimension: The dimension number to concatenate.
split_dimension: The dimension number to split.
split_count: The number of splits, this number must equal to the sub-group

size(group_assignment.get_shape()[1])

Returns The exchanged result.

func Angle ¶ added in v1.4.0

func Angle(scope *Scope, input tf.Output, optional ...AngleAttr) (output tf.Output)

Returns the argument of a complex number.

Given a tensor `input` of complex numbers, this operation returns a tensor of type `float` that is the argument of each element in `input`. All elements in `input` must be complex numbers of the form \$a + bj\$, where *a* is the real part and *b* is the imaginary part.

The argument returned by this operation is of the form \$atan2(b, a)\$.

For example:

``` # tensor 'input' is [-2.25 + 4.75j, 3.25 + 5.75j] tf.math.angle(input) ==> [2.0132, 1.056] ```

@compatibility(numpy) Equivalent to np.angle. @end_compatibility

func AnonymousHashTable ¶

func AnonymousHashTable(scope *Scope, key_dtype tf.DataType, value_dtype tf.DataType) (table_handle tf.Output)

Creates a uninitialized anonymous hash table.

This op creates a new anonymous hash table (as a resource) everytime it is executed, with the specified dtype of its keys and values, returning the resource handle. Before using the table you will have to initialize it. After initialization the table will be immutable. The table is anonymous in the sense that it can only be accessed by the returned resource handle (e.g. it cannot be looked up by a name in a resource manager). The table will be automatically deleted when all resource handles pointing to it are gone.

Arguments:

key_dtype: Type of the table keys.
value_dtype: Type of the table values.

Returns The resource handle to the newly created hash-table resource.

func AnonymousIterator ¶ added in v1.11.0

func AnonymousIterator(scope *Scope, output_types []tf.DataType, output_shapes []tf.Shape) (handle tf.Output)

A container for an iterator resource.

Returns A handle to the iterator that can be passed to a "MakeIterator" or "IteratorGetNext" op. In contrast to Iterator, AnonymousIterator prevents resource sharing by name, and does not keep a reference to the resource container.

func AnonymousIteratorV2 ¶ added in v1.12.1

func AnonymousIteratorV2(scope *Scope, output_types []tf.DataType, output_shapes []tf.Shape) (handle tf.Output, deleter tf.Output)

A container for an iterator resource.

Returns:

handle: A handle to the iterator that can be passed to a "MakeIterator" or

"IteratorGetNext" op. In contrast to Iterator, AnonymousIterator prevents resource sharing by name, and does not keep a reference to the resource container.

deleter: A variant deleter that should be passed into the op that deletes the iterator.

func AnonymousIteratorV3 ¶

func AnonymousIteratorV3(scope *Scope, output_types []tf.DataType, output_shapes []tf.Shape) (handle tf.Output)

A container for an iterator resource.

Returns A handle to the iterator that can be passed to a "MakeIterator" or "IteratorGetNext" op. In contrast to Iterator, AnonymousIterator prevents resource sharing by name, and does not keep a reference to the resource container.

func AnonymousMultiDeviceIterator ¶

func AnonymousMultiDeviceIterator(scope *Scope, devices []string, output_types []tf.DataType, output_shapes []tf.Shape) (handle tf.Output, deleter tf.Output)

A container for a multi device iterator resource.

Returns:

handle: A handle to a multi device iterator that can be passed to a

"MultiDeviceIteratorGetNextFromShard" op. In contrast to MultiDeviceIterator, AnonymousIterator prevents resource sharing by name, and does not keep a reference to the resource container.

deleter: A variant deleter that should be passed into the op that deletes the iterator.

func AnonymousMultiDeviceIteratorV3 ¶

func AnonymousMultiDeviceIteratorV3(scope *Scope, devices []string, output_types []tf.DataType, output_shapes []tf.Shape) (handle tf.Output)

A container for a multi device iterator resource.

Returns A handle to a multi device iterator that can be passed to a "MultiDeviceIteratorGetNextFromShard" op. In contrast to MultiDeviceIterator, AnonymousIterator prevents resource sharing by name, and does not keep a reference to the resource container.

func AnonymousMutableDenseHashTable ¶

func AnonymousMutableDenseHashTable(scope *Scope, empty_key tf.Output, deleted_key tf.Output, value_dtype tf.DataType, optional ...AnonymousMutableDenseHashTableAttr) (table_handle tf.Output)

Creates an empty anonymous mutable hash table that uses tensors as the backing store.

This op creates a new anonymous mutable hash table (as a resource) everytime it is executed, with the specified dtype of its keys and values, returning the resource handle. Each value must be a scalar. Data can be inserted into the table using the insert operations. It does not support the initialization operation.

It uses "open addressing" with quadratic reprobing to resolve collisions.

The table is anonymous in the sense that it can only be accessed by the returned resource handle (e.g. it cannot be looked up by a name in a resource manager). The table will be automatically deleted when all resource handles pointing to it are gone.

Arguments:

empty_key: The key used to represent empty key buckets internally. Must not

be used in insert or lookup operations.

value_dtype: Type of the table values.

Returns The resource handle to the newly created hash-table resource.

func AnonymousMutableHashTable ¶

func AnonymousMutableHashTable(scope *Scope, key_dtype tf.DataType, value_dtype tf.DataType) (table_handle tf.Output)

Creates an empty anonymous mutable hash table.

This op creates a new anonymous mutable hash table (as a resource) everytime it is executed, with the specified dtype of its keys and values, returning the resource handle. Each value must be a scalar. Data can be inserted into the table using the insert operations. It does not support the initialization operation. The table is anonymous in the sense that it can only be accessed by the returned resource handle (e.g. it cannot be looked up by a name in a resource manager). The table will be automatically deleted when all resource handles pointing to it are gone.

Arguments:

key_dtype: Type of the table keys.
value_dtype: Type of the table values.

Returns The resource handle to the newly created hash-table resource.

func AnonymousMutableHashTableOfTensors ¶

func AnonymousMutableHashTableOfTensors(scope *Scope, key_dtype tf.DataType, value_dtype tf.DataType, optional ...AnonymousMutableHashTableOfTensorsAttr) (table_handle tf.Output)

Creates an empty anonymous mutable hash table of vector values.

This op creates a new anonymous mutable hash table (as a resource) everytime it is executed, with the specified dtype of its keys and values, returning the resource handle. Each value must be a vector. Data can be inserted into the table using the insert operations. It does not support the initialization operation. The table is anonymous in the sense that it can only be accessed by the returned resource handle (e.g. it cannot be looked up by a name in a resource manager). The table will be automatically deleted when all resource handles pointing to it are gone.

Arguments:

key_dtype: Type of the table keys.
value_dtype: Type of the table values.

Returns The resource handle to the newly created hash-table resource.

func Any ¶ added in v1.1.0

func Any(scope *Scope, input tf.Output, axis tf.Output, optional ...AnyAttr) (output tf.Output)

Computes the "logical or" of elements across dimensions of a tensor.

Reduces `input` along the dimensions given in `axis`. Unless `keep_dims` is true, the rank of the tensor is reduced by 1 for each entry in `axis`. If `keep_dims` is true, the reduced dimensions are retained with length 1.

Arguments:

input: The tensor to reduce.
axis: The dimensions to reduce. Must be in the range

`[-rank(input), rank(input))`.

Returns The reduced tensor.

func ApproxTopK ¶

func ApproxTopK(scope *Scope, input tf.Output, k int64, optional ...ApproxTopKAttr) (values tf.Output, indices tf.Output)

Returns min/max k values and their indices of the input operand in an approximate manner.

See https://arxiv.org/abs/2206.14286 for the algorithm details. This op is only optimized on TPU currently.

Arguments:

input: Array to search. Must be at least 1-D of the floating type
k: Specifies the number of min/max-k.

Returns:

values: The min/max k values along the `reduction_dimension` of the `input` operand.

The dimension are the same as the `input` operand except for the `reduction_dimension`: when `aggregate_to_topk` is true, the reduction dimension is `k`; otherwise, it is greater equals to `k` where the size is implementation-defined.

indices: The indices of `values` along the `reduction_dimension` of the `input` operand.

func ApproximateEqual ¶ added in v1.1.0

func ApproximateEqual(scope *Scope, x tf.Output, y tf.Output, optional ...ApproximateEqualAttr) (z tf.Output)

Returns the truth value of abs(x-y) < tolerance element-wise.

func ArgMax ¶ added in v1.1.0

func ArgMax(scope *Scope, input tf.Output, dimension tf.Output, optional ...ArgMaxAttr) (output tf.Output)

Returns the index with the largest value across dimensions of a tensor.

Note that in case of ties the identity of the return value is not guaranteed.

Usage:

```python
import tensorflow as tf
a = [1, 10, 26.9, 2.8, 166.32, 62.3]
b = tf.math.argmax(input = a)
c = tf.keras.backend.eval(b)
# c = 4
# here a[4] = 166.32 which is the largest element of a across axis 0
```

Arguments:

dimension: int16, int32 or int64, must be in the range `[-rank(input), rank(input))`.

Describes which dimension of the input Tensor to reduce across. For vectors, use dimension = 0.

func ArgMin ¶ added in v1.1.0

func ArgMin(scope *Scope, input tf.Output, dimension tf.Output, optional ...ArgMinAttr) (output tf.Output)

Returns the index with the smallest value across dimensions of a tensor.

Note that in case of ties the identity of the return value is not guaranteed.

Usage:

```python
import tensorflow as tf
a = [1, 10, 26.9, 2.8, 166.32, 62.3]
b = tf.math.argmin(input = a)
c = tf.keras.backend.eval(b)
# c = 0
# here a[0] = 1 which is the smallest element of a across axis 0
```

Arguments:

dimension: int32 or int64, must be in the range `[-rank(input), rank(input))`.

Describes which dimension of the input Tensor to reduce across. For vectors, use dimension = 0.

func AsString ¶ added in v1.1.0

func AsString(scope *Scope, input tf.Output, optional ...AsStringAttr) (output tf.Output)

Converts each entry in the given tensor to strings.

Supports many numeric types and boolean.

For Unicode, see the [https://www.tensorflow.org/tutorials/representation/unicode](Working with Unicode text) tutorial.

Examples:

>>> tf.strings.as_string([3, 2]) <tf.Tensor: shape=(2,), dtype=string, numpy=array([b'3', b'2'], dtype=object)> >>> tf.strings.as_string([3.1415926, 2.71828], precision=2).numpy() array([b'3.14', b'2.72'], dtype=object)

func Asin ¶ added in v1.1.0

func Asin(scope *Scope, x tf.Output) (y tf.Output)

Computes the trignometric inverse sine of x element-wise.

The `tf.math.asin` operation returns the inverse of `tf.math.sin`, such that if `y = tf.math.sin(x)` then, `x = tf.math.asin(y)`.

**Note**: The output of `tf.math.asin` will lie within the invertible range of sine, i.e [-pi/2, pi/2].

For example:

```python # Note: [1.047, 0.785] ~= [(pi/3), (pi/4)] x = tf.constant([1.047, 0.785]) y = tf.math.sin(x) # [0.8659266, 0.7068252]

tf.math.asin(y) # [1.047, 0.785] = x ```

func Asinh ¶ added in v1.3.0

func Asinh(scope *Scope, x tf.Output) (y tf.Output)

Computes inverse hyperbolic sine of x element-wise.

Given an input tensor, this function computes inverse hyperbolic sine
for every element in the tensor. Both input and output has a range of
`[-inf, inf]`.

```python
x = tf.constant([-float("inf"), -2, -0.5, 1, 1.2, 200, 10000, float("inf")])
tf.math.asinh(x) ==> [-inf -1.4436355 -0.4812118 0.8813736 1.0159732 5.991471 9.903487 inf]
```

func Assert ¶ added in v1.1.0

func Assert(scope *Scope, condition tf.Output, data []tf.Output, optional ...AssertAttr) (o *tf.Operation)

Asserts that the given condition is true.

If `condition` evaluates to false, print the list of tensors in `data`. `summarize` determines how many entries of the tensors to print.

Arguments:

condition: The condition to evaluate.
data: The tensors to print out when condition is false.

Returns the created operation.

func AssertNextDataset ¶

func AssertNextDataset(scope *Scope, input_dataset tf.Output, transformations tf.Output, output_types []tf.DataType, output_shapes []tf.Shape) (handle tf.Output)

A transformation that asserts which transformations happen next.

This transformation checks whether the camel-case names (i.e. "FlatMap", not "flat_map") of the transformations following this transformation match the list of names in the `transformations` argument. If there is a mismatch, the transformation raises an exception.

The check occurs when iterating over the contents of the dataset, which means that the check happens *after* any static optimizations are applied to the dataset graph.

Arguments:

input_dataset: A variant tensor representing the input dataset.

`AssertNextDataset` passes through the outputs of its input dataset.

transformations: A `tf.string` vector `tf.Tensor` identifying the transformations that are

expected to happen next.

func AssertPrevDataset ¶

func AssertPrevDataset(scope *Scope, input_dataset tf.Output, transformations tf.Output, output_types []tf.DataType, output_shapes []tf.Shape) (handle tf.Output)

A transformation that asserts which transformations happened previously.

This transformation checks the names and, optionally, the attribute name-value pairs in the `transformations` argument against those of the transformations that preceded this transformation. If there is a mismatch, the transformation raises an exception.

The check occurs when iterating over the contents of the dataset, which means that the check happens *after* any static optimizations are applied to the dataset graph.

Arguments:

input_dataset: A variant tensor representing the input dataset.

`AssertPrevDataset` passes through the outputs of its input dataset.

transformations: A `tf.string` vector `tf.Tensor` identifying the transformations, with optional

attribute name-value pairs, that are expected to have happened previously.

func AssignAddVariableOp ¶ added in v1.1.0

func AssignAddVariableOp(scope *Scope, resource tf.Output, value tf.Output) (o *tf.Operation)

Adds a value to the current value of a variable.

Any ReadVariableOp with a control dependency on this op is guaranteed to see the incremented value or a subsequent newer one.

Arguments:

resource: handle to the resource in which to store the variable.
value: the value by which the variable will be incremented.

Returns the created operation.

func AssignSubVariableOp ¶ added in v1.1.0

func AssignSubVariableOp(scope *Scope, resource tf.Output, value tf.Output) (o *tf.Operation)

Subtracts a value from the current value of a variable.

Any ReadVariableOp with a control dependency on this op is guaranteed to see the decremented value or a subsequent newer one.

Arguments:

resource: handle to the resource in which to store the variable.
value: the value by which the variable will be incremented.

Returns the created operation.

func AssignVariableOp ¶ added in v1.1.0

func AssignVariableOp(scope *Scope, resource tf.Output, value tf.Output, optional ...AssignVariableOpAttr) (o *tf.Operation)

Assigns a new value to a variable.

Any ReadVariableOp with a control dependency on this op is guaranteed to return this value or a subsequent newer value of the variable.

Arguments:

resource: handle to the resource in which to store the variable.
value: the value to set the new tensor to use.

Returns the created operation.

func AssignVariableXlaConcatND ¶

func AssignVariableXlaConcatND(scope *Scope, resource tf.Output, inputs []tf.Output, num_concats []int64, optional ...AssignVariableXlaConcatNDAttr) (o *tf.Operation)

Concats input tensor across all dimensions.

An op which merges slices the input tensor based on the given num_splits attribute, strips paddings optionally, and writes the merged tensor without paddings to the resource variable.

This op may be generated via the TPU bridge.

For example, with `input` tensor: ``` [[0, 1],

[4, 5]]

[[2, 3],

[6, 7]]

[[8, 9],

[12, 13]]

[[10, 11],

[14, 15]]

``` `num_splits`: ``` [2, 2] ``` and `paddings`: ``` [1, 1] ``` the expected `outputs` is: ``` [[0, 1, 2],

[4, 5, 6],
[8, 9, 10]]

```

Arguments:

	resource: Resource variable for concatenated input tensors across all dimensions.
  }
  in_arg {
    name: "inputs"
    description: <<END

Input tensor slices in row-major order to merge across all dimensions. All inputs must have the same shape.

}
out_arg {
  name: "output"
  description: <<END

Output tensor formed from merging input slices based on num_concats defined.

num_concats: Number of ways to merge per dimension.

Returns the created operation.

func Atan ¶ added in v1.1.0

func Atan(scope *Scope, x tf.Output) (y tf.Output)

Computes the trignometric inverse tangent of x element-wise.

The `tf.math.atan` operation returns the inverse of `tf.math.tan`, such that if `y = tf.math.tan(x)` then, `x = tf.math.atan(y)`.

**Note**: The output of `tf.math.atan` will lie within the invertible range of tan, i.e (-pi/2, pi/2).

For example:

```python # Note: [1.047, 0.785] ~= [(pi/3), (pi/4)] x = tf.constant([1.047, 0.785]) y = tf.math.tan(x) # [1.731261, 0.99920404]

tf.math.atan(y) # [1.047, 0.785] = x ```

func Atan2 ¶ added in v1.2.0

func Atan2(scope *Scope, y tf.Output, x tf.Output) (z tf.Output)

Computes arctangent of `y/x` element-wise, respecting signs of the arguments.

This is the angle \$ \theta \in [-\pi, \pi] \$ such that \\[ x = r \cos(\theta) \\] and \\[ y = r \sin(\theta) \\] where \$r = \sqrt{x^2 + y^2} \$.

For example:

>>> x = [1., 1.] >>> y = [1., -1.] >>> print((tf.math.atan2(y,x) * (180 / np.pi)).numpy()) [ 45. -45.]

func Atanh ¶ added in v1.3.0

func Atanh(scope *Scope, x tf.Output) (y tf.Output)

Computes inverse hyperbolic tangent of x element-wise.

Given an input tensor, this function computes inverse hyperbolic tangent
for every element in the tensor. Input range is `[-1,1]` and output range is
`[-inf, inf]`. If input is `-1`, output will be `-inf` and if the
input is `1`, output will be `inf`. Values outside the range will have
`nan` as output.

```python
x = tf.constant([-float("inf"), -1, -0.5, 1, 0, 0.5, 10, float("inf")])
tf.math.atanh(x) ==> [nan -inf -0.54930615 inf  0. 0.54930615 nan nan]
```

func AudioSpectrogram ¶ added in v1.2.0

func AudioSpectrogram(scope *Scope, input tf.Output, window_size int64, stride int64, optional ...AudioSpectrogramAttr) (spectrogram tf.Output)

Produces a visualization of audio data over time.

Spectrograms are a standard way of representing audio information as a series of slices of frequency information, one slice for each window of time. By joining these together into a sequence, they form a distinctive fingerprint of the sound over time.

This op expects to receive audio data as an input, stored as floats in the range -1 to 1, together with a window width in samples, and a stride specifying how far to move the window between slices. From this it generates a three dimensional output. The first dimension is for the channels in the input, so a stereo audio input would have two here for example. The second dimension is time, with successive frequency slices. The third dimension has an amplitude value for each frequency during that time slice.

This means the layout when converted and saved as an image is rotated 90 degrees clockwise from a typical spectrogram. Time is descending down the Y axis, and the frequency decreases from left to right.

Each value in the result represents the square root of the sum of the real and imaginary parts of an FFT on the current window of samples. In this way, the lowest dimension represents the power of each frequency in the current window, and adjacent windows are concatenated in the next dimension.

To get a more intuitive and visual look at what this operation does, you can run tensorflow/examples/wav_to_spectrogram to read in an audio file and save out the resulting spectrogram as a PNG image.

Arguments:

input: Float representation of audio data.
window_size: How wide the input window is in samples. For the highest efficiency

this should be a power of two, but other values are accepted.

stride: How widely apart the center of adjacent sample windows should be.

Returns 3D representation of the audio frequencies as an image.

func AudioSummary ¶ added in v1.1.0

func AudioSummary(scope *Scope, tag tf.Output, tensor tf.Output, sample_rate float32, optional ...AudioSummaryAttr) (summary tf.Output)

Outputs a `Summary` protocol buffer with audio.

DEPRECATED at GraphDef version 15: Use AudioSummaryV2.

The summary has up to `max_outputs` summary values containing audio. The audio is built from `tensor` which must be 3-D with shape `[batch_size, frames, channels]` or 2-D with shape `[batch_size, frames]`. The values are assumed to be in the range of `[-1.0, 1.0]` with a sample rate of `sample_rate`.

The `tag` argument is a scalar `Tensor` of type `string`. It is used to build the `tag` of the summary values:

If `max_outputs` is 1, the summary value tag is '*tag*/audio'.
If `max_outputs` is greater than 1, the summary value tags are generated sequentially as '*tag*/audio/0', '*tag*/audio/1', etc.

Arguments:

tag: Scalar. Used to build the `tag` attribute of the summary values.
tensor: 2-D of shape `[batch_size, frames]`.
sample_rate: The sample rate of the signal in hertz.

Returns Scalar. Serialized `Summary` protocol buffer.

func AudioSummaryV2 ¶ added in v1.1.0

func AudioSummaryV2(scope *Scope, tag tf.Output, tensor tf.Output, sample_rate tf.Output, optional ...AudioSummaryV2Attr) (summary tf.Output)

Outputs a `Summary` protocol buffer with audio.

The summary has up to `max_outputs` summary values containing audio. The audio is built from `tensor` which must be 3-D with shape `[batch_size, frames, channels]` or 2-D with shape `[batch_size, frames]`. The values are assumed to be in the range of `[-1.0, 1.0]` with a sample rate of `sample_rate`.

The `tag` argument is a scalar `Tensor` of type `string`. It is used to build the `tag` of the summary values:

If `max_outputs` is 1, the summary value tag is '*tag*/audio'.
If `max_outputs` is greater than 1, the summary value tags are generated sequentially as '*tag*/audio/0', '*tag*/audio/1', etc.

Arguments:

tag: Scalar. Used to build the `tag` attribute of the summary values.
tensor: 2-D of shape `[batch_size, frames]`.
sample_rate: The sample rate of the signal in hertz.

Returns Scalar. Serialized `Summary` protocol buffer.

func AutoShardDataset ¶ added in v1.15.0

func AutoShardDataset(scope *Scope, input_dataset tf.Output, num_workers tf.Output, index tf.Output, output_types []tf.DataType, output_shapes []tf.Shape, optional ...AutoShardDatasetAttr) (handle tf.Output)

Creates a dataset that shards the input dataset.

Creates a dataset that shards the input dataset by num_workers, returning a sharded dataset for the index-th worker. This attempts to automatically shard a dataset by examining the Dataset graph and inserting a shard op before the inputs to a reader Dataset (e.g. CSVDataset, TFRecordDataset).

This dataset will throw a NotFound error if we cannot shard the dataset automatically.

Arguments:

input_dataset: A variant tensor representing the input dataset.
num_workers: A scalar representing the number of workers to distribute this dataset across.
index: A scalar representing the index of the current worker out of num_workers.

func AvgPool ¶ added in v1.1.0

func AvgPool(scope *Scope, value tf.Output, ksize []int64, strides []int64, padding string, optional ...AvgPoolAttr) (output tf.Output)

Performs average pooling on the input.

Each entry in `output` is the mean of the corresponding size `ksize` window in `value`.

Arguments:

value: 4-D with shape `[batch, height, width, channels]`.
ksize: The size of the sliding window for each dimension of `value`.
strides: The stride of the sliding window for each dimension of `value`.
padding: The type of padding algorithm to use.

Returns The average pooled output tensor.

func AvgPool3D ¶ added in v1.1.0

func AvgPool3D(scope *Scope, input tf.Output, ksize []int64, strides []int64, padding string, optional ...AvgPool3DAttr) (output tf.Output)

Performs 3D average pooling on the input.

Each entry in `output` is the mean of the corresponding size `ksize` window in `value`.

Arguments:

input: Shape `[batch, depth, rows, cols, channels]` tensor to pool over.
ksize: 1-D tensor of length 5. The size of the window for each dimension of

the input tensor. Must have `ksize[0] = ksize[4] = 1`.

strides: 1-D tensor of length 5. The stride of the sliding window for each

dimension of `input`. Must have `strides[0] = strides[4] = 1`.

padding: The type of padding algorithm to use.

Returns The average pooled output tensor.

func AvgPool3DGrad ¶ added in v1.1.0

func AvgPool3DGrad(scope *Scope, orig_input_shape tf.Output, grad tf.Output, ksize []int64, strides []int64, padding string, optional ...AvgPool3DGradAttr) (output tf.Output)

Computes gradients of average pooling function.

Arguments:

orig_input_shape: The original input dimensions.
grad: Output backprop of shape `[batch, depth, rows, cols, channels]`.
ksize: 1-D tensor of length 5. The size of the window for each dimension of

the input tensor. Must have `ksize[0] = ksize[4] = 1`.

strides: 1-D tensor of length 5. The stride of the sliding window for each

dimension of `input`. Must have `strides[0] = strides[4] = 1`.

padding: The type of padding algorithm to use.

Returns The backprop for input.

func AvgPoolGrad ¶ added in v1.1.0

func AvgPoolGrad(scope *Scope, orig_input_shape tf.Output, grad tf.Output, ksize []int64, strides []int64, padding string, optional ...AvgPoolGradAttr) (output tf.Output)

Computes gradients of the average pooling function.

Arguments:

orig_input_shape: 1-D.  Shape of the original input to `avg_pool`.
grad: 4-D with shape `[batch, height, width, channels]`.  Gradients w.r.t.

the output of `avg_pool`.

ksize: The size of the sliding window for each dimension of the input.
strides: The stride of the sliding window for each dimension of the input.
padding: The type of padding algorithm to use.

Returns 4-D. Gradients w.r.t. the input of `avg_pool`.

func Batch ¶ added in v1.9.0

func Batch(scope *Scope, in_tensors []tf.Output, num_batch_threads int64, max_batch_size int64, batch_timeout_micros int64, grad_timeout_micros int64, optional ...BatchAttr) (batched_tensors []tf.Output, batch_index tf.Output, id tf.Output)

Batches all input tensors nondeterministically.

When many instances of this Op are being run concurrently with the same container/shared_name in the same device, some will output zero-shaped Tensors and others will output Tensors of size up to max_batch_size.

All Tensors in in_tensors are batched together (so, for example, labels and features should be batched with a single instance of this operation.

Each invocation of batch emits an `id` scalar which will be used to identify this particular invocation when doing unbatch or its gradient.

Each op which emits a non-empty batch will also emit a non-empty batch_index Tensor, which, is a [K, 3] matrix where each row contains the invocation's id, start, and length of elements of each set of Tensors present in batched_tensors.

Batched tensors are concatenated along the first dimension, and all tensors in in_tensors must have the first dimension of the same size.

in_tensors: The tensors to be batched. num_batch_threads: Number of scheduling threads for processing batches of work.

Determines the number of batches processed in parallel.

max_batch_size: Batch sizes will never be bigger than this. batch_timeout_micros: Maximum number of microseconds to wait before outputting

an incomplete batch.

allowed_batch_sizes: Optional list of allowed batch sizes. If left empty, does

nothing. Otherwise, supplies a list of batch sizes, causing the op to pad
batches up to one of those sizes. The entries must increase monotonically, and
the final entry must equal max_batch_size.

grad_timeout_micros: The timeout to use for the gradient. See Unbatch. batched_tensors: Either empty tensors or a batch of concatenated Tensors. batch_index: If out_tensors is non-empty, has information to invert it. container: Controls the scope of sharing of this batch. id: always contains a scalar with a unique ID for this invocation of Batch. shared_name: Concurrently running instances of batch in the same device with the

same container and shared_name will batch their elements together. If left
empty, the op name will be used as the shared name.

T: the types of tensors to be batched.

func BatchDataset ¶ added in v1.2.0

func BatchDataset(scope *Scope, input_dataset tf.Output, batch_size tf.Output, output_types []tf.DataType, output_shapes []tf.Shape, optional ...BatchDatasetAttr) (handle tf.Output)

Creates a dataset that batches `batch_size` elements from `input_dataset`.

Arguments:

batch_size: A scalar representing the number of elements to accumulate in a

batch.

func BatchDatasetV2 ¶ added in v1.11.0

func BatchDatasetV2(scope *Scope, input_dataset tf.Output, batch_size tf.Output, drop_remainder tf.Output, output_types []tf.DataType, output_shapes []tf.Shape, optional ...BatchDatasetV2Attr) (handle tf.Output)

Creates a dataset that batches `batch_size` elements from `input_dataset`.

Arguments:

batch_size: A scalar representing the number of elements to accumulate in a batch.
drop_remainder: A scalar representing whether the last batch should be dropped in case its size

is smaller than desired.

func BatchMatMul ¶ added in v1.1.0

func BatchMatMul(scope *Scope, x tf.Output, y tf.Output, optional ...BatchMatMulAttr) (output tf.Output)

Multiplies slices of two tensors in batches.

Multiplies all slices of `Tensor` `x` and `y` (each slice can be viewed as an element of a batch), and arranges the individual results in a single output tensor of the same batch size. Each of the individual slices can optionally be adjointed (to adjoint a matrix means to transpose and conjugate it) before multiplication by setting the `adj_x` or `adj_y` flag to `True`, which are by default `False`.

The input tensors `x` and `y` are 2-D or higher with shape `[..., r_x, c_x]` and `[..., r_y, c_y]`.

The output tensor is 2-D or higher with shape `[..., r_o, c_o]`, where:

r_o = c_x if adj_x else r_x
c_o = r_y if adj_y else c_y

It is computed as:

output[..., :, :] = matrix(x[..., :, :]) * matrix(y[..., :, :])

Arguments:

x: 2-D or higher with shape `[..., r_x, c_x]`.
y: 2-D or higher with shape `[..., r_y, c_y]`.

Returns 3-D or higher with shape `[..., r_o, c_o]`

func BatchMatMulV2 ¶ added in v1.12.1

func BatchMatMulV2(scope *Scope, x tf.Output, y tf.Output, optional ...BatchMatMulV2Attr) (output tf.Output)

Multiplies slices of two tensors in batches.

Multiplies all slices of `Tensor` `x` and `y` (each slice can be viewed as an element of a batch), and arranges the individual results in a single output tensor of the same batch size. Each of the individual slices can optionally be adjointed (to adjoint a matrix means to transpose and conjugate it) before multiplication by setting the `adj_x` or `adj_y` flag to `True`, which are by default `False`.

The input tensors `x` and `y` are 2-D or higher with shape `[..., r_x, c_x]` and `[..., r_y, c_y]`.

The output tensor is 2-D or higher with shape `[..., r_o, c_o]`, where:

r_o = c_x if adj_x else r_x
c_o = r_y if adj_y else c_y

It is computed as:

output[..., :, :] = matrix(x[..., :, :]) * matrix(y[..., :, :])

*NOTE*: `BatchMatMulV2` supports broadcasting in the batch dimensions. More about broadcasting [here](http://docs.scipy.org/doc/numpy/user/basics.broadcasting.html).

Arguments:

x: 2-D or higher with shape `[..., r_x, c_x]`.
y: 2-D or higher with shape `[..., r_y, c_y]`.

Returns 3-D or higher with shape `[..., r_o, c_o]`

func BatchMatMulV3 ¶

func BatchMatMulV3(scope *Scope, x tf.Output, y tf.Output, Tout tf.DataType, optional ...BatchMatMulV3Attr) (output tf.Output)

Multiplies slices of two tensors in batches.

Multiplies all slices of `Tensor` `x` and `y` (each slice can be viewed as an element of a batch), and arranges the individual results in a single output tensor of the same batch size. Each of the individual slices can optionally be adjointed (to adjoint a matrix means to transpose and conjugate it) before multiplication by setting the `adj_x` or `adj_y` flag to `True`, which are by default `False`.

The input tensors `x` and `y` are 2-D or higher with shape `[..., r_x, c_x]` and `[..., r_y, c_y]`.

The output tensor is 2-D or higher with shape `[..., r_o, c_o]`, where:

r_o = c_x if adj_x else r_x
c_o = r_y if adj_y else c_y

It is computed as:

output[..., :, :] = matrix(x[..., :, :]) * matrix(y[..., :, :])

*NOTE*: `BatchMatMulV3` supports broadcasting in the batch dimensions. More about broadcasting [here](http://docs.scipy.org/doc/numpy/user/basics.broadcasting.html).

Arguments:

x: 2-D or higher with shape `[..., r_x, c_x]`.
y: 2-D or higher with shape `[..., r_y, c_y]`.
Tout: If not spcified, Tout is the same type to input type.

Returns 3-D or higher with shape `[..., r_o, c_o]`

func BatchNormWithGlobalNormalization ¶ added in v1.1.0

func BatchNormWithGlobalNormalization(scope *Scope, t tf.Output, m tf.Output, v tf.Output, beta tf.Output, gamma tf.Output, variance_epsilon float32, scale_after_normalization bool) (result tf.Output)

Batch normalization.

DEPRECATED at GraphDef version 9: Use tf.nn.batch_normalization()

This op is deprecated. Prefer `tf.nn.batch_normalization`.

Arguments:

t: A 4D input Tensor.
m: A 1D mean Tensor with size matching the last dimension of t.

This is the first output from tf.nn.moments, or a saved moving average thereof.

v: A 1D variance Tensor with size matching the last dimension of t.

This is the second output from tf.nn.moments, or a saved moving average thereof.

beta: A 1D beta Tensor with size matching the last dimension of t.

An offset to be added to the normalized tensor.

gamma: A 1D gamma Tensor with size matching the last dimension of t.

If "scale_after_normalization" is true, this tensor will be multiplied with the normalized tensor.

variance_epsilon: A small float number to avoid dividing by 0.
scale_after_normalization: A bool indicating whether the resulted tensor

needs to be multiplied with gamma.

func BatchNormWithGlobalNormalizationGrad ¶ added in v1.1.0

func BatchNormWithGlobalNormalizationGrad(scope *Scope, t tf.Output, m tf.Output, v tf.Output, gamma tf.Output, backprop tf.Output, variance_epsilon float32, scale_after_normalization bool) (dx tf.Output, dm tf.Output, dv tf.Output, db tf.Output, dg tf.Output)

Gradients for batch normalization.

DEPRECATED at GraphDef version 9: Use tf.nn.batch_normalization()

This op is deprecated. See `tf.nn.batch_normalization`.

Arguments:

t: A 4D input Tensor.
m: A 1D mean Tensor with size matching the last dimension of t.

This is the first output from tf.nn.moments, or a saved moving average thereof.

v: A 1D variance Tensor with size matching the last dimension of t.

This is the second output from tf.nn.moments, or a saved moving average thereof.

gamma: A 1D gamma Tensor with size matching the last dimension of t.

If "scale_after_normalization" is true, this Tensor will be multiplied with the normalized Tensor.

backprop: 4D backprop Tensor.
variance_epsilon: A small float number to avoid dividing by 0.
scale_after_normalization: A bool indicating whether the resulted tensor

needs to be multiplied with gamma.

Returns:

dx: 4D backprop tensor for input.
dm: 1D backprop tensor for mean.
dv: 1D backprop tensor for variance.
db: 1D backprop tensor for beta.
dg: 1D backprop tensor for gamma.

func BatchToSpace ¶ added in v1.1.0

func BatchToSpace(scope *Scope, input tf.Output, crops tf.Output, block_size int64) (output tf.Output)

BatchToSpace for 4-D tensors of type T.

This is a legacy version of the more general BatchToSpaceND.

Rearranges (permutes) data from batch into blocks of spatial data, followed by cropping. This is the reverse transformation of SpaceToBatch. More specifically, this op outputs a copy of the input tensor where values from the `batch` dimension are moved in spatial blocks to the `height` and `width` dimensions, followed by cropping along the `height` and `width` dimensions.

Arguments:

input: 4-D tensor with shape

`[batch*block_size*block_size, height_pad/block_size, width_pad/block_size,

depth]`. Note that the batch size of the input tensor must be divisible by

`block_size * block_size`.

crops: 2-D tensor of non-negative integers with shape `[2, 2]`. It specifies

how many elements to crop from the intermediate result across the spatial dimensions as follows:

crops = [[crop_top, crop_bottom], [crop_left, crop_right]]

Returns 4-D with shape `[batch, height, width, depth]`, where:

height = height_pad - crop_top - crop_bottom
width = width_pad - crop_left - crop_right

The attr `block_size` must be greater than one. It indicates the block size.

Some examples:

(1) For the following input of shape `[4, 1, 1, 1]` and block_size of 2:

``` [[[[1]]], [[[2]]], [[[3]]], [[[4]]]] ```

The output tensor has shape `[1, 2, 2, 1]` and value:

``` x = [[[[1], [2]], [[3], [4]]]] ```

(2) For the following input of shape `[4, 1, 1, 3]` and block_size of 2:

``` [[[[1, 2, 3]]], [[[4, 5, 6]]], [[[7, 8, 9]]], [[[10, 11, 12]]]] ```

The output tensor has shape `[1, 2, 2, 3]` and value:

``` x = [[[[1, 2, 3], [4, 5, 6]],

[[7, 8, 9], [10, 11, 12]]]]

```

(3) For the following input of shape `[4, 2, 2, 1]` and block_size of 2:

``` x = [[[[1], [3]], [[9], [11]]],

[[[2], [4]], [[10], [12]]],
[[[5], [7]], [[13], [15]]],
[[[6], [8]], [[14], [16]]]]

```

The output tensor has shape `[1, 4, 4, 1]` and value:

``` x = [[[[1], [2], [3], [4]],

[[5],   [6],  [7],  [8]],
[[9],  [10], [11],  [12]],
[[13], [14], [15],  [16]]]]

```

(4) For the following input of shape `[8, 1, 2, 1]` and block_size of 2:

``` x = [[[[1], [3]]], [[[9], [11]]], [[[2], [4]]], [[[10], [12]]],

[[[5], [7]]], [[[13], [15]]], [[[6], [8]]], [[[14], [16]]]]

```

The output tensor has shape `[2, 2, 4, 1]` and value:

``` x = [[[[1], [3]], [[5], [7]]],

[[[2], [4]], [[10], [12]]],
[[[5], [7]], [[13], [15]]],
[[[6], [8]], [[14], [16]]]]

```

func BatchToSpaceND ¶ added in v1.1.0

func BatchToSpaceND(scope *Scope, input tf.Output, block_shape tf.Output, crops tf.Output) (output tf.Output)

BatchToSpace for N-D tensors of type T.

This operation reshapes the "batch" dimension 0 into `M + 1` dimensions of shape `block_shape + [batch]`, interleaves these blocks back into the grid defined by the spatial dimensions `[1, ..., M]`, to obtain a result with the same rank as the input. The spatial dimensions of this intermediate result are then optionally cropped according to `crops` to produce the output. This is the reverse of SpaceToBatch. See below for a precise description.

Arguments:

input: N-D with shape `input_shape = [batch] + spatial_shape + remaining_shape`,

where spatial_shape has M dimensions.

	block_shape: 1-D with shape `[M]`, all values must be >= 1.
	crops: 2-D with shape `[M, 2]`, all values must be >= 0.
  `crops[i] = [crop_start, crop_end]` specifies the amount to crop from input
  dimension `i + 1`, which corresponds to spatial dimension `i`.  It is
  required that
  `crop_start[i] + crop_end[i] <= block_shape[i] * input_shape[i + 1]`.

This operation is equivalent to the following steps:

Reshape `input` to `reshaped` of shape: [block_shape[0], ..., block_shape[M-1], batch / prod(block_shape), input_shape[1], ..., input_shape[N-1]]
Permute dimensions of `reshaped` to produce `permuted` of shape [batch / prod(block_shape),
input_shape[1], block_shape[0], ..., input_shape[M], block_shape[M-1],
input_shape[M+1], ..., input_shape[N-1]]
Reshape `permuted` to produce `reshaped_permuted` of shape [batch / prod(block_shape),
input_shape[1] * block_shape[0], ..., input_shape[M] * block_shape[M-1],
input_shape[M+1], ..., input_shape[N-1]]
Crop the start and end of dimensions `[1, ..., M]` of `reshaped_permuted` according to `crops` to produce the output of shape: [batch / prod(block_shape),
input_shape[1] * block_shape[0] - crops[0,0] - crops[0,1], ..., input_shape[M] * block_shape[M-1] - crops[M-1,0] - crops[M-1,1],
input_shape[M+1], ..., input_shape[N-1]]

Some examples:

(1) For the following input of shape `[4, 1, 1, 1]`, `block_shape = [2, 2]`, and

`crops = [[0, 0], [0, 0]]`:

``` [[[[1]]], [[[2]]], [[[3]]], [[[4]]]] ```

The output tensor has shape `[1, 2, 2, 1]` and value:

``` x = [[[[1], [2]], [[3], [4]]]] ```

(2) For the following input of shape `[4, 1, 1, 3]`, `block_shape = [2, 2]`, and

`crops = [[0, 0], [0, 0]]`:

``` [[[[1, 2, 3]]], [[[4, 5, 6]]], [[[7, 8, 9]]], [[[10, 11, 12]]]] ```

The output tensor has shape `[1, 2, 2, 3]` and value:

``` x = [[[[1, 2, 3], [4, 5, 6]],

[[7, 8, 9], [10, 11, 12]]]]

```

(3) For the following input of shape `[4, 2, 2, 1]`, `block_shape = [2, 2]`, and

`crops = [[0, 0], [0, 0]]`:

``` x = [[[[1], [3]], [[9], [11]]],

[[[2], [4]], [[10], [12]]],
[[[5], [7]], [[13], [15]]],
[[[6], [8]], [[14], [16]]]]

```

The output tensor has shape `[1, 4, 4, 1]` and value:

``` x = [[[[1], [2], [3], [4]],

[[5],   [6],  [7],  [8]],
[[9],  [10], [11],  [12]],
[[13], [14], [15],  [16]]]]

```

(4) For the following input of shape `[8, 1, 3, 1]`, `block_shape = [2, 2]`, and

`crops = [[0, 0], [2, 0]]`:

``` x = [[[[0], [1], [3]]], [[[0], [9], [11]]],

[[[0], [2], [4]]], [[[0], [10], [12]]],
[[[0], [5], [7]]], [[[0], [13], [15]]],
[[[0], [6], [8]]], [[[0], [14], [16]]]]

```

The output tensor has shape `[2, 2, 4, 1]` and value:

``` x = [[[[1], [2], [3], [4]],

 [[5],   [6],  [7],  [8]]],
[[[9],  [10], [11],  [12]],
 [[13], [14], [15],  [16]]]]

```

func Betainc ¶ added in v1.1.0

func Betainc(scope *Scope, a tf.Output, b tf.Output, x tf.Output) (z tf.Output)

Compute the regularized incomplete beta integral \$I_x(a, b)\$.

The regularized incomplete beta integral is defined as:

\$I_x(a, b) = \frac{B(x; a, b)}{B(a, b)}\$

where

\$B(x; a, b) = \int_0^x t^{a-1} (1 - t)^{b-1} dt\$

is the incomplete beta function and \$B(a, b)\$ is the *complete* beta function.

func BiasAdd ¶ added in v1.1.0

func BiasAdd(scope *Scope, value tf.Output, bias tf.Output, optional ...BiasAddAttr) (output tf.Output)

Adds `bias` to `value`.

This is a special case of `tf.add` where `bias` is restricted to be 1-D. Broadcasting is supported, so `value` may have any number of dimensions.

Arguments:

value: Any number of dimensions.
bias: 1-D with size the last dimension of `value`.

Returns Broadcasted sum of `value` and `bias`.

func BiasAddGrad ¶ added in v1.1.0

func BiasAddGrad(scope *Scope, out_backprop tf.Output, optional ...BiasAddGradAttr) (output tf.Output)

The backward operation for "BiasAdd" on the "bias" tensor.

It accumulates all the values from out_backprop into the feature dimension. For NHWC data format, the feature dimension is the last. For NCHW data format, the feature dimension is the third-to-last.

Arguments:

out_backprop: Any number of dimensions.

Returns 1-D with size the feature dimension of `out_backprop`.

func BiasAddV1 ¶ added in v1.1.0

func BiasAddV1(scope *Scope, value tf.Output, bias tf.Output) (output tf.Output)

Adds `bias` to `value`.

This is a deprecated version of BiasAdd and will be soon removed.

This is a special case of `tf.add` where `bias` is restricted to be 1-D. Broadcasting is supported, so `value` may have any number of dimensions.

Arguments:

value: Any number of dimensions.
bias: 1-D with size the last dimension of `value`.

Returns Broadcasted sum of `value` and `bias`.

func Bincount ¶ added in v1.1.0

func Bincount(scope *Scope, arr tf.Output, size tf.Output, weights tf.Output) (bins tf.Output)

Counts the number of occurrences of each value in an integer array.

Outputs a vector with length `size` and the same dtype as `weights`. If `weights` are empty, then index `i` stores the number of times the value `i` is counted in `arr`. If `weights` are non-empty, then index `i` stores the sum of the value in `weights` at each index where the corresponding value in `arr` is `i`.

Values in `arr` outside of the range [0, size) are ignored.

Arguments:

arr: int32 `Tensor`.
size: non-negative int32 scalar `Tensor`.
weights: is an int32, int64, float32, or float64 `Tensor` with the same

shape as `arr`, or a length-0 `Tensor`, in which case it acts as all weights equal to 1.

Returns 1D `Tensor` with length equal to `size`. The counts or summed weights for each value in the range [0, size).

func Bitcast ¶ added in v1.1.0

func Bitcast(scope *Scope, input tf.Output, type_ tf.DataType) (output tf.Output)

Bitcasts a tensor from one type to another without copying data.

Given a tensor `input`, this operation returns a tensor that has the same buffer data as `input` with datatype `type`.

If the input datatype `T` is larger than the output datatype `type` then the shape changes from [...] to [..., sizeof(`T`)/sizeof(`type`)].

If `T` is smaller than `type`, the operator requires that the rightmost dimension be equal to sizeof(`type`)/sizeof(`T`). The shape then goes from [..., sizeof(`type`)/sizeof(`T`)] to [...].

tf.bitcast() and tf.cast() work differently when real dtype is casted as a complex dtype (e.g. tf.complex64 or tf.complex128) as tf.cast() make imaginary part 0 while tf.bitcast() gives module error. For example,

Example 1:

>>> a = [1., 2., 3.] >>> equality_bitcast = tf.bitcast(a, tf.complex128) Traceback (most recent call last): ... InvalidArgumentError: Cannot bitcast from 1 to 18 [Op:Bitcast] >>> equality_cast = tf.cast(a, tf.complex128) >>> print(equality_cast) tf.Tensor([1.+0.j 2.+0.j 3.+0.j], shape=(3,), dtype=complex128)

Example 2:

>>> tf.bitcast(tf.constant(0xffffffff, dtype=tf.uint32), tf.uint8) <tf.Tensor: shape=(4,), dtype=uint8, numpy=array([255, 255, 255, 255], dtype=uint8)>

Example 3:

>>> x = [1., 2., 3.] >>> y = [0., 2., 3.] >>> equality= tf.equal(x,y) >>> equality_cast = tf.cast(equality,tf.float32) >>> equality_bitcast = tf.bitcast(equality_cast,tf.uint8) >>> print(equality) tf.Tensor([False True True], shape=(3,), dtype=bool) >>> print(equality_cast) tf.Tensor([0. 1. 1.], shape=(3,), dtype=float32) >>> print(equality_bitcast) tf.Tensor(

[[  0   0   0   0]
 [  0   0 128  63]
 [  0   0 128  63]], shape=(3, 4), dtype=uint8)

*NOTE*: Bitcast is implemented as a low-level cast, so machines with different endian orderings will give different results. A copy from input buffer to output buffer is made on BE machines when types are of different sizes in order to get the same casting results as on LE machines.

func BitwiseAnd ¶ added in v1.3.0

func BitwiseAnd(scope *Scope, x tf.Output, y tf.Output) (z tf.Output)

Elementwise computes the bitwise AND of `x` and `y`.

The result will have those bits set, that are set in both `x` and `y`. The computation is performed on the underlying representations of `x` and `y`.

For example:

```python import tensorflow as tf from tensorflow.python.ops import bitwise_ops dtype_list = [tf.int8, tf.int16, tf.int32, tf.int64,

tf.uint8, tf.uint16, tf.uint32, tf.uint64]

for dtype in dtype_list:

lhs = tf.constant([0, 5, 3, 14], dtype=dtype)
rhs = tf.constant([5, 0, 7, 11], dtype=dtype)
exp = tf.constant([0, 0, 3, 10], dtype=tf.float32)

res = bitwise_ops.bitwise_and(lhs, rhs)
tf.assert_equal(tf.cast(res, tf.float32), exp) # TRUE

```

func BitwiseOr ¶ added in v1.3.0

func BitwiseOr(scope *Scope, x tf.Output, y tf.Output) (z tf.Output)

Elementwise computes the bitwise OR of `x` and `y`.

The result will have those bits set, that are set in `x`, `y` or both. The computation is performed on the underlying representations of `x` and `y`.

For example:

```python import tensorflow as tf from tensorflow.python.ops import bitwise_ops dtype_list = [tf.int8, tf.int16, tf.int32, tf.int64,

tf.uint8, tf.uint16, tf.uint32, tf.uint64]

for dtype in dtype_list:

lhs = tf.constant([0, 5, 3, 14], dtype=dtype)
rhs = tf.constant([5, 0, 7, 11], dtype=dtype)
exp = tf.constant([5, 5, 7, 15], dtype=tf.float32)

res = bitwise_ops.bitwise_or(lhs, rhs)
tf.assert_equal(tf.cast(res,  tf.float32), exp)  # TRUE

```

func BitwiseXor ¶ added in v1.3.0

func BitwiseXor(scope *Scope, x tf.Output, y tf.Output) (z tf.Output)

Elementwise computes the bitwise XOR of `x` and `y`.

The result will have those bits set, that are different in `x` and `y`. The computation is performed on the underlying representations of `x` and `y`.

For example:

```python import tensorflow as tf from tensorflow.python.ops import bitwise_ops dtype_list = [tf.int8, tf.int16, tf.int32, tf.int64,

tf.uint8, tf.uint16, tf.uint32, tf.uint64]

for dtype in dtype_list:

lhs = tf.constant([0, 5, 3, 14], dtype=dtype)
rhs = tf.constant([5, 0, 7, 11], dtype=dtype)
exp = tf.constant([5, 5, 4, 5],  dtype=tf.float32)

res = bitwise_ops.bitwise_xor(lhs, rhs)
tf.assert_equal(tf.cast(res, tf.float32), exp) # TRUE

```

func BlockLSTM ¶ added in v1.15.0

func BlockLSTM(scope *Scope, seq_len_max tf.Output, x tf.Output, cs_prev tf.Output, h_prev tf.Output, w tf.Output, wci tf.Output, wcf tf.Output, wco tf.Output, b tf.Output, optional ...BlockLSTMAttr) (i tf.Output, cs tf.Output, f tf.Output, o tf.Output, ci tf.Output, co tf.Output, h tf.Output)

Computes the LSTM cell forward propagation for all the time steps.

This is equivalent to applying LSTMBlockCell in a loop, like so:

```python for x1 in unpack(x):

i1, cs1, f1, o1, ci1, co1, h1 = LSTMBlock(
  x1, cs_prev, h_prev, w, wci, wcf, wco, b)
cs_prev = cs1
h_prev = h1
i.append(i1)
cs.append(cs1)
f.append(f1)
o.append(o1)
ci.append(ci1)
co.append(co1)
h.append(h1)

return pack(i), pack(cs), pack(f), pack(o), pack(ci), pack(ch), pack(h) ```

Arguments:

seq_len_max: Maximum time length actually used by this input. Outputs are padded

with zeros beyond this length.

x: The sequence input to the LSTM, shape (timelen, batch_size, num_inputs).
cs_prev: Value of the initial cell state.
h_prev: Initial output of cell (to be used for peephole).
w: The weight matrix.
wci: The weight matrix for input gate peephole connection.
wcf: The weight matrix for forget gate peephole connection.
wco: The weight matrix for output gate peephole connection.
b: The bias vector.

Returns:

i: The input gate over the whole time sequence.
cs: The cell state before the tanh over the whole time sequence.
f: The forget gate over the whole time sequence.
o: The output gate over the whole time sequence.
ci: The cell input over the whole time sequence.
co: The cell after the tanh over the whole time sequence.
h: The output h vector over the whole time sequence.

func BlockLSTMGrad ¶ added in v1.15.0

func BlockLSTMGrad(scope *Scope, seq_len_max tf.Output, x tf.Output, cs_prev tf.Output, h_prev tf.Output, w tf.Output, wci tf.Output, wcf tf.Output, wco tf.Output, b tf.Output, i tf.Output, cs tf.Output, f tf.Output, o tf.Output, ci tf.Output, co tf.Output, h tf.Output, cs_grad tf.Output, h_grad tf.Output, use_peephole bool) (x_grad tf.Output, cs_prev_grad tf.Output, h_prev_grad tf.Output, w_grad tf.Output, wci_grad tf.Output, wcf_grad tf.Output, wco_grad tf.Output, b_grad tf.Output)

Computes the LSTM cell backward propagation for the entire time sequence.

This implementation is to be used in conjunction of LSTMBlock.

Arguments:

seq_len_max: Maximum time length actually used by this input. Outputs are padded

with zeros beyond this length.

x: The sequence input to the LSTM, shape (timelen, batch_size, num_inputs).
cs_prev: Value of the initial cell state.
h_prev: Initial output of cell (to be used for peephole).
w: The weight matrix.
wci: The weight matrix for input gate peephole connection.
wcf: The weight matrix for forget gate peephole connection.
wco: The weight matrix for output gate peephole connection.
b: The bias vector.
i: The input gate over the whole time sequence.
cs: The cell state before the tanh over the whole time sequence.
f: The forget gate over the whole time sequence.
o: The output gate over the whole time sequence.
ci: The cell input over the whole time sequence.
co: The cell after the tanh over the whole time sequence.
h: The output h vector over the whole time sequence.
cs_grad: The current gradient of cs.
h_grad: The gradient of h vector.
use_peephole: Whether to use peephole weights.

Returns:

x_grad: The gradient of x to be back-propped.
cs_prev_grad: The gradient of cs_prev to be back-propped.
h_prev_grad: The gradient of h_prev to be back-propped.
w_grad: The gradient for w to be back-propped.
wci_grad: The gradient for wci to be back-propped.
wcf_grad: The gradient for wcf to be back-propped.
wco_grad: The gradient for wco to be back-propped.
b_grad: The gradient for w to be back-propped.

func BlockLSTMGradV2 ¶

func BlockLSTMGradV2(scope *Scope, seq_len_max tf.Output, x tf.Output, cs_prev tf.Output, h_prev tf.Output, w tf.Output, wci tf.Output, wcf tf.Output, wco tf.Output, b tf.Output, i tf.Output, cs tf.Output, f tf.Output, o tf.Output, ci tf.Output, co tf.Output, h tf.Output, cs_grad tf.Output, h_grad tf.Output, use_peephole bool) (x_grad tf.Output, cs_prev_grad tf.Output, h_prev_grad tf.Output, w_grad tf.Output, wci_grad tf.Output, wcf_grad tf.Output, wco_grad tf.Output, b_grad tf.Output)

Computes the LSTM cell backward propagation for the entire time sequence.

This implementation is to be used in conjunction of BlockLSTMV2.

Arguments:

seq_len_max: Maximum time length actually used by this input. Outputs are padded

with zeros beyond this length.

x: The sequence input to the LSTM, shape (timelen, batch_size, num_inputs).
cs_prev: Value of the initial cell state.
h_prev: Initial output of cell (to be used for peephole).
w: The weight matrix.
wci: The weight matrix for input gate peephole connection.
wcf: The weight matrix for forget gate peephole connection.
wco: The weight matrix for output gate peephole connection.
b: The bias vector.
i: The input gate over the whole time sequence.
cs: The cell state before the tanh over the whole time sequence.
f: The forget gate over the whole time sequence.
o: The output gate over the whole time sequence.
ci: The cell input over the whole time sequence.
co: The cell after the tanh over the whole time sequence.
h: The output h vector over the whole time sequence.
cs_grad: The current gradient of cs.
h_grad: The gradient of h vector.
use_peephole: Whether to use peephole weights.

Returns:

x_grad: The gradient of x to be back-propped.
cs_prev_grad: The gradient of cs_prev to be back-propped.
h_prev_grad: The gradient of h_prev to be back-propped.
w_grad: The gradient for w to be back-propped.
wci_grad: The gradient for wci to be back-propped.
wcf_grad: The gradient for wcf to be back-propped.
wco_grad: The gradient for wco to be back-propped.
b_grad: The gradient for w to be back-propped.

func BlockLSTMV2 ¶

func BlockLSTMV2(scope *Scope, seq_len_max tf.Output, x tf.Output, cs_prev tf.Output, h_prev tf.Output, w tf.Output, wci tf.Output, wcf tf.Output, wco tf.Output, b tf.Output, optional ...BlockLSTMV2Attr) (i tf.Output, cs tf.Output, f tf.Output, o tf.Output, ci tf.Output, co tf.Output, h tf.Output)

Computes the LSTM cell forward propagation for all the time steps.

This is equivalent to applying LSTMBlockCell in a loop, like so:

```python for x1 in unpack(x):

i1, cs1, f1, o1, ci1, co1, h1 = LSTMBlock(
  x1, cs_prev, h_prev, w, wci, wcf, wco, b)
cs_prev = cs1
h_prev = h1
i.append(i1)
cs.append(cs1)
f.append(f1)
o.append(o1)
ci.append(ci1)
co.append(co1)
h.append(h1)

return pack(i), pack(cs), pack(f), pack(o), pack(ci), pack(ch), pack(h)

Note that unlike LSTMBlockCell (and BlockLSTM) which uses ICFO gate layout, this op uses IFCO. So in order for the following snippet to be equivalent all gate-related outputs should be reordered. ```

Arguments:

seq_len_max: Maximum time length actually used by this input. Outputs are padded

with zeros beyond this length.

x: The sequence input to the LSTM, shape (timelen, batch_size, num_inputs).
cs_prev: Value of the initial cell state.
h_prev: Initial output of cell (to be used for peephole).
w: The weight matrix.
wci: The weight matrix for input gate peephole connection.
wcf: The weight matrix for forget gate peephole connection.
wco: The weight matrix for output gate peephole connection.
b: The bias vector.

Returns:

i: The input gate over the whole time sequence.
cs: The cell state before the tanh over the whole time sequence.
f: The forget gate over the whole time sequence.
o: The output gate over the whole time sequence.
ci: The cell input over the whole time sequence.
co: The cell after the tanh over the whole time sequence.
h: The output h vector over the whole time sequence.

func BoostedTreesAggregateStats ¶ added in v1.12.1

func BoostedTreesAggregateStats(scope *Scope, node_ids tf.Output, gradients tf.Output, hessians tf.Output, feature tf.Output, max_splits int64, num_buckets int64) (stats_summary tf.Output)

Aggregates the summary of accumulated stats for the batch.

The summary stats contains gradients and hessians accumulated for each node, feature dimension id and bucket.

Arguments:

node_ids: int32; Rank 1 Tensor containing node ids for each example, shape [batch_size].
gradients: float32; Rank 2 Tensor (shape=[batch_size, logits_dimension]) with gradients for each example.
hessians: float32; Rank 2 Tensor (shape=[batch_size, hessian_dimension]) with hessians for each example.
feature: int32; Rank 2 feature Tensors (shape=[batch_size, feature_dimension]).
max_splits: int; the maximum number of splits possible in the whole tree.
num_buckets: int; equals to the maximum possible value of bucketized feature.

Returns output Rank 4 Tensor (shape=[splits, feature_dimension, buckets, logits_dimension + hessian_dimension]) containing accumulated stats for each node, feature dimension and bucket.

func BoostedTreesBucketize ¶ added in v1.12.0

func BoostedTreesBucketize(scope *Scope, float_values []tf.Output, bucket_boundaries []tf.Output) (buckets []tf.Output)

Bucketize each feature based on bucket boundaries.

An op that returns a list of float tensors, where each tensor represents the bucketized values for a single feature.

Arguments:

float_values: float; List of Rank 1 Tensor each containing float values for a single feature.
bucket_boundaries: float; List of Rank 1 Tensors each containing the bucket boundaries for a single

feature.

Returns int; List of Rank 1 Tensors each containing the bucketized values for a single feature.

func BoostedTreesCalculateBestFeatureSplit ¶ added in v1.12.1

func BoostedTreesCalculateBestFeatureSplit(scope *Scope, node_id_range tf.Output, stats_summary tf.Output, l1 tf.Output, l2 tf.Output, tree_complexity tf.Output, min_node_weight tf.Output, logits_dimension int64, optional ...BoostedTreesCalculateBestFeatureSplitAttr) (node_ids tf.Output, gains tf.Output, feature_dimensions tf.Output, thresholds tf.Output, left_node_contribs tf.Output, right_node_contribs tf.Output, split_with_default_directions tf.Output)

Calculates gains for each feature and returns the best possible split information for the feature.

The split information is the best threshold (bucket id), gains and left/right node contributions per node for each feature.

It is possible that not all nodes can be split on each feature. Hence, the list of possible nodes can differ between the features. Therefore, we return `node_ids_list` for each feature, containing the list of nodes that this feature can be used to split.

In this manner, the output is the best split per features and per node, so that it needs to be combined later to produce the best split for each node (among all possible features).

The output shapes are compatible in a way that the first dimension of all tensors are the same and equal to the number of possible split nodes for each feature.

Arguments:

node_id_range: A Rank 1 tensor (shape=[2]) to specify the range [first, last) of node ids to process within `stats_summary_list`. The nodes are iterated between the two nodes specified by the tensor, as like `for node_id in range(node_id_range[0], node_id_range[1])` (Note that the last index node_id_range[1] is exclusive).
stats_summary: A Rank 4 tensor (#shape=[max_splits, feature_dims, bucket, stats_dims]) for accumulated stats summary (gradient/hessian) per node, per dimension, per buckets for each feature.

The first dimension of the tensor is the maximum number of splits, and thus not all elements of it will be used, but only the indexes specified by node_ids will be used.

l1: l1 regularization factor on leaf weights, per instance based.
l2: l2 regularization factor on leaf weights, per instance based.
tree_complexity: adjustment to the gain, per leaf based.
min_node_weight: minimum avg of hessians in a node before required for the node to be considered for splitting.
logits_dimension: The dimension of logit, i.e., number of classes.

Returns:

node_ids: A Rank 1 tensors indicating possible split node ids for each feature. The length of the list is num_features, but each tensor has different size as each feature provides different possible nodes. See above for details like shapes and sizes.
gains: A Rank 1 tensors indicating the best gains for each feature to split for certain nodes. See above for details like shapes and sizes.
feature_dimensions: A Rank 1 tensors indicating the best feature dimension for each feature to split for certain nodes if the feature is multi-dimension. See above for details like shapes and sizes.
thresholds: A Rank 1 tensors indicating the bucket id to compare with (as a threshold) for split in each node. See above for details like shapes and sizes.
left_node_contribs: A Rank 2 tensors indicating the contribution of the left nodes when branching from parent nodes (given by the tensor element in the output node_ids_list) to the left direction by the given threshold for each feature. This value will be used to make the left node value by adding to the parent node value. Second dimension size is 1 for 1-dimensional logits, but would be larger for multi-class problems. See above for details like shapes and sizes.
right_node_contribs: A Rank 2 tensors, with the same shape/conditions as left_node_contribs_list, but just that the value is for the right node.
split_with_default_directions: A Rank 1 tensors indicating the which direction to go if data is missing. See above for details like shapes and sizes.

Inequality with default left returns 0, inequality with default right returns 1, equality with default right returns 2.

func BoostedTreesCalculateBestFeatureSplitV2 ¶

func BoostedTreesCalculateBestFeatureSplitV2(scope *Scope, node_id_range tf.Output, stats_summaries_list []tf.Output, split_types tf.Output, candidate_feature_ids tf.Output, l1 tf.Output, l2 tf.Output, tree_complexity tf.Output, min_node_weight tf.Output, logits_dimension int64) (node_ids tf.Output, gains tf.Output, feature_ids tf.Output, feature_dimensions tf.Output, thresholds tf.Output, left_node_contribs tf.Output, right_node_contribs tf.Output, split_with_default_directions tf.Output)

Calculates gains for each feature and returns the best possible split information for each node. However, if no split is found, then no split information is returned for that node.

The split information is the best threshold (bucket id), gains and left/right node contributions per node for each feature.

It is possible that not all nodes can be split on each feature. Hence, the list of possible nodes can differ between the features. Therefore, we return `node_ids_list` for each feature, containing the list of nodes that this feature can be used to split.

In this manner, the output is the best split per features and per node, so that it needs to be combined later to produce the best split for each node (among all possible features).

The output shapes are compatible in a way that the first dimension of all tensors are the same and equal to the number of possible split nodes for each feature.

Arguments:

node_id_range: A Rank 1 tensor (shape=[2]) to specify the range [first, last) of node ids to process within `stats_summary_list`. The nodes are iterated between the two nodes specified by the tensor, as like `for node_id in range(node_id_range[0], node_id_range[1])` (Note that the last index node_id_range[1] is exclusive).
stats_summaries_list: A list of Rank 4 tensor (#shape=[max_splits, feature_dims, bucket, stats_dims]) for accumulated stats summary (gradient/hessian) per node, per dimension, per buckets for each feature.

The first dimension of the tensor is the maximum number of splits, and thus not all elements of it will be used, but only the indexes specified by node_ids will be used.

split_types: A Rank 1 tensor indicating if this Op should perform inequality split or equality split per feature.
candidate_feature_ids: Rank 1 tensor with ids for each feature. This is the real id of the feature.
l1: l1 regularization factor on leaf weights, per instance based.
l2: l2 regularization factor on leaf weights, per instance based.
tree_complexity: adjustment to the gain, per leaf based.
min_node_weight: minimum avg of hessians in a node before required for the node to be considered for splitting.
logits_dimension: The dimension of logit, i.e., number of classes.

Returns:

node_ids: A Rank 1 tensors indicating possible split node ids for each feature. The length of the list is num_features, but each tensor has different size as each feature provides different possible nodes. See above for details like shapes and sizes.
gains: A Rank 1 tensor indicating the best gains for each feature to split for certain nodes. See above for details like shapes and sizes.
feature_ids: A Rank 1 tensors indicating the best feature id for each node. See above for details like shapes and sizes.
feature_dimensions: A Rank 1 tensors indicating the best feature dimension for each feature to split for certain nodes if the feature is multi-dimension. See above for details like shapes and sizes.
thresholds: A Rank 1 tensors indicating the bucket id to compare with (as a threshold) for split in each node. See above for details like shapes and sizes.
left_node_contribs: A Rank 2 tensors indicating the contribution of the left nodes when branching from parent nodes (given by the tensor element in the output node_ids_list) to the left direction by the given threshold for each feature. This value will be used to make the left node value by adding to the parent node value. Second dimension size is 1 for 1-dimensional logits, but would be larger for multi-class problems. See above for details like shapes and sizes.
right_node_contribs: A Rank 2 tensors, with the same shape/conditions as left_node_contribs_list, but just that the value is for the right node.
split_with_default_directions: A Rank 1 tensors indicating the which direction to go if data is missing. See above for details like shapes and sizes.

Inequality with default left returns 0, inequality with default right returns 1, equality with default right returns 2.

func BoostedTreesCalculateBestGainsPerFeature ¶ added in v1.9.0

func BoostedTreesCalculateBestGainsPerFeature(scope *Scope, node_id_range tf.Output, stats_summary_list []tf.Output, l1 tf.Output, l2 tf.Output, tree_complexity tf.Output, min_node_weight tf.Output, max_splits int64) (node_ids_list []tf.Output, gains_list []tf.Output, thresholds_list []tf.Output, left_node_contribs_list []tf.Output, right_node_contribs_list []tf.Output)

Calculates gains for each feature and returns the best possible split information for the feature.

The split information is the best threshold (bucket id), gains and left/right node contributions per node for each feature.

It is possible that not all nodes can be split on each feature. Hence, the list of possible nodes can differ between the features. Therefore, we return `node_ids_list` for each feature, containing the list of nodes that this feature can be used to split.

In this manner, the output is the best split per features and per node, so that it needs to be combined later to produce the best split for each node (among all possible features).

The length of output lists are all of the same length, `num_features`. The output shapes are compatible in a way that the first dimension of all tensors of all lists are the same and equal to the number of possible split nodes for each feature.

Arguments:

node_id_range: A Rank 1 tensor (shape=[2]) to specify the range [first, last) of node ids to process within `stats_summary_list`. The nodes are iterated between the two nodes specified by the tensor, as like `for node_id in range(node_id_range[0], node_id_range[1])` (Note that the last index node_id_range[1] is exclusive).
stats_summary_list: A list of Rank 3 tensor (#shape=[max_splits, bucket, 2]) for accumulated stats summary (gradient/hessian) per node per buckets for each feature. The first dimension of the tensor is the maximum number of splits, and thus not all elements of it will be used, but only the indexes specified by node_ids will be used.
l1: l1 regularization factor on leaf weights, per instance based.
l2: l2 regularization factor on leaf weights, per instance based.
tree_complexity: adjustment to the gain, per leaf based.
min_node_weight: minimum avg of hessians in a node before required for the node to be considered for splitting.
max_splits: the number of nodes that can be split in the whole tree. Used as a dimension of output tensors.

Returns:

node_ids_list: An output list of Rank 1 tensors indicating possible split node ids for each feature. The length of the list is num_features, but each tensor has different size as each feature provides different possible nodes. See above for details like shapes and sizes.
gains_list: An output list of Rank 1 tensors indicating the best gains for each feature to split for certain nodes. See above for details like shapes and sizes.
thresholds_list: An output list of Rank 1 tensors indicating the bucket id to compare with (as a threshold) for split in each node. See above for details like shapes and sizes.
left_node_contribs_list: A list of Rank 2 tensors indicating the contribution of the left nodes when branching from parent nodes (given by the tensor element in the output node_ids_list) to the left direction by the given threshold for each feature. This value will be used to make the left node value by adding to the parent node value. Second dimension size is 1 for 1-dimensional logits, but would be larger for multi-class problems. See above for details like shapes and sizes.
right_node_contribs_list: A list of Rank 2 tensors, with the same shape/conditions as left_node_contribs_list, but just that the value is for the right node.

func BoostedTreesCenterBias ¶ added in v1.11.0

func BoostedTreesCenterBias(scope *Scope, tree_ensemble_handle tf.Output, mean_gradients tf.Output, mean_hessians tf.Output, l1 tf.Output, l2 tf.Output) (continue_centering tf.Output)

Calculates the prior from the training data (the bias) and fills in the first node with the logits' prior. Returns a boolean indicating whether to continue centering.

Arguments:

tree_ensemble_handle: Handle to the tree ensemble.
mean_gradients: A tensor with shape=[logits_dimension] with mean of gradients for a first node.
mean_hessians: A tensor with shape=[logits_dimension] mean of hessians for a first node.
l1: l1 regularization factor on leaf weights, per instance based.
l2: l2 regularization factor on leaf weights, per instance based.

Returns Bool, whether to continue bias centering.

func BoostedTreesCreateEnsemble ¶ added in v1.9.0

func BoostedTreesCreateEnsemble(scope *Scope, tree_ensemble_handle tf.Output, stamp_token tf.Output, tree_ensemble_serialized tf.Output) (o *tf.Operation)

Creates a tree ensemble model and returns a handle to it.

Arguments:

tree_ensemble_handle: Handle to the tree ensemble resource to be created.
stamp_token: Token to use as the initial value of the resource stamp.
tree_ensemble_serialized: Serialized proto of the tree ensemble.

Returns the created operation.

func BoostedTreesCreateQuantileStreamResource ¶ added in v1.12.0

func BoostedTreesCreateQuantileStreamResource(scope *Scope, quantile_stream_resource_handle tf.Output, epsilon tf.Output, num_streams tf.Output, optional ...BoostedTreesCreateQuantileStreamResourceAttr) (o *tf.Operation)

Create the Resource for Quantile Streams.

Arguments:

quantile_stream_resource_handle: resource; Handle to quantile stream resource.
epsilon: float; The required approximation error of the stream resource.
num_streams: int; The number of streams managed by the resource that shares the same epsilon.

Returns the created operation.

func BoostedTreesDeserializeEnsemble ¶ added in v1.9.0

func BoostedTreesDeserializeEnsemble(scope *Scope, tree_ensemble_handle tf.Output, stamp_token tf.Output, tree_ensemble_serialized tf.Output) (o *tf.Operation)

Deserializes a serialized tree ensemble config and replaces current tree

ensemble.

Arguments:

tree_ensemble_handle: Handle to the tree ensemble.
stamp_token: Token to use as the new value of the resource stamp.
tree_ensemble_serialized: Serialized proto of the ensemble.

Returns the created operation.

func BoostedTreesEnsembleResourceHandleOp ¶ added in v1.9.0

func BoostedTreesEnsembleResourceHandleOp(scope *Scope, optional ...BoostedTreesEnsembleResourceHandleOpAttr) (resource tf.Output)

Creates a handle to a BoostedTreesEnsembleResource

func BoostedTreesExampleDebugOutputs ¶ added in v1.11.0

func BoostedTreesExampleDebugOutputs(scope *Scope, tree_ensemble_handle tf.Output, bucketized_features []tf.Output, logits_dimension int64) (examples_debug_outputs_serialized tf.Output)

Debugging/model interpretability outputs for each example.

It traverses all the trees and computes debug metrics for individual examples, such as getting split feature ids and logits after each split along the decision path used to compute directional feature contributions.

Arguments:

bucketized_features: A list of rank 1 Tensors containing bucket id for each

feature.

logits_dimension: scalar, dimension of the logits, to be used for constructing the protos in

examples_debug_outputs_serialized.

Returns Output rank 1 Tensor containing a proto serialized as a string for each example.

func BoostedTreesFlushQuantileSummaries ¶

func BoostedTreesFlushQuantileSummaries(scope *Scope, quantile_stream_resource_handle tf.Output, num_features int64) (summaries []tf.Output)

Flush the quantile summaries from each quantile stream resource.

An op that outputs a list of quantile summaries of a quantile stream resource. Each summary Tensor is rank 2, containing summaries (value, weight, min_rank, max_rank) for a single feature.

Arguments:

quantile_stream_resource_handle: resource handle referring to a QuantileStreamResource.

func BoostedTreesGetEnsembleStates ¶ added in v1.9.0

func BoostedTreesGetEnsembleStates(scope *Scope, tree_ensemble_handle tf.Output) (stamp_token tf.Output, num_trees tf.Output, num_finalized_trees tf.Output, num_attempted_layers tf.Output, last_layer_nodes_range tf.Output)

Retrieves the tree ensemble resource stamp token, number of trees and growing statistics.

Arguments:

tree_ensemble_handle: Handle to the tree ensemble.

Returns:

stamp_token: Stamp token of the tree ensemble resource.
num_trees: The number of trees in the tree ensemble resource.
num_finalized_trees: The number of trees that were finished successfully.
num_attempted_layers: The number of layers we attempted to build (but not necessarily succeeded).
last_layer_nodes_range: Rank size 2 tensor that contains start and end ids of the nodes in the latest

layer.

func BoostedTreesMakeQuantileSummaries ¶ added in v1.12.0

func BoostedTreesMakeQuantileSummaries(scope *Scope, float_values []tf.Output, example_weights tf.Output, epsilon tf.Output) (summaries []tf.Output)

Makes the summary of quantiles for the batch.

An op that takes a list of tensors (one tensor per feature) and outputs the quantile summaries for each tensor.

Arguments:

float_values: float; List of Rank 1 Tensors each containing values for a single feature.
example_weights: float; Rank 1 Tensor with weights per instance.
epsilon: float; The required maximum approximation error.

Returns float; List of Rank 2 Tensors each containing the quantile summary (value, weight, min_rank, max_rank) of a single feature.

func BoostedTreesMakeStatsSummary ¶ added in v1.9.0

func BoostedTreesMakeStatsSummary(scope *Scope, node_ids tf.Output, gradients tf.Output, hessians tf.Output, bucketized_features_list []tf.Output, max_splits int64, num_buckets int64) (stats_summary tf.Output)

Makes the summary of accumulated stats for the batch.

The summary stats contains gradients and hessians accumulated into the corresponding node and bucket for each example.

Arguments:

node_ids: int32 Rank 1 Tensor containing node ids, which each example falls into for the requested layer.
gradients: float32; Rank 2 Tensor (shape=[#examples, 1]) for gradients.
hessians: float32; Rank 2 Tensor (shape=[#examples, 1]) for hessians.
bucketized_features_list: int32 list of Rank 1 Tensors, each containing the bucketized feature (for each feature column).
max_splits: int; the maximum number of splits possible in the whole tree.
num_buckets: int; equals to the maximum possible value of bucketized feature.

Returns output Rank 4 Tensor (shape=[#features, #splits, #buckets, 2]) containing accumulated stats put into the corresponding node and bucket. The first index of 4th dimension refers to gradients, and the second to hessians.

func BoostedTreesPredict ¶ added in v1.9.0

func BoostedTreesPredict(scope *Scope, tree_ensemble_handle tf.Output, bucketized_features []tf.Output, logits_dimension int64) (logits tf.Output)

Runs multiple additive regression ensemble predictors on input instances and

computes the logits. It is designed to be used during prediction. It traverses all the trees and calculates the final score for each instance.

Arguments:

bucketized_features: A list of rank 1 Tensors containing bucket id for each

feature.

logits_dimension: scalar, dimension of the logits, to be used for partial logits

shape.

Returns Output rank 2 Tensor containing logits for each example.

func BoostedTreesQuantileStreamResourceAddSummaries ¶ added in v1.12.0

func BoostedTreesQuantileStreamResourceAddSummaries(scope *Scope, quantile_stream_resource_handle tf.Output, summaries []tf.Output) (o *tf.Operation)

Add the quantile summaries to each quantile stream resource.

An op that adds a list of quantile summaries to a quantile stream resource. Each summary Tensor is rank 2, containing summaries (value, weight, min_rank, max_rank) for a single feature.

Arguments:

quantile_stream_resource_handle: resource handle referring to a QuantileStreamResource.
summaries: string; List of Rank 2 Tensor each containing the summaries for a single feature.

Returns the created operation.

func BoostedTreesQuantileStreamResourceDeserialize ¶ added in v1.12.1

func BoostedTreesQuantileStreamResourceDeserialize(scope *Scope, quantile_stream_resource_handle tf.Output, bucket_boundaries []tf.Output) (o *tf.Operation)

Deserialize bucket boundaries and ready flag into current QuantileAccumulator.

An op that deserializes bucket boundaries and are boundaries ready flag into current QuantileAccumulator.

Arguments:

quantile_stream_resource_handle: resource handle referring to a QuantileStreamResource.
bucket_boundaries: float; List of Rank 1 Tensors each containing the bucket boundaries for a feature.

Returns the created operation.

func BoostedTreesQuantileStreamResourceFlush ¶ added in v1.12.0

func BoostedTreesQuantileStreamResourceFlush(scope *Scope, quantile_stream_resource_handle tf.Output, num_buckets tf.Output, optional ...BoostedTreesQuantileStreamResourceFlushAttr) (o *tf.Operation)

Flush the summaries for a quantile stream resource.

An op that flushes the summaries for a quantile stream resource.

Arguments:

quantile_stream_resource_handle: resource handle referring to a QuantileStreamResource.
num_buckets: int; approximate number of buckets unless using generate_quantiles.

Returns the created operation.

func BoostedTreesQuantileStreamResourceGetBucketBoundaries ¶ added in v1.12.0

func BoostedTreesQuantileStreamResourceGetBucketBoundaries(scope *Scope, quantile_stream_resource_handle tf.Output, num_features int64) (bucket_boundaries []tf.Output)

Generate the bucket boundaries for each feature based on accumulated summaries.

An op that returns a list of float tensors for a quantile stream resource. Each tensor is Rank 1 containing bucket boundaries for a single feature.

Arguments:

quantile_stream_resource_handle: resource handle referring to a QuantileStreamResource.
num_features: inferred int; number of features to get bucket boundaries for.

Returns float; List of Rank 1 Tensors each containing the bucket boundaries for a feature.

func BoostedTreesQuantileStreamResourceHandleOp ¶ added in v1.12.0

func BoostedTreesQuantileStreamResourceHandleOp(scope *Scope, optional ...BoostedTreesQuantileStreamResourceHandleOpAttr) (resource tf.Output)

Creates a handle to a BoostedTreesQuantileStreamResource.

func BoostedTreesSerializeEnsemble ¶ added in v1.9.0

func BoostedTreesSerializeEnsemble(scope *Scope, tree_ensemble_handle tf.Output) (stamp_token tf.Output, tree_ensemble_serialized tf.Output)

Serializes the tree ensemble to a proto.

Arguments:

tree_ensemble_handle: Handle to the tree ensemble.

Returns:

stamp_token: Stamp token of the tree ensemble resource.
tree_ensemble_serialized: Serialized proto of the ensemble.

func BoostedTreesSparseAggregateStats ¶ added in v1.15.0

func BoostedTreesSparseAggregateStats(scope *Scope, node_ids tf.Output, gradients tf.Output, hessians tf.Output, feature_indices tf.Output, feature_values tf.Output, feature_shape tf.Output, max_splits int64, num_buckets int64) (stats_summary_indices tf.Output, stats_summary_values tf.Output, stats_summary_shape tf.Output)

Aggregates the summary of accumulated stats for the batch.

The summary stats contains gradients and hessians accumulated for each node, bucket and dimension id.

Arguments:

node_ids: int32; Rank 1 Tensor containing node ids for each example, shape [batch_size].
gradients: float32; Rank 2 Tensor (shape=[batch_size, logits_dimension]) with gradients for each example.
hessians: float32; Rank 2 Tensor (shape=[batch_size, hessian_dimension]) with hessians for each example.
feature_indices: int32; Rank 2 indices of feature sparse Tensors (shape=[number of sparse entries, 2]).

Number of sparse entries across all instances from the batch. The first value is the index of the instance, the second is dimension of the feature. The second axis can only have 2 values, i.e., the input dense version of Tensor can only be matrix.

feature_values: int32; Rank 1 values of feature sparse Tensors (shape=[number of sparse entries]).

Number of sparse entries across all instances from the batch. The first value is the index of the instance, the second is dimension of the feature.

feature_shape: int32; Rank 1 dense shape of feature sparse Tensors (shape=[2]).

The first axis can only have 2 values, [batch_size, feature_dimension].

max_splits: int; the maximum number of splits possible in the whole tree.
num_buckets: int; equals to the maximum possible value of bucketized feature + 1.

Returns:

stats_summary_indices: int32; Rank 2 indices of summary sparse Tensors (shape=[number of non zero statistics, 4])

The second axis can only be 4 including node id, feature dimension, bucket id, and statistics_dimension. statistics_dimension = logits_dimension + hessian_dimension.

stats_summary_values: output Rank 1 Tensor (shape=[number of non zero statistics])
stats_summary_shape: output Rank 1 Tensor (shape=[4])

The tensor has following 4 values: [max_splits, feature_dimension, num_buckets, statistics_dimension], where statistics_dimension = gradient_dimension + hessian_dimension. gradient_dimension is the same as label_dimension, i.e., the output space. hessian_dimension can be the same as logits dimension when diagonal hessian is used, or label_dimension^2 when full hessian is used.

func BoostedTreesSparseCalculateBestFeatureSplit ¶ added in v1.15.0

func BoostedTreesSparseCalculateBestFeatureSplit(scope *Scope, node_id_range tf.Output, stats_summary_indices tf.Output, stats_summary_values tf.Output, stats_summary_shape tf.Output, l1 tf.Output, l2 tf.Output, tree_complexity tf.Output, min_node_weight tf.Output, logits_dimension int64, optional ...BoostedTreesSparseCalculateBestFeatureSplitAttr) (node_ids tf.Output, gains tf.Output, feature_dimensions tf.Output, thresholds tf.Output, left_node_contribs tf.Output, right_node_contribs tf.Output, split_with_default_directions tf.Output)

Calculates gains for each feature and returns the best possible split information for the feature.

The split information is the best threshold (bucket id), gains and left/right node contributions per node for each feature.

It is possible that not all nodes can be split on each feature. Hence, the list of possible nodes can differ between the features. Therefore, we return `node_ids_list` for each feature, containing the list of nodes that this feature can be used to split.

In this manner, the output is the best split per features and per node, so that it needs to be combined later to produce the best split for each node (among all possible features).

The output shapes are compatible in a way that the first dimension of all tensors are the same and equal to the number of possible split nodes for each feature.

Arguments:

node_id_range: A Rank 1 tensor (shape=[2]) to specify the range [first, last) of node ids to process within `stats_summary_list`. The nodes are iterated between the two nodes specified by the tensor, as like `for node_id in range(node_id_range[0], node_id_range[1])` (Note that the last index node_id_range[1] is exclusive).
stats_summary_indices: A Rank 2 int64 tensor of dense shape [N, 4] (N specifies the number of non-zero values) for accumulated stats summary (gradient/hessian) per node per bucket for each feature. The second dimension contains node id, feature dimension, bucket id, and stats dim.

stats dim is the sum of logits dimension and hessian dimension, hessian dimension can either be logits dimension if diagonal hessian is used, or logits dimension^2 if full hessian is used.

stats_summary_values: A Rank 1 float tensor of dense shape [N] (N specifies the number of non-zero values), which supplies the values for each element in summary_indices.
stats_summary_shape: A Rank 1 float tensor of dense shape [4], which specifies the dense shape of the sparse tensor, which is [num tree nodes, feature dimensions, num buckets, stats dim].
l1: l1 regularization factor on leaf weights, per instance based.
l2: l2 regularization factor on leaf weights, per instance based.
tree_complexity: adjustment to the gain, per leaf based.
min_node_weight: minimum avg of hessians in a node before required for the node to be considered for splitting.
logits_dimension: The dimension of logit, i.e., number of classes.

Returns:

node_ids: A Rank 1 tensor indicating possible node ids that can be split.
gains: A Rank 1 tensor indicating the best gains to split each node.
feature_dimensions: A Rank 1 tensor indicating the best feature dimension for each feature to split for each node.
thresholds: A Rank 1 tensor indicating the bucket id to compare with (as a threshold) for split in each node.
left_node_contribs: A Rank 2 tensor indicating the contribution of the left nodes when branching from parent nodes to the left direction by the given threshold for each feature.

This value will be used to make the left node value by adding to the parent node value. Second dimension size is logits dimension.

right_node_contribs: A Rank 2 tensor, with the same shape/conditions as left_node_contribs_list, but just that the value is for the right node.
split_with_default_directions: A Rank 1 tensor indicating which direction to go if data is missing.

Inequality with default left returns 0, inequality with default right returns 1, equality with default right returns 2.

func BoostedTreesTrainingPredict ¶ added in v1.9.0

func BoostedTreesTrainingPredict(scope *Scope, tree_ensemble_handle tf.Output, cached_tree_ids tf.Output, cached_node_ids tf.Output, bucketized_features []tf.Output, logits_dimension int64) (partial_logits tf.Output, tree_ids tf.Output, node_ids tf.Output)

Runs multiple additive regression ensemble predictors on input instances and

computes the update to cached logits. It is designed to be used during training. It traverses the trees starting from cached tree id and cached node id and calculates the updates to be pushed to the cache.

Arguments:

cached_tree_ids: Rank 1 Tensor containing cached tree ids which is the starting

tree of prediction.

cached_node_ids: Rank 1 Tensor containing cached node id which is the starting

node of prediction.

bucketized_features: A list of rank 1 Tensors containing bucket id for each

feature.

logits_dimension: scalar, dimension of the logits, to be used for partial logits

shape.

Returns:

partial_logits: Rank 2 Tensor containing logits update (with respect to cached

values stored) for each example.

tree_ids: Rank 1 Tensor containing new tree ids for each example.
node_ids: Rank 1 Tensor containing new node ids in the new tree_ids.

func BoostedTreesUpdateEnsemble ¶ added in v1.9.0

func BoostedTreesUpdateEnsemble(scope *Scope, tree_ensemble_handle tf.Output, feature_ids tf.Output, node_ids []tf.Output, gains []tf.Output, thresholds []tf.Output, left_node_contribs []tf.Output, right_node_contribs []tf.Output, max_depth tf.Output, learning_rate tf.Output, pruning_mode int64) (o *tf.Operation)

Updates the tree ensemble by either adding a layer to the last tree being grown

or by starting a new tree.

Arguments:

tree_ensemble_handle: Handle to the ensemble variable.
feature_ids: Rank 1 tensor with ids for each feature. This is the real id of

the feature that will be used in the split.

node_ids: List of rank 1 tensors representing the nodes for which this feature

has a split.

gains: List of rank 1 tensors representing the gains for each of the feature's

split.

thresholds: List of rank 1 tensors representing the thesholds for each of the

feature's split.

left_node_contribs: List of rank 2 tensors with left leaf contribs for each of

the feature's splits. Will be added to the previous node values to constitute the values of the left nodes.

right_node_contribs: List of rank 2 tensors with right leaf contribs for each

of the feature's splits. Will be added to the previous node values to constitute the values of the right nodes.

max_depth: Max depth of the tree to build.
learning_rate: shrinkage const for each new tree.
pruning_mode: 0-No pruning, 1-Pre-pruning, 2-Post-pruning.

Returns the created operation.

func BoostedTreesUpdateEnsembleV2 ¶

func BoostedTreesUpdateEnsembleV2(scope *Scope, tree_ensemble_handle tf.Output, feature_ids []tf.Output, dimension_ids []tf.Output, node_ids []tf.Output, gains []tf.Output, thresholds []tf.Output, left_node_contribs []tf.Output, right_node_contribs []tf.Output, split_types []tf.Output, max_depth tf.Output, learning_rate tf.Output, pruning_mode tf.Output, optional ...BoostedTreesUpdateEnsembleV2Attr) (o *tf.Operation)

Updates the tree ensemble by adding a layer to the last tree being grown

or by starting a new tree.

Arguments:

tree_ensemble_handle: Handle to the ensemble variable.
feature_ids: Rank 1 tensor with ids for each feature. This is the real id of

the feature that will be used in the split.

dimension_ids: List of rank 1 tensors representing the dimension in each feature.
node_ids: List of rank 1 tensors representing the nodes for which this feature

has a split.

gains: List of rank 1 tensors representing the gains for each of the feature's

split.

thresholds: List of rank 1 tensors representing the thesholds for each of the

feature's split.

left_node_contribs: List of rank 2 tensors with left leaf contribs for each of

the feature's splits. Will be added to the previous node values to constitute the values of the left nodes.

right_node_contribs: List of rank 2 tensors with right leaf contribs for each

of the feature's splits. Will be added to the previous node values to constitute the values of the right nodes.

split_types: List of rank 1 tensors representing the split type for each feature.
max_depth: Max depth of the tree to build.
learning_rate: shrinkage const for each new tree.
pruning_mode: 0-No pruning, 1-Pre-pruning, 2-Post-pruning.

Returns the created operation.

func BroadcastArgs ¶ added in v1.1.0

func BroadcastArgs(scope *Scope, s0 tf.Output, s1 tf.Output) (r0 tf.Output)

Return the shape of s0 op s1 with broadcast.

Given `s0` and `s1`, tensors that represent shapes, compute `r0`, the broadcasted shape. `s0`, `s1` and `r0` are all integer vectors.

func BroadcastGradientArgs ¶ added in v1.1.0

func BroadcastGradientArgs(scope *Scope, s0 tf.Output, s1 tf.Output) (r0 tf.Output, r1 tf.Output)

Return the reduction indices for computing gradients of s0 op s1 with broadcast.

This is typically used by gradient computations for a broadcasting operation.

func BroadcastTo ¶ added in v1.11.0

func BroadcastTo(scope *Scope, input tf.Output, shape tf.Output) (output tf.Output)

Broadcast an array for a compatible shape.

Broadcasting is the process of making arrays to have compatible shapes for arithmetic operations. Two shapes are compatible if for each dimension pair they are either equal or one of them is one.

For example:

>>> x = tf.constant([[1, 2, 3]]) # Shape (1, 3,) >>> y = tf.broadcast_to(x, [2, 3]) >>> print(y) tf.Tensor(

[[1 2 3]
 [1 2 3]], shape=(2, 3), dtype=int32)

In the above example, the input Tensor with the shape of `[1, 3]` is broadcasted to output Tensor with shape of `[2, 3]`.

When broadcasting, if a tensor has fewer axes than necessary its shape is padded on the left with ones. So this gives the same result as the previous example:

>>> x = tf.constant([1, 2, 3]) # Shape (3,) >>> y = tf.broadcast_to(x, [2, 3])

When doing broadcasted operations such as multiplying a tensor by a scalar, broadcasting (usually) confers some time or space benefit, as the broadcasted tensor is never materialized.

However, `broadcast_to` does not carry with it any such benefits. The newly-created tensor takes the full memory of the broadcasted shape. (In a graph context, `broadcast_to` might be fused to subsequent operation and then be optimized away, however.)

Arguments:

input: A Tensor to broadcast.
shape: An 1-D `int` Tensor. The shape of the desired output.

Returns A Tensor.

func Bucketize ¶ added in v1.2.0

func Bucketize(scope *Scope, input tf.Output, boundaries []float32) (output tf.Output)

Bucketizes 'input' based on 'boundaries'.

For example, if the inputs are

boundaries = [0, 10, 100]
input = [[-5, 10000]
         [150,   10]
         [5,    100]]

then the output will be

output = [[0, 3]
          [3, 2]
          [1, 3]]

Arguments:

input: Any shape of Tensor contains with int or float type.
boundaries: A sorted list of floats gives the boundary of the buckets.

Returns Same shape with 'input', each value of input replaced with bucket index.

@compatibility(numpy) Equivalent to np.digitize. @end_compatibility

func BytesProducedStatsDataset ¶ added in v1.5.0

func BytesProducedStatsDataset(scope *Scope, input_dataset tf.Output, tag tf.Output, output_types []tf.DataType, output_shapes []tf.Shape) (handle tf.Output)

Records the bytes size of each element of `input_dataset` in a StatsAggregator.

func CSRSparseMatrixComponents ¶

func CSRSparseMatrixComponents(scope *Scope, csr_sparse_matrix tf.Output, index tf.Output, type_ tf.DataType) (row_ptrs tf.Output, col_inds tf.Output, values tf.Output)

Reads out the CSR components at batch `index`.

This op is meant only for debugging / testing, and its interface is not expected to be stable.

Arguments:

csr_sparse_matrix: A batched CSRSparseMatrix.
index: The index in `csr_sparse_matrix`'s batch.

Returns:

row_ptrs: An array containing CSR matrix row pointers.
col_inds: An array containing CSR matrix column indices.
values: An array containing CSR matrix nonzero values.

func CSRSparseMatrixToDense ¶

func CSRSparseMatrixToDense(scope *Scope, sparse_input tf.Output, type_ tf.DataType) (dense_output tf.Output)

Convert a (possibly batched) CSRSparseMatrix to dense.

Arguments:

sparse_input: A batched CSRSparseMatrix.

Returns A dense tensor.

func CSRSparseMatrixToSparseTensor ¶

func CSRSparseMatrixToSparseTensor(scope *Scope, sparse_matrix tf.Output, type_ tf.DataType) (indices tf.Output, values tf.Output, dense_shape tf.Output)

Converts a (possibly batched) CSRSparesMatrix to a SparseTensor.

Arguments:

sparse_matrix: A (possibly batched) CSRSparseMatrix.

Returns:

indices: SparseTensor indices.
values: SparseTensor values.
dense_shape: SparseTensor dense shape.

func CTCBeamSearchDecoder ¶ added in v1.1.0

func CTCBeamSearchDecoder(scope *Scope, inputs tf.Output, sequence_length tf.Output, beam_width int64, top_paths int64, optional ...CTCBeamSearchDecoderAttr) (decoded_indices []tf.Output, decoded_values []tf.Output, decoded_shape []tf.Output, log_probability tf.Output)

Performs beam search decoding on the logits given in input.

A note about the attribute merge_repeated: For the beam search decoder, this means that if consecutive entries in a beam are the same, only the first of these is emitted. That is, when the top path is "A B B B B", "A B" is returned if merge_repeated = True but "A B B B B" is returned if merge_repeated = False.

Arguments:

inputs: 3-D, shape: `(max_time x batch_size x num_classes)`, the logits.
sequence_length: A vector containing sequence lengths, size `(batch)`.
beam_width: A scalar >= 0 (beam search beam width).
top_paths: A scalar >= 0, <= beam_width (controls output size).

Returns:

decoded_indices: A list (length: top_paths) of indices matrices.  Matrix j,

size `(total_decoded_outputs[j] x 2)`, has indices of a `SparseTensor<int64, 2>`. The rows store: [batch, time].

decoded_values: A list (length: top_paths) of values vectors.  Vector j,

size `(length total_decoded_outputs[j])`, has the values of a `SparseTensor<int64, 2>`. The vector stores the decoded classes for beam j.

decoded_shape: A list (length: top_paths) of shape vector.  Vector j,

size `(2)`, stores the shape of the decoded `SparseTensor[j]`. Its values are: `[batch_size, max_decoded_length[j]]`.

log_probability: A matrix, shaped: `(batch_size x top_paths)`.  The

sequence log-probabilities.

func CTCGreedyDecoder ¶ added in v1.1.0

func CTCGreedyDecoder(scope *Scope, inputs tf.Output, sequence_length tf.Output, optional ...CTCGreedyDecoderAttr) (decoded_indices tf.Output, decoded_values tf.Output, decoded_shape tf.Output, log_probability tf.Output)

Performs greedy decoding on the logits given in inputs.

A note about the attribute merge_repeated: if enabled, when consecutive logits' maximum indices are the same, only the first of these is emitted. Labeling the blank '*', the sequence "A B B * B B" becomes "A B B" if merge_repeated = True and "A B B B B" if merge_repeated = False.

Regardless of the value of merge_repeated, if the maximum index of a given time and batch corresponds to the blank, index `(num_classes - 1)`, no new element is emitted.

Arguments:

inputs: 3-D, shape: `(max_time x batch_size x num_classes)`, the logits.
sequence_length: A vector containing sequence lengths, size `(batch_size)`.

Returns:

decoded_indices: Indices matrix, size `(total_decoded_outputs x 2)`,

of a `SparseTensor<int64, 2>`. The rows store: [batch, time].

decoded_values: Values vector, size: `(total_decoded_outputs)`,

of a `SparseTensor<int64, 2>`. The vector stores the decoded classes.

decoded_shape: Shape vector, size `(2)`, of the decoded SparseTensor.

Values are: `[batch_size, max_decoded_length]`.

log_probability: Matrix, size `(batch_size x 1)`, containing sequence

log-probabilities.

func CTCLoss ¶ added in v1.1.0

func CTCLoss(scope *Scope, inputs tf.Output, labels_indices tf.Output, labels_values tf.Output, sequence_length tf.Output, optional ...CTCLossAttr) (loss tf.Output, gradient tf.Output)

Calculates the CTC Loss (log probability) for each batch entry. Also calculates

the gradient. This class performs the softmax operation for you, so inputs should be e.g. linear projections of outputs by an LSTM.

Arguments:

inputs: 3-D, shape: `(max_time x batch_size x num_classes)`, the logits.
labels_indices: The indices of a `SparseTensor<int32, 2>`.

`labels_indices(i, :) == [b, t]` means `labels_values(i)` stores the id for `(batch b, time t)`.

labels_values: The values (labels) associated with the given batch and time.
sequence_length: A vector containing sequence lengths (batch).

Returns:

loss: A vector (batch) containing log-probabilities.
gradient: The gradient of `loss`.  3-D, shape:

`(max_time x batch_size x num_classes)`.

func CTCLossV2 ¶

func CTCLossV2(scope *Scope, inputs tf.Output, labels_indices tf.Output, labels_values tf.Output, sequence_length tf.Output, optional ...CTCLossV2Attr) (loss tf.Output, gradient tf.Output)

Calculates the CTC Loss (log probability) for each batch entry. Also calculates

the gradient. This class performs the softmax operation for you, so inputs should be e.g. linear projections of outputs by an LSTM.

Arguments:

inputs: 3-D, shape: `(max_time x batch_size x num_classes)`, the logits. Default blank

label is 0 rather num_classes - 1.

labels_indices: The indices of a `SparseTensor<int32, 2>`.

`labels_indices(i, :) == [b, t]` means `labels_values(i)` stores the id for `(batch b, time t)`.

labels_values: The values (labels) associated with the given batch and time.
sequence_length: A vector containing sequence lengths (batch).

Returns:

loss: A vector (batch) containing log-probabilities.
gradient: The gradient of `loss`.  3-D, shape:

`(max_time x batch_size x num_classes)`.

func CacheDataset ¶ added in v1.3.0

func CacheDataset(scope *Scope, input_dataset tf.Output, filename tf.Output, output_types []tf.DataType, output_shapes []tf.Shape, optional ...CacheDatasetAttr) (handle tf.Output)

Creates a dataset that caches elements from `input_dataset`.

A CacheDataset will iterate over the input_dataset, and store tensors. If the cache already exists, the cache will be used. If the cache is inappropriate (e.g. cannot be opened, contains tensors of the wrong shape / size), an error will the returned when used.

Arguments:

filename: A path on the filesystem where we should cache the dataset. Note: this

will be a directory.

func Cast ¶ added in v1.1.0

func Cast(scope *Scope, x tf.Output, DstT tf.DataType, optional ...CastAttr) (y tf.Output)

Cast x of type SrcT to y of DstT.

func Ceil ¶ added in v1.1.0

func Ceil(scope *Scope, x tf.Output) (y tf.Output)

Returns element-wise smallest integer not less than x.

func CheckNumerics ¶ added in v1.1.0

func CheckNumerics(scope *Scope, tensor tf.Output, message string) (output tf.Output)

Checks a tensor for NaN and Inf values.

When run, reports an `InvalidArgument` error if `tensor` has any values that are not a number (NaN) or infinity (Inf). Otherwise, returns the input tensor.

Example usage:

``` python a = tf.Variable(1.0) tf.debugging.check_numerics(a, message=”)

b = tf.Variable(np.nan) try:

tf.debugging.check_numerics(b, message='Checking b')

except Exception as e:

assert "Checking b : Tensor had NaN values" in e.message

c = tf.Variable(np.inf) try:

tf.debugging.check_numerics(c, message='Checking c')

except Exception as e:

assert "Checking c : Tensor had Inf values" in e.message

```

Arguments:

message: Prefix of the error message.

func CheckNumericsV2 ¶

func CheckNumericsV2(scope *Scope, tensor tf.Output, message string) (output tf.Output)

Checks a tensor for NaN, -Inf and +Inf values.

When run, reports an `InvalidArgument` error if `tensor` has any values that are not a number (NaN) or infinity (Inf). Otherwise, returns the input tensor. Unlike CheckNumerics (V1), CheckNumericsV2 distinguishes -Inf and +Inf in the errors it throws.

Arguments:

message: Prefix of the error message.

func Cholesky ¶ added in v1.1.0

func Cholesky(scope *Scope, input tf.Output) (output tf.Output)

Computes the Cholesky decomposition of one or more square matrices.

The input is a tensor of shape `[..., M, M]` whose inner-most 2 dimensions form square matrices.

The input has to be symmetric and positive definite. Only the lower-triangular part of the input will be used for this operation. The upper-triangular part will not be read.

The output is a tensor of the same shape as the input containing the Cholesky decompositions for all input submatrices `[..., :, :]`.

**Note**: The gradient computation on GPU is faster for large matrices but not for large batch dimensions when the submatrices are small. In this case it might be faster to use the CPU.

Arguments:

input: Shape is `[..., M, M]`.

Returns Shape is `[..., M, M]`.

func CholeskyGrad ¶ added in v1.1.0

func CholeskyGrad(scope *Scope, l tf.Output, grad tf.Output) (output tf.Output)

Computes the reverse mode backpropagated gradient of the Cholesky algorithm.

For an explanation see "Differentiation of the Cholesky algorithm" by Iain Murray http://arxiv.org/abs/1602.07527.

Arguments:

l: Output of batch Cholesky algorithm l = cholesky(A). Shape is `[..., M, M]`.

Algorithm depends only on lower triangular part of the innermost matrices of this tensor.

grad: df/dl where f is some scalar function. Shape is `[..., M, M]`.

Algorithm depends only on lower triangular part of the innermost matrices of this tensor.

Returns Symmetrized version of df/dA . Shape is `[..., M, M]`

func ClipByValue ¶ added in v1.9.0

func ClipByValue(scope *Scope, t tf.Output, clip_value_min tf.Output, clip_value_max tf.Output) (output tf.Output)

Clips tensor values to a specified min and max.

Given a tensor `t`, this operation returns a tensor of the same type and shape as `t` with its values clipped to `clip_value_min` and `clip_value_max`. Any values less than `clip_value_min` are set to `clip_value_min`. Any values greater than `clip_value_max` are set to `clip_value_max`.

Arguments:

t: A `Tensor`.
clip_value_min: A 0-D (scalar) `Tensor`, or a `Tensor` with the same shape

as `t`. The minimum value to clip by.

clip_value_max: A 0-D (scalar) `Tensor`, or a `Tensor` with the same shape

as `t`. The maximum value to clip by.

Returns A clipped `Tensor` with the same shape as input 't'.

func CollateTPUEmbeddingMemory ¶

func CollateTPUEmbeddingMemory(scope *Scope, memory_configs []tf.Output) (merged_memory_config tf.Output)

An op that merges the string-encoded memory config protos from all hosts.

Arguments:

memory_configs: String-encoded memory config protos containing metadata about

the memory allocations reserved for TPUEmbedding across all hosts.

func CollectiveAllToAllV2 ¶

func CollectiveAllToAllV2(scope *Scope, input tf.Output, group_size tf.Output, group_key tf.Output, instance_key tf.Output, ordering_token []tf.Output, optional ...CollectiveAllToAllV2Attr) (data tf.Output)

Mutually exchanges multiple tensors of identical type and shape.

`is_stateless` means each op does not need control dependencies to other collective ops. In this case, keys that are unique at runtime (e.g. `instance_key`) should be used to distinguish collective groups.

func CollectiveAllToAllV3 ¶

func CollectiveAllToAllV3(scope *Scope, input tf.Output, communicator tf.Output, group_assignment tf.Output, optional ...CollectiveAllToAllV3Attr) (data tf.Output)

Mutually exchanges multiple tensors of identical type and shape.

func CollectiveAssignGroupV2 ¶

func CollectiveAssignGroupV2(scope *Scope, group_assignment tf.Output, device_index tf.Output, base_key tf.Output) (group_size tf.Output, group_key tf.Output)

Assign group keys based on group assignment.

func CollectiveBcastRecv ¶ added in v1.9.0

func CollectiveBcastRecv(scope *Scope, T tf.DataType, group_size int64, group_key int64, instance_key int64, shape tf.Shape, optional ...CollectiveBcastRecvAttr) (data tf.Output)

Receives a tensor value broadcast from another device.

func CollectiveBcastRecvV2 ¶

func CollectiveBcastRecvV2(scope *Scope, group_size tf.Output, group_key tf.Output, instance_key tf.Output, shape tf.Output, T tf.DataType, optional ...CollectiveBcastRecvV2Attr) (data tf.Output)

Receives a tensor value broadcast from another device.

func CollectiveBcastSend ¶ added in v1.9.0

func CollectiveBcastSend(scope *Scope, input tf.Output, group_size int64, group_key int64, instance_key int64, shape tf.Shape, optional ...CollectiveBcastSendAttr) (data tf.Output)

Broadcasts a tensor value to one or more other devices.

func CollectiveBcastSendV2 ¶

func CollectiveBcastSendV2(scope *Scope, input tf.Output, group_size tf.Output, group_key tf.Output, instance_key tf.Output, optional ...CollectiveBcastSendV2Attr) (data tf.Output)

Broadcasts a tensor value to one or more other devices.

func CollectiveGather ¶ added in v1.12.1

func CollectiveGather(scope *Scope, input tf.Output, group_size int64, group_key int64, instance_key int64, shape tf.Shape, optional ...CollectiveGatherAttr) (data tf.Output)

Mutually accumulates multiple tensors of identical type and shape.

func CollectiveGatherV2 ¶

func CollectiveGatherV2(scope *Scope, input tf.Output, group_size tf.Output, group_key tf.Output, instance_key tf.Output, ordering_token []tf.Output, optional ...CollectiveGatherV2Attr) (data tf.Output)

Mutually accumulates multiple tensors of identical type and shape.

`is_stateless` means each op does not need control dependencies to other collective ops. In this case, keys that are unique at runtime (e.g. `instance_key`) should be used to distinguish collective groups.

func CollectiveInitializeCommunicator ¶

func CollectiveInitializeCommunicator(scope *Scope, group_key tf.Output, rank tf.Output, group_size tf.Output, optional ...CollectiveInitializeCommunicatorAttr) (communicator tf.Output)

Initializes a group for collective operations.

func CollectivePermute ¶ added in v1.12.1

func CollectivePermute(scope *Scope, input tf.Output, source_target_pairs tf.Output) (output tf.Output)

An Op to permute tensors across replicated TPU instances.

Each instance supplies its own input.

For example, suppose there are 4 TPU instances: `[A, B, C, D]`. Passing source_target_pairs=`[[0,1],[1,2],[2,3],[3,0]]` gets the outputs: `[D, A, B, C]`.

Arguments:

input: The local input to be permuted. Currently only supports float and

bfloat16.

source_target_pairs: A tensor with shape [num_pairs, 2].

Returns The permuted input.

func CollectiveReduce ¶ added in v1.9.0

func CollectiveReduce(scope *Scope, input tf.Output, group_size int64, group_key int64, instance_key int64, merge_op string, final_op string, subdiv_offsets []int64, optional ...CollectiveReduceAttr) (data tf.Output)

Mutually reduces multiple tensors of identical type and shape.

func CollectiveReduceScatterV2 ¶

func CollectiveReduceScatterV2(scope *Scope, input tf.Output, group_size tf.Output, group_key tf.Output, instance_key tf.Output, ordering_token []tf.Output, merge_op string, final_op string, optional ...CollectiveReduceScatterV2Attr) (data tf.Output)

Mutually reduces multiple tensors of identical type and shape and scatters the result.

`is_stateless` means each op does not need control dependencies to other collective ops. In this case, keys that are unique at runtime (e.g. `instance_key`) should be used to distinguish collective groups.

func CollectiveReduceV2 ¶

func CollectiveReduceV2(scope *Scope, input tf.Output, group_size tf.Output, group_key tf.Output, instance_key tf.Output, ordering_token []tf.Output, merge_op string, final_op string, optional ...CollectiveReduceV2Attr) (data tf.Output)

Mutually reduces multiple tensors of identical type and shape.

`is_stateless` means each op does not need control dependencies to other collective ops. In this case, keys that are unique at runtime (e.g. `instance_key`) should be used to distinguish collective groups.

func CollectiveReduceV3 ¶

func CollectiveReduceV3(scope *Scope, input tf.Output, communicator tf.Output, group_assignment tf.Output, reduction string, optional ...CollectiveReduceV3Attr) (data tf.Output)

Mutually reduces multiple tensors of identical type and shape.

func CombinedNonMaxSuppression ¶ added in v1.12.1

func CombinedNonMaxSuppression(scope *Scope, boxes tf.Output, scores tf.Output, max_output_size_per_class tf.Output, max_total_size tf.Output, iou_threshold tf.Output, score_threshold tf.Output, optional ...CombinedNonMaxSuppressionAttr) (nmsed_boxes tf.Output, nmsed_scores tf.Output, nmsed_classes tf.Output, valid_detections tf.Output)

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. Prunes away boxes that have high intersection-over-union (IOU) overlap with previously selected boxes. Bounding boxes are supplied as [y1, x1, y2, x2], where (y1, x1) and (y2, x2) are the coordinates of any diagonal pair of box corners and the coordinates can be provided as normalized (i.e., lying in the interval [0, 1]) or absolute. Note that this algorithm is agnostic to where the origin is in the coordinate system. Also note that this algorithm is invariant to orthogonal transformations and translations of the coordinate system; thus translating or reflections of the coordinate system result in the same boxes being selected by the algorithm. The output of this operation is the final boxes, scores and classes tensor returned after performing non_max_suppression.

Arguments:

boxes: A 4-D float tensor of shape `[batch_size, num_boxes, q, 4]`. If `q` is 1 then

same boxes are used for all classes otherwise, if `q` is equal to number of classes, class-specific boxes are used.

scores: A 3-D float tensor of shape `[batch_size, num_boxes, num_classes]`

representing a single score corresponding to each box (each row of boxes).

max_output_size_per_class: A scalar integer tensor representing the maximum number of

boxes to be selected by non max suppression per class

max_total_size: An int32 scalar representing the maximum number of boxes retained over all

classes. Note that setting this value to a large number may result in OOM error depending on the system workload.

iou_threshold: A 0-D float tensor representing the threshold for deciding whether

boxes overlap too much with respect to IOU.

score_threshold: A 0-D float tensor representing the threshold for deciding when to remove

boxes based on score.

Returns:

nmsed_boxes: A [batch_size, max_detections, 4] float32 tensor

containing the non-max suppressed boxes.

nmsed_scores: A [batch_size, max_detections] float32 tensor

containing the scores for the boxes.

nmsed_classes: A [batch_size, max_detections] float32 tensor

containing the classes for the boxes.

valid_detections: A [batch_size] int32 tensor indicating the number of

valid detections per batch item. Only the top num_detections[i] entries in nms_boxes[i], nms_scores[i] and nms_class[i] are valid. The rest of the entries are zero paddings.

func Complex ¶ added in v1.1.0

func Complex(scope *Scope, real tf.Output, imag tf.Output, optional ...ComplexAttr) (out tf.Output)

Converts two real numbers to a complex number.

Given a tensor `real` representing the real part of a complex number, and a tensor `imag` representing the imaginary part of a complex number, this operation returns complex numbers elementwise of the form \$a + bj\$, where *a* represents the `real` part and *b* represents the `imag` part.

The input tensors `real` and `imag` must have the same shape.

For example:

``` # tensor 'real' is [2.25, 3.25] # tensor `imag` is [4.75, 5.75] tf.complex(real, imag) ==> [[2.25 + 4.75j], [3.25 + 5.75j]] ```

func ComplexAbs ¶ added in v1.1.0

func ComplexAbs(scope *Scope, x tf.Output, optional ...ComplexAbsAttr) (y tf.Output)

Computes the complex absolute value of a tensor.

Given a tensor `x` of complex numbers, this operation returns a tensor of type `float` or `double` that is the absolute value of each element in `x`. All elements in `x` must be complex numbers of the form \$a + bj\$. The absolute value is computed as \$ \sqrt{a^2 + b^2}\$.

For example:

>>> x = tf.complex(3.0, 4.0) >>> print((tf.raw_ops.ComplexAbs(x=x, Tout=tf.dtypes.float32, name=None)).numpy()) 5.0

func CompositeTensorVariantFromComponents ¶

func CompositeTensorVariantFromComponents(scope *Scope, components []tf.Output, metadata string) (encoded tf.Output)

Encodes an `ExtensionType` value into a `variant` scalar Tensor.

Returns a scalar variant tensor containing a single `CompositeTensorVariant` with the specified Tensor components and TypeSpec.

Arguments:

components: The component tensors for the extension type value.
metadata: String serialization for the TypeSpec.  (Note: the encoding for the TypeSpec

may change in future versions of TensorFlow.)

Returns A `variant` Tensor that containing the encoded value.

func CompositeTensorVariantToComponents ¶

func CompositeTensorVariantToComponents(scope *Scope, encoded tf.Output, metadata string, Tcomponents []tf.DataType) (components []tf.Output)

Decodes a `variant` scalar Tensor into an `ExtensionType` value.

Returns the Tensor components encoded in a `CompositeTensorVariant`.

Raises an error if `type_spec_proto` doesn't match the TypeSpec in `encoded`.

Arguments:

encoded: A scalar `variant` Tensor containing an encoded ExtensionType value.
metadata: String serialization for the TypeSpec.  Must be compatible with the

`TypeSpec` contained in `encoded`. (Note: the encoding for the TypeSpec may change in future versions of TensorFlow.)

Tcomponents: Expected dtypes for components.

Returns The component tensors for the ExtensionType value in `encoded`.

func CompressElement ¶

func CompressElement(scope *Scope, components []tf.Output) (compressed tf.Output)

Compresses a dataset element.

func ComputeAccidentalHits ¶ added in v1.1.0

func ComputeAccidentalHits(scope *Scope, true_classes tf.Output, sampled_candidates tf.Output, num_true int64, optional ...ComputeAccidentalHitsAttr) (indices tf.Output, ids tf.Output, weights tf.Output)

Computes the ids of the positions in sampled_candidates that match true_labels.

When doing log-odds NCE, the result of this op should be passed through a SparseToDense op, then added to the logits of the sampled candidates. This has the effect of 'removing' the sampled labels that match the true labels by making the classifier sure that they are sampled labels.

Arguments:

true_classes: The true_classes output of UnpackSparseLabels.
sampled_candidates: The sampled_candidates output of CandidateSampler.
num_true: Number of true labels per context.

Returns:

indices: A vector of indices corresponding to rows of true_candidates.
ids: A vector of IDs of positions in sampled_candidates that match a true_label

for the row with the corresponding index in indices.

weights: A vector of the same length as indices and ids, in which each element

is -FLOAT_MAX.

func ComputeBatchSize ¶

func ComputeBatchSize(scope *Scope, input_dataset tf.Output) (batch_size tf.Output)

Computes the static batch size of a dataset sans partial batches.

func ComputeDedupDataSize ¶

func ComputeDedupDataSize(scope *Scope, config string) (num_elements tf.Output)

An op computes the size of the deduplication data from embedding core and returns the updated config.

This op is to compute size of the deduplication data so to provide this information to the op that computes the tuple mask of deduplication data can have static output shape.

Arguments:

config: Serialized TPUEmbeddingConfiguration proto.

Returns The size of the deduplicated data from infeed.

func ComputeDedupDataTupleMask ¶

func ComputeDedupDataTupleMask(scope *Scope, config string) (output_shape tf.Output)

An op computes tuple mask of deduplication data from embedding core.

The deduplication data receiving from embedding core is a Tensor with type=DT_VARIANT. The tensor itself is an XLA nested tuple, whose elements are rank 1 tensors. This op is to represents types and length of these elements.

Arguments:

config: Serialized TPUEmbeddingConfiguration proto.

Returns A 2-D int tensor represent mask of deduplication data tuple generated by `XlaRecvTPUEmbeddingDeduplicationData`. The tuple has several integer and float type 1-D tensor tuple elements. The first dimenion of this output_shape 2-D tensor is tensor type of tuple elements, `0` represents integer tensor, `1` represents float tensor. The second dimension of `output_shape` gives length of each tuple element.

func Concat ¶ added in v1.1.0

func Concat(scope *Scope, concat_dim tf.Output, values []tf.Output) (output tf.Output)

Concatenates tensors along one dimension.

Arguments:

concat_dim: 0-D.  The dimension along which to concatenate.  Must be in the

range [0, rank(values)).

values: The `N` Tensors to concatenate. Their ranks and types must match,

and their sizes must match in all dimensions except `concat_dim`.

Returns A `Tensor` with the concatenation of values stacked along the `concat_dim` dimension. This tensor's shape matches that of `values` except in `concat_dim` where it has the sum of the sizes.

func ConcatOffset ¶ added in v1.1.0

func ConcatOffset(scope *Scope, concat_dim tf.Output, shape []tf.Output) (offset []tf.Output)

Computes offsets of concat inputs within its output.

For example:

>>> x = [2, 2, 7] >>> y = [2, 3, 7] >>> z = [2, 9, 7] >>> offsets = concat_offset(1, [x, y, z]) >>> [list(off.numpy()) for off in offsets] [[0, 0, 0], [0, 2, 0], [0, 5, 0]]

This is typically used by gradient computations for a concat operation.

Arguments:

concat_dim: The dimension along which to concatenate.
shape: The `N` int32 or int64 vectors representing shape of tensors being concatenated.

Returns The `N` vectors representing the starting offset of input tensors within the concatenated output with type matching `shape`.

func ConcatV2 ¶ added in v1.1.0

func ConcatV2(scope *Scope, values []tf.Output, axis tf.Output) (output tf.Output)

Concatenates tensors along one dimension.

Arguments:

values: List of `N` Tensors to concatenate. Their ranks and types must match,

and their sizes must match in all dimensions except `concat_dim`.

axis: 0-D.  The dimension along which to concatenate.  Must be in the

range [-rank(values), rank(values)).

Returns A `Tensor` with the concatenation of values stacked along the `concat_dim` dimension. This tensor's shape matches that of `values` except in `concat_dim` where it has the sum of the sizes.

func ConcatenateDataset ¶ added in v1.3.0

func ConcatenateDataset(scope *Scope, input_dataset tf.Output, another_dataset tf.Output, output_types []tf.DataType, output_shapes []tf.Shape, optional ...ConcatenateDatasetAttr) (handle tf.Output)

Creates a dataset that concatenates `input_dataset` with `another_dataset`.

func ConfigureAndInitializeGlobalTPU ¶

func ConfigureAndInitializeGlobalTPU(scope *Scope, optional ...ConfigureAndInitializeGlobalTPUAttr) (output tf.Output)

An op that sets up the centralized structures for a distributed TPU system.

Returns A vector containing the global TPU id of each TPU on the host.

func ConfigureDistributedTPU ¶ added in v1.12.1

func ConfigureDistributedTPU(scope *Scope, optional ...ConfigureDistributedTPUAttr) (topology tf.Output)

Sets up the centralized structures for a distributed TPU system.

Returns A serialized tensorflow.tpu.TopologyProto that describes the TPU topology.

func ConfigureTPUEmbedding ¶

func ConfigureTPUEmbedding(scope *Scope, config string) (o *tf.Operation)

Sets up TPUEmbedding in a distributed TPU system.

Arguments:

config: Serialized tensorflow.tpu.TPUEmbeddingConfiguration that

describes the embedding lookups of the program.

Returns the created operation.

func ConfigureTPUEmbeddingHost ¶

func ConfigureTPUEmbeddingHost(scope *Scope, common_config tf.Output, memory_config tf.Output, config string) (network_config tf.Output)

An op that configures the TPUEmbedding software on a host.

Arguments:

common_config: A string-encoded common configuration proto containing metadata

about the TPUEmbedding partitioner output.

memory_config: A string-encoded memory config proto containing metadata about

the memory allocations reserved for TPUEmbedding.

config: An TPUEmbeddingConfiguration proto serialized to a string,

describing the desired TPUEmbedding configuration.

Returns A string containing metadata about the hostname and RPC port used for communication with this host.

func ConfigureTPUEmbeddingMemory ¶

func ConfigureTPUEmbeddingMemory(scope *Scope, common_config tf.Output) (memory_config tf.Output)

An op that configures the TPUEmbedding software on a host.

Arguments:

common_config: A string-encoded CommonConfiguration proto containing metadata

about the TPUEmbedding partitioner output and the HBM size (in bytes) required for operation.

Returns A string-encoded memory configuration containing metadata about the memory allocations reserved for TPUEmbedding.

func Conj ¶ added in v1.1.0

func Conj(scope *Scope, input tf.Output) (output tf.Output)

Returns the complex conjugate of a complex number.

Given a tensor `input` of complex numbers, this operation returns a tensor of complex numbers that are the complex conjugate of each element in `input`. The complex numbers in `input` must be of the form \$a + bj\$, where *a* is the real part and *b* is the imaginary part.

The complex conjugate returned by this operation is of the form \$a - bj\$.

For example:

``` # tensor 'input' is [-2.25 + 4.75j, 3.25 + 5.75j] tf.conj(input) ==> [-2.25 - 4.75j, 3.25 - 5.75j] ```

func ConjugateTranspose ¶ added in v1.5.0

func ConjugateTranspose(scope *Scope, x tf.Output, perm tf.Output) (y tf.Output)

Shuffle dimensions of x according to a permutation and conjugate the result.

The output `y` has the same rank as `x`. The shapes of `x` and `y` satisfy:

`y.shape[i] == x.shape[perm[i]] for i in [0, 1, ..., rank(x) - 1]`
`y[i,j,k,...,s,t,u] == conj(x[perm[i], perm[j], perm[k],...,perm[s], perm[t], perm[u]])`

func ConnectTPUEmbeddingHosts ¶

func ConnectTPUEmbeddingHosts(scope *Scope, network_configs []tf.Output) (o *tf.Operation)

An op that sets up communication between TPUEmbedding host software instances

after ConfigureTPUEmbeddingHost has been called on each host.

Arguments:

network_configs: Strings containing metadata about the hostname and RPC port

used for communication with all hosts.

Returns the created operation.

func Const ¶

func Const(scope *Scope, value interface{}) (output tf.Output)

Const adds an operation to graph that produces value as output.

func ConsumeMutexLock ¶ added in v1.9.0

func ConsumeMutexLock(scope *Scope, mutex_lock tf.Output) (o *tf.Operation)

This op consumes a lock created by `MutexLock`.

This op exists to consume a tensor created by `MutexLock` (other than direct control dependencies). It should be the only that consumes the tensor, and will raise an error if it is not. Its only purpose is to keep the mutex lock tensor alive until it is consumed by this op.

**NOTE**: This operation must run on the same device as its input. This may be enforced via the `colocate_with` mechanism.

Arguments:

mutex_lock: A tensor returned by `MutexLock`.

Returns the created operation.

func ControlTrigger ¶ added in v1.1.0

func ControlTrigger(scope *Scope) (o *tf.Operation)

Does nothing. Serves as a control trigger for scheduling.

Only useful as a placeholder for control edges.

Returns the created operation.

func Conv ¶

func Conv(scope *Scope, input tf.Output, filter tf.Output, strides []int64, padding string, optional ...ConvAttr) (output tf.Output)

Computes a N-D convolution given (N+1+batch_dims)-D `input` and (N+2)-D `filter` tensors.

General function for computing a N-D convolution. It is required that `1 <= N <= 3`.

Arguments:

input: Tensor of type T and shape `batch_shape + spatial_shape + [in_channels]` in the

case that `channels_last_format = true` or shape `batch_shape + [in_channels] + spatial_shape` if `channels_last_format = false`. spatial_shape is N-dimensional with `N=2` or `N=3`. Also note that `batch_shape` is dictated by the parameter `batch_dims` and defaults to 1.

filter: An `(N+2)-D` Tensor with the same type as `input` and shape

`spatial_filter_shape + [in_channels, out_channels]`, where spatial_filter_shape is N-dimensional with `N=2` or `N=3`.

strides: 1-D tensor of length `N+2`. The stride of the sliding window for each

dimension of `input`. Must have `strides[0] = strides[N+1] = 1`.

padding: The type of padding algorithm to use.

Returns A (N+1+batch_dims)-D tensor. The dimension order is determined by the value of `channels_last_format`, see below for details.

func Conv2D ¶ added in v1.1.0

func Conv2D(scope *Scope, input tf.Output, filter tf.Output, strides []int64, padding string, optional ...Conv2DAttr) (output tf.Output)

Computes a 2-D convolution given 4-D `input` and `filter` tensors.

Given an input tensor of shape `[batch, in_height, in_width, in_channels]` and a filter / kernel tensor of shape `[filter_height, filter_width, in_channels, out_channels]`, this op performs the following:

Flattens the filter to a 2-D matrix with shape `[filter_height * filter_width * in_channels, output_channels]`.
Extracts image patches from the input tensor to form a *virtual* tensor of shape `[batch, out_height, out_width, filter_height * filter_width * in_channels]`.
For each patch, right-multiplies the filter matrix and the image patch vector.

In detail, with the default NHWC format,

output[b, i, j, k] =
    sum_{di, dj, q} input[b, strides[1] * i + di, strides[2] * j + dj, q] *
                    filter[di, dj, q, k]

Must have `strides[0] = strides[3] = 1`. For the most common case of the same horizontal and vertices strides, `strides = [1, stride, stride, 1]`.

Arguments:

input: A 4-D tensor. The dimension order is interpreted according to the value

of `data_format`, see below for details.

filter: A 4-D tensor of shape

`[filter_height, filter_width, in_channels, out_channels]`

strides: 1-D tensor of length 4.  The stride of the sliding window for each

dimension of `input`. The dimension order is determined by the value of `data_format`, see below for details.

padding: The type of padding algorithm to use.

Returns A 4-D tensor. The dimension order is determined by the value of `data_format`, see below for details.

func Conv2DBackpropFilter ¶ added in v1.1.0

func Conv2DBackpropFilter(scope *Scope, input tf.Output, filter_sizes tf.Output, out_backprop tf.Output, strides []int64, padding string, optional ...Conv2DBackpropFilterAttr) (output tf.Output)

Computes the gradients of convolution with respect to the filter.

Arguments:

input: 4-D with shape `[batch, in_height, in_width, in_channels]`.
filter_sizes: An integer vector representing the tensor shape of `filter`,

where `filter` is a 4-D `[filter_height, filter_width, in_channels, out_channels]` tensor.

out_backprop: 4-D with shape `[batch, out_height, out_width, out_channels]`.

Gradients w.r.t. the output of the convolution.

strides: The stride of the sliding window for each dimension of the input

of the convolution. Must be in the same order as the dimension specified with format.

padding: The type of padding algorithm to use.

Returns 4-D with shape `[filter_height, filter_width, in_channels, out_channels]`. Gradient w.r.t. the `filter` input of the convolution.

func Conv2DBackpropFilterV2 ¶

func Conv2DBackpropFilterV2(scope *Scope, input tf.Output, filter tf.Output, out_backprop tf.Output, strides []int64, padding string, optional ...Conv2DBackpropFilterV2Attr) (output tf.Output)

Computes the gradients of convolution with respect to the filter.

Arguments:

input: 4-D with shape `[batch, in_height, in_width, in_channels]`.
filter: 4-D with shape `[filter_height, filter_width, in_channels, out_channels]`.

Only shape of tensor is used.

out_backprop: 4-D with shape `[batch, out_height, out_width, out_channels]`.

Gradients w.r.t. the output of the convolution.

strides: The stride of the sliding window for each dimension of the input

of the convolution. Must be in the same order as the dimension specified with format.

padding: The type of padding algorithm to use.

Returns 4-D with shape `[filter_height, filter_width, in_channels, out_channels]`. Gradient w.r.t. the `filter` input of the convolution.

func Conv2DBackpropInput ¶ added in v1.1.0

func Conv2DBackpropInput(scope *Scope, input_sizes tf.Output, filter tf.Output, out_backprop tf.Output, strides []int64, padding string, optional ...Conv2DBackpropInputAttr) (output tf.Output)

Computes the gradients of convolution with respect to the input.

Arguments:

input_sizes: An integer vector representing the shape of `input`,

where `input` is a 4-D `[batch, height, width, channels]` tensor.

filter: 4-D with shape

`[filter_height, filter_width, in_channels, out_channels]`.

out_backprop: 4-D with shape `[batch, out_height, out_width, out_channels]`.

Gradients w.r.t. the output of the convolution.

strides: The stride of the sliding window for each dimension of the input

of the convolution. Must be in the same order as the dimension specified with format.

padding: The type of padding algorithm to use.

Returns 4-D with shape `[batch, in_height, in_width, in_channels]`. Gradient w.r.t. the input of the convolution.

func Conv2DBackpropInputV2 ¶

func Conv2DBackpropInputV2(scope *Scope, input tf.Output, filter tf.Output, out_backprop tf.Output, strides []int64, padding string, optional ...Conv2DBackpropInputV2Attr) (output tf.Output)

Computes the gradients of convolution with respect to the input.

Arguments:

input: 4-D with shape `[batch, in_height, in_width, in_channels]`.

Only shape of tensor is used.

filter: 4-D with shape

`[filter_height, filter_width, in_channels, out_channels]`.

out_backprop: 4-D with shape `[batch, out_height, out_width, out_channels]`.

Gradients w.r.t. the output of the convolution.

strides: The stride of the sliding window for each dimension of the input

of the convolution. Must be in the same order as the dimension specified with format.

padding: The type of padding algorithm to use.

Returns 4-D with shape `[batch, in_height, in_width, in_channels]`. Gradient w.r.t. the input of the convolution.

func Conv3D ¶ added in v1.1.0

func Conv3D(scope *Scope, input tf.Output, filter tf.Output, strides []int64, padding string, optional ...Conv3DAttr) (output tf.Output)

Computes a 3-D convolution given 5-D `input` and `filter` tensors.

In signal processing, cross-correlation is a measure of similarity of two waveforms as a function of a time-lag applied to one of them. This is also known as a sliding dot product or sliding inner-product.

Our Conv3D implements a form of cross-correlation.

Arguments:

input: Shape `[batch, in_depth, in_height, in_width, in_channels]`.
filter: Shape `[filter_depth, filter_height, filter_width, in_channels,

out_channels]`. `in_channels` must match between `input` and `filter`.

strides: 1-D tensor of length 5. The stride of the sliding window for each

dimension of `input`. Must have `strides[0] = strides[4] = 1`.

padding: The type of padding algorithm to use.

func Conv3DBackpropFilter ¶ added in v1.1.0

func Conv3DBackpropFilter(scope *Scope, input tf.Output, filter tf.Output, out_backprop tf.Output, strides []int64, padding string, optional ...Conv3DBackpropFilterAttr) (output tf.Output)

Computes the gradients of 3-D convolution with respect to the filter.

DEPRECATED at GraphDef version 10: Use Conv3DBackpropFilterV2

Arguments:

input: Shape `[batch, depth, rows, cols, in_channels]`.
filter: Shape `[depth, rows, cols, in_channels, out_channels]`.

`in_channels` must match between `input` and `filter`.

out_backprop: Backprop signal of shape `[batch, out_depth, out_rows, out_cols,

out_channels]`.

strides: 1-D tensor of length 5. The stride of the sliding window for each

dimension of `input`. Must have `strides[0] = strides[4] = 1`.

padding: The type of padding algorithm to use.

func Conv3DBackpropFilterV2 ¶ added in v1.1.0

func Conv3DBackpropFilterV2(scope *Scope, input tf.Output, filter_sizes tf.Output, out_backprop tf.Output, strides []int64, padding string, optional ...Conv3DBackpropFilterV2Attr) (output tf.Output)

Computes the gradients of 3-D convolution with respect to the filter.

Arguments:

input: Shape `[batch, depth, rows, cols, in_channels]`.
filter_sizes: An integer vector representing the tensor shape of `filter`,

where `filter` is a 5-D `[filter_depth, filter_height, filter_width, in_channels, out_channels]` tensor.

out_backprop: Backprop signal of shape `[batch, out_depth, out_rows, out_cols,

out_channels]`.

strides: 1-D tensor of length 5. The stride of the sliding window for each

dimension of `input`. Must have `strides[0] = strides[4] = 1`.

padding: The type of padding algorithm to use.

func Conv3DBackpropInput ¶ added in v1.1.0

func Conv3DBackpropInput(scope *Scope, input tf.Output, filter tf.Output, out_backprop tf.Output, strides []int64, padding string, optional ...Conv3DBackpropInputAttr) (output tf.Output)

Computes the gradients of 3-D convolution with respect to the input.

DEPRECATED at GraphDef version 10: Use Conv3DBackpropInputV2

Arguments:

input: Shape `[batch, depth, rows, cols, in_channels]`.
filter: Shape `[depth, rows, cols, in_channels, out_channels]`.

`in_channels` must match between `input` and `filter`.

out_backprop: Backprop signal of shape `[batch, out_depth, out_rows, out_cols,

out_channels]`.

strides: 1-D tensor of length 5. The stride of the sliding window for each

dimension of `input`. Must have `strides[0] = strides[4] = 1`.

padding: The type of padding algorithm to use.

func Conv3DBackpropInputV2 ¶ added in v1.1.0

func Conv3DBackpropInputV2(scope *Scope, input_sizes tf.Output, filter tf.Output, out_backprop tf.Output, strides []int64, padding string, optional ...Conv3DBackpropInputV2Attr) (output tf.Output)

Computes the gradients of 3-D convolution with respect to the input.

Arguments:

input_sizes: An integer vector representing the tensor shape of `input`,

where `input` is a 5-D `[batch, depth, rows, cols, in_channels]` tensor.

filter: Shape `[depth, rows, cols, in_channels, out_channels]`.

`in_channels` must match between `input` and `filter`.

out_backprop: Backprop signal of shape `[batch, out_depth, out_rows, out_cols,

out_channels]`.

strides: 1-D tensor of length 5. The stride of the sliding window for each

dimension of `input`. Must have `strides[0] = strides[4] = 1`.

padding: The type of padding algorithm to use.

func Copy ¶

func Copy(scope *Scope, input tf.Output, optional ...CopyAttr) (output tf.Output)

Copy a tensor from CPU-to-CPU or GPU-to-GPU.

Performs CPU-to-CPU or GPU-to-GPU deep-copying of tensor, depending on the device on which the tensor is allocated. N.B.: If the all downstream attached debug ops are disabled given the current gRPC gating status, the output will simply forward the input tensor without deep-copying. See the documentation of Debug* ops for more details.

Unlike the CopyHost Op, this op does not have HostMemory constraint on its input or output.

Arguments:

input: Input tensor.

func CopyHost ¶

func CopyHost(scope *Scope, input tf.Output, optional ...CopyHostAttr) (output tf.Output)

Copy a tensor to host.

Performs CPU-to-CPU deep-copying of tensor. N.B.: If the all downstream attached debug ops are disabled given the current gRPC gating status, the output will simply forward the input tensor without deep-copying. See the documentation of Debug* ops for more details.

Unlike the Copy Op, this op has HostMemory constraint on its input or output.

Arguments:

input: Input tensor.

func Cos ¶ added in v1.1.0

func Cos(scope *Scope, x tf.Output) (y tf.Output)

Computes cos of x element-wise.

Given an input tensor, this function computes cosine of every
element in the tensor. Input range is `(-inf, inf)` and
output range is `[-1,1]`. If input lies outside the boundary, `nan`
is returned.

```python
x = tf.constant([-float("inf"), -9, -0.5, 1, 1.2, 200, 10000, float("inf")])
tf.math.cos(x) ==> [nan -0.91113025 0.87758255 0.5403023 0.36235774 0.48718765 -0.95215535 nan]
```

func Cosh ¶ added in v1.3.0

func Cosh(scope *Scope, x tf.Output) (y tf.Output)

Computes hyperbolic cosine of x element-wise.

Given an input tensor, this function computes hyperbolic cosine of every
element in the tensor. Input range is `[-inf, inf]` and output range
is `[1, inf]`.

```python
x = tf.constant([-float("inf"), -9, -0.5, 1, 1.2, 2, 10, float("inf")])
tf.math.cosh(x) ==> [inf 4.0515420e+03 1.1276259e+00 1.5430807e+00 1.8106556e+00 3.7621956e+00 1.1013233e+04 inf]
```

func CropAndResize ¶ added in v1.1.0

func CropAndResize(scope *Scope, image tf.Output, boxes tf.Output, box_ind tf.Output, crop_size tf.Output, optional ...CropAndResizeAttr) (crops tf.Output)

Extracts crops from the input image tensor and resizes them.

Extracts crops from the input image tensor and resizes them using bilinear sampling or nearest neighbor sampling (possibly with aspect ratio change) to a common output size specified by `crop_size`. This is more general than the `crop_to_bounding_box` op which extracts a fixed size slice from the input image and does not allow resizing or aspect ratio change.

Returns a tensor with `crops` from the input `image` at positions defined at the bounding box locations in `boxes`. The cropped boxes are all resized (with bilinear or nearest neighbor interpolation) to a fixed `size = [crop_height, crop_width]`. The result is a 4-D tensor `[num_boxes, crop_height, crop_width, depth]`. The resizing is corner aligned. In particular, if `boxes = [[0, 0, 1, 1]]`, the method will give identical results to using `tf.image.resize_bilinear()` or `tf.image.resize_nearest_neighbor()`(depends on the `method` argument) with `align_corners=True`.

Arguments:

image: A 4-D tensor of shape `[batch, image_height, image_width, depth]`.

Both `image_height` and `image_width` need to be positive.

boxes: A 2-D tensor of shape `[num_boxes, 4]`. The `i`-th row of the tensor

specifies the coordinates of a box in the `box_ind[i]` image and is specified in normalized coordinates `[y1, x1, y2, x2]`. A normalized coordinate value of `y` is mapped to the image coordinate at `y * (image_height - 1)`, so as the `[0, 1]` interval of normalized image height is mapped to `[0, image_height - 1]` in image height coordinates. We do allow `y1` > `y2`, in which case the sampled crop is an up-down flipped version of the original image. The width dimension is treated similarly. Normalized coordinates outside the `[0, 1]` range are allowed, in which case we use `extrapolation_value` to extrapolate the input image values.

box_ind: A 1-D tensor of shape `[num_boxes]` with int32 values in `[0, batch)`.

The value of `box_ind[i]` specifies the image that the `i`-th box refers to.

crop_size: A 1-D tensor of 2 elements, `size = [crop_height, crop_width]`. All

cropped image patches are resized to this size. The aspect ratio of the image content is not preserved. Both `crop_height` and `crop_width` need to be positive.

Returns A 4-D tensor of shape `[num_boxes, crop_height, crop_width, depth]`.

func CropAndResizeGradBoxes ¶ added in v1.1.0

func CropAndResizeGradBoxes(scope *Scope, grads tf.Output, image tf.Output, boxes tf.Output, box_ind tf.Output, optional ...CropAndResizeGradBoxesAttr) (output tf.Output)

Computes the gradient of the crop_and_resize op wrt the input boxes tensor.

Arguments:

grads: A 4-D tensor of shape `[num_boxes, crop_height, crop_width, depth]`.
image: A 4-D tensor of shape `[batch, image_height, image_width, depth]`.

Both `image_height` and `image_width` need to be positive.

boxes: A 2-D tensor of shape `[num_boxes, 4]`. The `i`-th row of the tensor

specifies the coordinates of a box in the `box_ind[i]` image and is specified in normalized coordinates `[y1, x1, y2, x2]`. A normalized coordinate value of `y` is mapped to the image coordinate at `y * (image_height - 1)`, so as the `[0, 1]` interval of normalized image height is mapped to `[0, image_height - 1] in image height coordinates. We do allow y1 > y2, in which case the sampled crop is an up-down flipped version of the original image. The width dimension is treated similarly. Normalized coordinates outside the `[0, 1]` range are allowed, in which case we use `extrapolation_value` to extrapolate the input image values.

box_ind: A 1-D tensor of shape `[num_boxes]` with int32 values in `[0, batch)`.

The value of `box_ind[i]` specifies the image that the `i`-th box refers to.

Returns A 2-D tensor of shape `[num_boxes, 4]`.

func CropAndResizeGradImage ¶ added in v1.1.0

func CropAndResizeGradImage(scope *Scope, grads tf.Output, boxes tf.Output, box_ind tf.Output, image_size tf.Output, T tf.DataType, optional ...CropAndResizeGradImageAttr) (output tf.Output)

Computes the gradient of the crop_and_resize op wrt the input image tensor.

Arguments:

grads: A 4-D tensor of shape `[num_boxes, crop_height, crop_width, depth]`.
boxes: A 2-D tensor of shape `[num_boxes, 4]`. The `i`-th row of the tensor

specifies the coordinates of a box in the `box_ind[i]` image and is specified in normalized coordinates `[y1, x1, y2, x2]`. A normalized coordinate value of `y` is mapped to the image coordinate at `y * (image_height - 1)`, so as the `[0, 1]` interval of normalized image height is mapped to `[0, image_height - 1] in image height coordinates. We do allow y1 > y2, in which case the sampled crop is an up-down flipped version of the original image. The width dimension is treated similarly. Normalized coordinates outside the `[0, 1]` range are allowed, in which case we use `extrapolation_value` to extrapolate the input image values.

box_ind: A 1-D tensor of shape `[num_boxes]` with int32 values in `[0, batch)`.

The value of `box_ind[i]` specifies the image that the `i`-th box refers to.

image_size: A 1-D tensor with value `[batch, image_height, image_width, depth]`

containing the original image size. Both `image_height` and `image_width` need to be positive.

Returns A 4-D tensor of shape `[batch, image_height, image_width, depth]`.

func Cross ¶ added in v1.1.0

func Cross(scope *Scope, a tf.Output, b tf.Output) (product tf.Output)

Compute the pairwise cross product.

`a` and `b` must be the same shape; they can either be simple 3-element vectors, or any shape where the innermost dimension is 3. In the latter case, each pair of corresponding 3-element vectors is cross-multiplied independently.

Arguments:

a: A tensor containing 3-element vectors.
b: Another tensor, of same type and shape as `a`.

Returns Pairwise cross product of the vectors in `a` and `b`.

func CrossReplicaSum ¶ added in v1.12.1

func CrossReplicaSum(scope *Scope, input tf.Output, group_assignment tf.Output) (output tf.Output)

An Op to sum inputs across replicated TPU instances.

Each instance supplies its own input.

For example, suppose there are 8 TPU instances: `[A, B, C, D, E, F, G, H]`. Passing group_assignment=`[[0,2,4,6],[1,3,5,7]]` sets `A, C, E, G` as group 0, and `B, D, F, H` as group 1. Thus we get the outputs: `[A+C+E+G, B+D+F+H, A+C+E+G, B+D+F+H, A+C+E+G, B+D+F+H, A+C+E+G, B+D+F+H]`.

Arguments:

input: The local input to the sum.
group_assignment: An int32 tensor with shape

[num_groups, num_replicas_per_group]. `group_assignment[i]` represents the replica ids in the ith subgroup.

Returns The sum of all the distributed inputs.

func CudnnRNN ¶ added in v1.9.0

func CudnnRNN(scope *Scope, input tf.Output, input_h tf.Output, input_c tf.Output, params tf.Output, optional ...CudnnRNNAttr) (output tf.Output, output_h tf.Output, output_c tf.Output, reserve_space tf.Output)

A RNN backed by cuDNN.

Computes the RNN from the input and initial states, with respect to the params buffer.

rnn_mode: Indicates the type of the RNN model. input_mode: Indicate whether there is a linear projection between the input and

the actual computation before the first layer. 'skip_input' is only allowed
when input_size == num_units; 'auto_select' implies 'skip_input' when
input_size == num_units; otherwise, it implies 'linear_input'.

direction: Indicates whether a bidirectional model will be used. Should be

"unidirectional" or "bidirectional".

dropout: Dropout probability. When set to 0., dropout is disabled. seed: The 1st part of a seed to initialize dropout. seed2: The 2nd part of a seed to initialize dropout. input: A 3-D tensor with the shape of [seq_length, batch_size, input_size]. input_h: A 3-D tensor with the shape of [num_layer * dir, batch_size,

num_units].

input_c: For LSTM, a 3-D tensor with the shape of

[num_layer * dir, batch, num_units]. For other models, it is ignored.

params: A 1-D tensor that contains the weights and biases in an opaque layout.

The size must be created through CudnnRNNParamsSize, and initialized
separately. Note that they might not be compatible across different
generations. So it is a good idea to save and restore

output: A 3-D tensor with the shape of [seq_length, batch_size,

dir * num_units].

output_h: The same shape has input_h. output_c: The same shape as input_c for LSTM. An empty tensor for other models. is_training: Indicates whether this operation is used for inference or

training.

reserve_space: An opaque tensor that can be used in backprop calculation. It

is only produced if is_training is false.

func CudnnRNNBackprop ¶ added in v1.9.0

func CudnnRNNBackprop(scope *Scope, input tf.Output, input_h tf.Output, input_c tf.Output, params tf.Output, output tf.Output, output_h tf.Output, output_c tf.Output, output_backprop tf.Output, output_h_backprop tf.Output, output_c_backprop tf.Output, reserve_space tf.Output, optional ...CudnnRNNBackpropAttr) (input_backprop tf.Output, input_h_backprop tf.Output, input_c_backprop tf.Output, params_backprop tf.Output)

Backprop step of CudnnRNN.

Compute the backprop of both data and weights in a RNN.

rnn_mode: Indicates the type of the RNN model. input_mode: Indicate whether there is a linear projection between the input and

the actual computation before the first layer. 'skip_input' is only allowed
when input_size == num_units; 'auto_select' implies 'skip_input' when
input_size == num_units; otherwise, it implies 'linear_input'.

direction: Indicates whether a bidirectional model will be used. Should be

"unidirectional" or "bidirectional".

dropout: Dropout probability. When set to 0., dropout is disabled. seed: The 1st part of a seed to initialize dropout. seed2: The 2nd part of a seed to initialize dropout. input: A 3-D tensor with the shape of [seq_length, batch_size, input_size]. input_h: A 3-D tensor with the shape of [num_layer * dir, batch_size,

num_units].

input_c: For LSTM, a 3-D tensor with the shape of

[num_layer * dir, batch, num_units]. For other models, it is ignored.

params: A 1-D tensor that contains the weights and biases in an opaque layout.

The size must be created through CudnnRNNParamsSize, and initialized
separately. Note that they might not be compatible across different
generations. So it is a good idea to save and restore

output: A 3-D tensor with the shape of [seq_length, batch_size,

dir * num_units].

output_h: The same shape has input_h. output_c: The same shape as input_c for LSTM. An empty tensor for other models. output_backprop: A 3-D tensor with the same shape as output in the forward pass. output_h_backprop: A 3-D tensor with the same shape as output_h in the forward

pass.

output_c_backprop: A 3-D tensor with the same shape as output_c in the forward

pass.

reserve_space: The same reserve_space produced in for forward operation. input_backprop: The backprop to input in the forward pass. Has the same shape

as input.

input_h_backprop: The backprop to input_h in the forward pass. Has the same

shape as input_h.

input_c_backprop: The backprop to input_c in the forward pass. Has the same

shape as input_c.

params_backprop: The backprop to the params buffer in the forward pass. Has the

same shape as params.

func CudnnRNNBackpropV2 ¶ added in v1.11.0

func CudnnRNNBackpropV2(scope *Scope, input tf.Output, input_h tf.Output, input_c tf.Output, params tf.Output, output tf.Output, output_h tf.Output, output_c tf.Output, output_backprop tf.Output, output_h_backprop tf.Output, output_c_backprop tf.Output, reserve_space tf.Output, host_reserved tf.Output, optional ...CudnnRNNBackpropV2Attr) (input_backprop tf.Output, input_h_backprop tf.Output, input_c_backprop tf.Output, params_backprop tf.Output)

Backprop step of CudnnRNN.

Compute the backprop of both data and weights in a RNN. Takes an extra

"host_reserved" inupt than CudnnRNNBackprop, which is used to determine RNN
cudnnRNNAlgo_t and cudnnMathType_t.

rnn_mode: Indicates the type of the RNN model. input_mode: Indicates whether there is a linear projection between the input and

the actual computation before the first layer. 'skip_input' is only allowed
when input_size == num_units; 'auto_select' implies 'skip_input' when
input_size == num_units; otherwise, it implies 'linear_input'.

direction: Indicates whether a bidirectional model will be used. Should be

"unidirectional" or "bidirectional".

dropout: Dropout probability. When set to 0., dropout is disabled. seed: The 1st part of a seed to initialize dropout. seed2: The 2nd part of a seed to initialize dropout. input: A 3-D tensor with the shape of [seq_length, batch_size, input_size]. input_h: A 3-D tensor with the shape of [num_layer * dir, batch_size,

num_units].

input_c: For LSTM, a 3-D tensor with the shape of

[num_layer * dir, batch, num_units]. For other models, it is ignored.

params: A 1-D tensor that contains the weights and biases in an opaque layout.

The size must be created through CudnnRNNParamsSize, and initialized
separately. Note that they might not be compatible across different
generations. So it is a good idea to save and restore

output: A 3-D tensor with the shape of [seq_length, batch_size,

dir * num_units].

output_h: The same shape has input_h. output_c: The same shape as input_c for LSTM. An empty tensor for other models. output_backprop: A 3-D tensor with the same shape as output in the forward pass. output_h_backprop: A 3-D tensor with the same shape as output_h in the forward

pass.

output_c_backprop: A 3-D tensor with the same shape as output_c in the forward

pass.

reserve_space: The same reserve_space produced in the forward operation. host_reserved: The same host_reserved produced in the forward operation. input_backprop: The backprop to input in the forward pass. Has the same shape

as input.

input_h_backprop: The backprop to input_h in the forward pass. Has the same

shape as input_h.

input_c_backprop: The backprop to input_c in the forward pass. Has the same

shape as input_c.

params_backprop: The backprop to the params buffer in the forward pass. Has the

same shape as params.

func CudnnRNNBackpropV3 ¶ added in v1.12.1

func CudnnRNNBackpropV3(scope *Scope, input tf.Output, input_h tf.Output, input_c tf.Output, params tf.Output, sequence_lengths tf.Output, output tf.Output, output_h tf.Output, output_c tf.Output, output_backprop tf.Output, output_h_backprop tf.Output, output_c_backprop tf.Output, reserve_space tf.Output, host_reserved tf.Output, optional ...CudnnRNNBackpropV3Attr) (input_backprop tf.Output, input_h_backprop tf.Output, input_c_backprop tf.Output, params_backprop tf.Output)

Backprop step of CudnnRNNV3.

Compute the backprop of both data and weights in a RNN. Takes an extra

"sequence_lengths" input than CudnnRNNBackprop.

rnn_mode: Indicates the type of the RNN model. input_mode: Indicates whether there is a linear projection between the input and

the actual computation before the first layer. 'skip_input' is only allowed
when input_size == num_units; 'auto_select' implies 'skip_input' when
input_size == num_units; otherwise, it implies 'linear_input'.

direction: Indicates whether a bidirectional model will be used. Should be

"unidirectional" or "bidirectional".

dropout: Dropout probability. When set to 0., dropout is disabled. seed: The 1st part of a seed to initialize dropout. seed2: The 2nd part of a seed to initialize dropout. input: If time_major is true, this is a 3-D tensor with the shape of

[seq_length, batch_size, input_size]. If time_major is false, the shape is
[batch_size, seq_length, input_size].

input_h: If time_major is true, this is a 3-D tensor with the shape of

[num_layer * dir, batch_size, num_units]. If time_major is false, the shape
is [batch_size, num_layer * dir, num_units].

input_c: For LSTM, a 3-D tensor with the shape of

[num_layer * dir, batch, num_units]. For other models, it is ignored.

params: A 1-D tensor that contains the weights and biases in an opaque layout.

The size must be created through CudnnRNNParamsSize, and initialized
separately. Note that they might not be compatible across different
generations. So it is a good idea to save and restore

sequence_lengths: a vector of lengths of each input sequence. output: If time_major is true, this is a 3-D tensor with the shape of

[seq_length, batch_size, dir * num_units]. If time_major is false, the
shape is [batch_size, seq_length, dir * num_units].

output_h: The same shape has input_h. output_c: The same shape as input_c for LSTM. An empty tensor for other models. output_backprop: A 3-D tensor with the same shape as output in the forward pass. output_h_backprop: A 3-D tensor with the same shape as output_h in the forward

pass.

output_c_backprop: A 3-D tensor with the same shape as output_c in the forward

pass.

time_major: Indicates whether the input/output format is time major or batch

major.

reserve_space: The same reserve_space produced in the forward operation. input_backprop: The backprop to input in the forward pass. Has the same shape

as input.

input_h_backprop: The backprop to input_h in the forward pass. Has the same

shape as input_h.

input_c_backprop: The backprop to input_c in the forward pass. Has the same

shape as input_c.

params_backprop: The backprop to the params buffer in the forward pass. Has the

same shape as params.

func CudnnRNNCanonicalToParams ¶ added in v1.9.0

func CudnnRNNCanonicalToParams(scope *Scope, num_layers tf.Output, num_units tf.Output, input_size tf.Output, weights []tf.Output, biases []tf.Output, optional ...CudnnRNNCanonicalToParamsAttr) (params tf.Output)

Converts CudnnRNN params from canonical form to usable form.

Writes a set of weights into the opaque params buffer so they can be used in upcoming training or inferences.

Note that the params buffer may not be compatible across different GPUs. So any save and restoration should be converted to and from the canonical weights and biases.

num_layers: Specifies the number of layers in the RNN model. num_units: Specifies the size of the hidden state. input_size: Specifies the size of the input state. weights: the canonical form of weights that can be used for saving

and restoration. They are more likely to be compatible across different
generations.

biases: the canonical form of biases that can be used for saving

and restoration. They are more likely to be compatible across different
generations.

num_params: number of parameter sets for all layers.

Each layer may contain multiple parameter sets, with each set consisting of
a weight matrix and a bias vector.

rnn_mode: Indicates the type of the RNN model. input_mode: Indicate whether there is a linear projection between the input and

The actual computation before the first layer. 'skip_input' is only allowed
when input_size == num_units; 'auto_select' implies 'skip_input' when
input_size == num_units; otherwise, it implies 'linear_input'.

direction: Indicates whether a bidirectional model will be used.

dir = (direction == bidirectional) ? 2 : 1

dropout: dropout probability. When set to 0., dropout is disabled. seed: the 1st part of a seed to initialize dropout. seed2: the 2nd part of a seed to initialize dropout.

func CudnnRNNCanonicalToParamsV2 ¶ added in v1.15.0

func CudnnRNNCanonicalToParamsV2(scope *Scope, num_layers tf.Output, num_units tf.Output, input_size tf.Output, weights []tf.Output, biases []tf.Output, optional ...CudnnRNNCanonicalToParamsV2Attr) (params tf.Output)

Converts CudnnRNN params from canonical form to usable form. It supports the projection in LSTM.

Writes a set of weights into the opaque params buffer so they can be used in upcoming training or inferences.

Note that the params buffer may not be compatible across different GPUs. So any save and restoration should be converted to and from the canonical weights and biases.

num_layers: Specifies the number of layers in the RNN model. num_units: Specifies the size of the hidden state. input_size: Specifies the size of the input state. weights: the canonical form of weights that can be used for saving

and restoration. They are more likely to be compatible across different
generations.

biases: the canonical form of biases that can be used for saving

and restoration. They are more likely to be compatible across different
generations.

num_params_weights: number of weight parameter matrix for all layers. num_params_biases: number of bias parameter vector for all layers. rnn_mode: Indicates the type of the RNN model. input_mode: Indicate whether there is a linear projection between the input and

The actual computation before the first layer. 'skip_input' is only allowed
when input_size == num_units; 'auto_select' implies 'skip_input' when
input_size == num_units; otherwise, it implies 'linear_input'.

direction: Indicates whether a bidirectional model will be used.

dir = (direction == bidirectional) ? 2 : 1

dropout: dropout probability. When set to 0., dropout is disabled. seed: the 1st part of a seed to initialize dropout. seed2: the 2nd part of a seed to initialize dropout. num_proj: The output dimensionality for the projection matrices. If None or 0,

no projection is performed.

func CudnnRNNParamsSize ¶ added in v1.9.0

func CudnnRNNParamsSize(scope *Scope, num_layers tf.Output, num_units tf.Output, input_size tf.Output, T tf.DataType, S tf.DataType, optional ...CudnnRNNParamsSizeAttr) (params_size tf.Output)

Computes size of weights that can be used by a Cudnn RNN model.

Return the params size that can be used by the Cudnn RNN model. Subsequent weight allocation and initialization should use this size.

num_layers: Specifies the number of layers in the RNN model. num_units: Specifies the size of the hidden state. input_size: Specifies the size of the input state. rnn_mode: Indicates the type of the RNN model. input_mode: Indicate whether there is a linear projection between the input and

The actual computation before the first layer. 'skip_input' is only allowed
when input_size == num_units; 'auto_select' implies 'skip_input' when
input_size == num_units; otherwise, it implies 'linear_input'.

direction: Indicates whether a bidirectional model will be used.

dir = (direction == bidirectional) ? 2 : 1

dropout: dropout probability. When set to 0., dropout is disabled. seed: the 1st part of a seed to initialize dropout. seed2: the 2nd part of a seed to initialize dropout. params_size: The size of the params buffer that should be allocated and

initialized for this RNN model. Note that this params buffer may not be
compatible across GPUs. Please use CudnnRNNParamsWeights and
CudnnRNNParamsBiases to save and restore them in a way that is compatible
across different runs.

func CudnnRNNParamsToCanonical ¶ added in v1.9.0

func CudnnRNNParamsToCanonical(scope *Scope, num_layers tf.Output, num_units tf.Output, input_size tf.Output, params tf.Output, num_params int64, optional ...CudnnRNNParamsToCanonicalAttr) (weights []tf.Output, biases []tf.Output)

Retrieves CudnnRNN params in canonical form.

Retrieves a set of weights from the opaque params buffer that can be saved and restored in a way compatible with future runs.

Note that the params buffer may not be compatible across different GPUs. So any save and restoration should be converted to and from the canonical weights and biases.

num_layers: Specifies the number of layers in the RNN model. num_units: Specifies the size of the hidden state. input_size: Specifies the size of the input state. num_params: number of parameter sets for all layers.

Each layer may contain multiple parameter sets, with each set consisting of
a weight matrix and a bias vector.

weights: the canonical form of weights that can be used for saving

and restoration. They are more likely to be compatible across different
generations.

biases: the canonical form of biases that can be used for saving

and restoration. They are more likely to be compatible across different
generations.

rnn_mode: Indicates the type of the RNN model. input_mode: Indicate whether there is a linear projection between the input and

The actual computation before the first layer. 'skip_input' is only allowed
when input_size == num_units; 'auto_select' implies 'skip_input' when
input_size == num_units; otherwise, it implies 'linear_input'.

direction: Indicates whether a bidirectional model will be used.

dir = (direction == bidirectional) ? 2 : 1

dropout: dropout probability. When set to 0., dropout is disabled. seed: the 1st part of a seed to initialize dropout. seed2: the 2nd part of a seed to initialize dropout.

func CudnnRNNParamsToCanonicalV2 ¶ added in v1.15.0

func CudnnRNNParamsToCanonicalV2(scope *Scope, num_layers tf.Output, num_units tf.Output, input_size tf.Output, params tf.Output, num_params_weights int64, num_params_biases int64, optional ...CudnnRNNParamsToCanonicalV2Attr) (weights []tf.Output, biases []tf.Output)

Retrieves CudnnRNN params in canonical form. It supports the projection in LSTM.

Retrieves a set of weights from the opaque params buffer that can be saved and restored in a way compatible with future runs.

Note that the params buffer may not be compatible across different GPUs. So any save and restoration should be converted to and from the canonical weights and biases.

num_layers: Specifies the number of layers in the RNN model. num_units: Specifies the size of the hidden state. input_size: Specifies the size of the input state. num_params_weights: number of weight parameter matrix for all layers. num_params_biases: number of bias parameter vector for all layers. weights: the canonical form of weights that can be used for saving

and restoration. They are more likely to be compatible across different
generations.

biases: the canonical form of biases that can be used for saving

and restoration. They are more likely to be compatible across different
generations.

rnn_mode: Indicates the type of the RNN model. input_mode: Indicate whether there is a linear projection between the input and

The actual computation before the first layer. 'skip_input' is only allowed
when input_size == num_units; 'auto_select' implies 'skip_input' when
input_size == num_units; otherwise, it implies 'linear_input'.

direction: Indicates whether a bidirectional model will be used.

dir = (direction == bidirectional) ? 2 : 1

dropout: dropout probability. When set to 0., dropout is disabled. seed: the 1st part of a seed to initialize dropout. seed2: the 2nd part of a seed to initialize dropout. num_proj: The output dimensionality for the projection matrices. If None or 0,

no projection is performed.

func CudnnRNNV2 ¶ added in v1.11.0

func CudnnRNNV2(scope *Scope, input tf.Output, input_h tf.Output, input_c tf.Output, params tf.Output, optional ...CudnnRNNV2Attr) (output tf.Output, output_h tf.Output, output_c tf.Output, reserve_space tf.Output, host_reserved tf.Output)

A RNN backed by cuDNN.

Computes the RNN from the input and initial states, with respect to the params buffer. Produces one extra output "host_reserved" than CudnnRNN.

rnn_mode: Indicates the type of the RNN model. input_mode: Indicates whether there is a linear projection between the input and

the actual computation before the first layer. 'skip_input' is only allowed
when input_size == num_units; 'auto_select' implies 'skip_input' when
input_size == num_units; otherwise, it implies 'linear_input'.

direction: Indicates whether a bidirectional model will be used. Should be

"unidirectional" or "bidirectional".

dropout: Dropout probability. When set to 0., dropout is disabled. seed: The 1st part of a seed to initialize dropout. seed2: The 2nd part of a seed to initialize dropout. input: A 3-D tensor with the shape of [seq_length, batch_size, input_size]. input_h: A 3-D tensor with the shape of [num_layer * dir, batch_size,

num_units].

input_c: For LSTM, a 3-D tensor with the shape of

[num_layer * dir, batch, num_units]. For other models, it is ignored.

params: A 1-D tensor that contains the weights and biases in an opaque layout.

The size must be created through CudnnRNNParamsSize, and initialized
separately. Note that they might not be compatible across different
generations. So it is a good idea to save and restore

output: A 3-D tensor with the shape of [seq_length, batch_size,

dir * num_units].

output_h: The same shape has input_h. output_c: The same shape as input_c for LSTM. An empty tensor for other models. is_training: Indicates whether this operation is used for inference or

training.

reserve_space: An opaque tensor that can be used in backprop calculation. It

is only produced if is_training is true.

host_reserved: An opaque tensor that can be used in backprop calculation. It is

only produced if is_training is true. It is output on host memory rather than
device memory.

func CudnnRNNV3 ¶ added in v1.12.1

func CudnnRNNV3(scope *Scope, input tf.Output, input_h tf.Output, input_c tf.Output, params tf.Output, sequence_lengths tf.Output, optional ...CudnnRNNV3Attr) (output tf.Output, output_h tf.Output, output_c tf.Output, reserve_space tf.Output, host_reserved tf.Output)

A RNN backed by cuDNN.

Computes the RNN from the input and initial states, with respect to the params buffer. Accepts one extra input "sequence_lengths" than CudnnRNN.

rnn_mode: Indicates the type of the RNN model. input_mode: Indicates whether there is a linear projection between the input and

the actual computation before the first layer. 'skip_input' is only allowed
when input_size == num_units; 'auto_select' implies 'skip_input' when
input_size == num_units; otherwise, it implies 'linear_input'.

direction: Indicates whether a bidirectional model will be used. Should be

"unidirectional" or "bidirectional".

dropout: Dropout probability. When set to 0., dropout is disabled. seed: The 1st part of a seed to initialize dropout. seed2: The 2nd part of a seed to initialize dropout. input: If time_major is true, this is a 3-D tensor with the shape of

[seq_length, batch_size, input_size]. If time_major is false, the shape is
[batch_size, seq_length, input_size].

input_h: If time_major is true, this is a 3-D tensor with the shape of

[num_layer * dir, batch_size, num_units]. If time_major is false, the shape
is [batch_size, num_layer * dir, num_units].

input_c: For LSTM, a 3-D tensor with the shape of

[num_layer * dir, batch, num_units]. For other models, it is ignored.

params: A 1-D tensor that contains the weights and biases in an opaque layout.

The size must be created through CudnnRNNParamsSize, and initialized
separately. Note that they might not be compatible across different
generations. So it is a good idea to save and restore

sequence_lengths: a vector of lengths of each input sequence. output: If time_major is true, this is a 3-D tensor with the shape of

[seq_length, batch_size, dir * num_units]. If time_major is false, the
shape is [batch_size, seq_length, dir * num_units].

output_h: The same shape has input_h. output_c: The same shape as input_c for LSTM. An empty tensor for other models. is_training: Indicates whether this operation is used for inference or

training.

time_major: Indicates whether the input/output format is time major or batch

major.

reserve_space: An opaque tensor that can be used in backprop calculation. It

is only produced if is_training is true.

func Cumprod ¶ added in v1.1.0

func Cumprod(scope *Scope, x tf.Output, axis tf.Output, optional ...CumprodAttr) (out tf.Output)

Compute the cumulative product of the tensor `x` along `axis`.

By default, this op performs an inclusive cumprod, which means that the first element of the input is identical to the first element of the output:

```python tf.cumprod([a, b, c]) # => [a, a * b, a * b * c] ```

By setting the `exclusive` kwarg to `True`, an exclusive cumprod is performed instead:

```python tf.cumprod([a, b, c], exclusive=True) # => [1, a, a * b] ```

By setting the `reverse` kwarg to `True`, the cumprod is performed in the opposite direction:

```python tf.cumprod([a, b, c], reverse=True) # => [a * b * c, b * c, c] ```

This is more efficient than using separate `tf.reverse` ops.

The `reverse` and `exclusive` kwargs can also be combined:

```python tf.cumprod([a, b, c], exclusive=True, reverse=True) # => [b * c, c, 1] ```

Arguments:

x: A `Tensor`. Must be one of the following types: `float32`, `float64`,

`int64`, `int32`, `uint8`, `uint16`, `int16`, `int8`, `complex64`, `complex128`, `qint8`, `quint8`, `qint32`, `half`.

axis: A `Tensor` of type `int32` (default: 0). Must be in the range

`[-rank(x), rank(x))`.

func Cumsum ¶ added in v1.1.0

func Cumsum(scope *Scope, x tf.Output, axis tf.Output, optional ...CumsumAttr) (out tf.Output)

Compute the cumulative sum of the tensor `x` along `axis`.

By default, this op performs an inclusive cumsum, which means that the first element of the input is identical to the first element of the output:

```python tf.cumsum([a, b, c]) # => [a, a + b, a + b + c] ```

By setting the `exclusive` kwarg to `True`, an exclusive cumsum is performed instead:

```python tf.cumsum([a, b, c], exclusive=True) # => [0, a, a + b] ```

By setting the `reverse` kwarg to `True`, the cumsum is performed in the opposite direction:

```python tf.cumsum([a, b, c], reverse=True) # => [a + b + c, b + c, c] ```

This is more efficient than using separate `tf.reverse` ops.

The `reverse` and `exclusive` kwargs can also be combined:

```python tf.cumsum([a, b, c], exclusive=True, reverse=True) # => [b + c, c, 0] ```

Arguments:

x: A `Tensor`. Must be one of the following types: `float32`, `float64`,

`int64`, `int32`, `uint8`, `uint16`, `int16`, `int8`, `complex64`, `complex128`, `qint8`, `quint8`, `qint32`, `half`.

axis: A `Tensor` of type `int32` (default: 0). Must be in the range

`[-rank(x), rank(x))`.

func CumulativeLogsumexp ¶

func CumulativeLogsumexp(scope *Scope, x tf.Output, axis tf.Output, optional ...CumulativeLogsumexpAttr) (out tf.Output)

Compute the cumulative product of the tensor `x` along `axis`.

By default, this op performs an inclusive cumulative log-sum-exp, which means that the first element of the input is identical to the first element of the output: ```python tf.math.cumulative_logsumexp([a, b, c]) # => [a, log(exp(a) + exp(b)), log(exp(a) + exp(b) + exp(c))] ```

By setting the `exclusive` kwarg to `True`, an exclusive cumulative log-sum-exp is performed instead: ```python tf.cumulative_logsumexp([a, b, c], exclusive=True) # => [-inf, a, log(exp(a) * exp(b))] ``` Note that the neutral element of the log-sum-exp operation is `-inf`, however, for performance reasons, the minimal value representable by the floating point type is used instead.

By setting the `reverse` kwarg to `True`, the cumulative log-sum-exp is performed in the opposite direction.

Arguments:

x: A `Tensor`. Must be one of the following types: `float16`, `float32`, `float64`.
axis: A `Tensor` of type `int32` (default: 0). Must be in the range

`[-rank(x), rank(x))`.

func DTensorSetGlobalTPUArray ¶

func DTensorSetGlobalTPUArray(scope *Scope, topology tf.Output) (o *tf.Operation)

An op that informs a host of the global ids of all the of TPUs in the system.

Arguments:

topology: A serialized tensorflow.tpu.TopologyProto that describes the TPU topology.

Returns the created operation.

func DataFormatDimMap ¶ added in v1.6.0

func DataFormatDimMap(scope *Scope, x tf.Output, optional ...DataFormatDimMapAttr) (y tf.Output)

Returns the dimension index in the destination data format given the one in

the source data format.

Arguments:

x: A Tensor with each element as a dimension index in source data format.

Must be in the range [-4, 4).

Returns A Tensor with each element as a dimension index in destination data format.

func DataFormatVecPermute ¶ added in v1.6.0

func DataFormatVecPermute(scope *Scope, x tf.Output, optional ...DataFormatVecPermuteAttr) (y tf.Output)

Permute input tensor from `src_format` to `dst_format`.

Given source and destination format strings of length n=4 or 5, the input tensor must be a vector of size n or n-2, or a 2D tensor of shape (n, 2) or (n-2, 2).

If the first dimension of the input tensor is n-2, it is assumed that non-spatial dimensions are omitted (i.e `N`, `C`).

For example, with `src_format` of `NHWC`, `dst_format` of `NCHW`, and input: ``` [1, 2, 3, 4] ``` , the output will be: ``` [1, 4, 2, 3] ``` With `src_format` of `NDHWC`, `dst_format` of `NCDHW`, and input: ``` [[1, 6], [2, 7], [3, 8], [4, 9], [5, 10]] ``` , the output will be: ``` [[1, 6], [5, 10], [2, 7], [3, 8], [4, 9]] ``` With `src_format` of `NHWC`, `dst_format` of `NCHW`, and input: ``` [1, 2] ``` , the output will be: ``` [1, 2] ```

Arguments:

x: Tensor of rank 1 or 2 in source data format.

Returns Tensor of rank 1 or 2 in destination data format.

func DataServiceDataset ¶

func DataServiceDataset(scope *Scope, dataset_id tf.Output, processing_mode tf.Output, address tf.Output, protocol tf.Output, job_name tf.Output, max_outstanding_requests tf.Output, iteration_counter tf.Output, output_types []tf.DataType, output_shapes []tf.Shape, optional ...DataServiceDatasetAttr) (handle tf.Output)

Creates a dataset that reads data from the tf.data service.

func DataServiceDatasetV2 ¶

func DataServiceDatasetV2(scope *Scope, dataset_id tf.Output, processing_mode tf.Output, address tf.Output, protocol tf.Output, job_name tf.Output, consumer_index tf.Output, num_consumers tf.Output, max_outstanding_requests tf.Output, iteration_counter tf.Output, output_types []tf.DataType, output_shapes []tf.Shape, optional ...DataServiceDatasetV2Attr) (handle tf.Output)

Creates a dataset that reads data from the tf.data service.

func DatasetCardinality ¶ added in v1.15.0

func DatasetCardinality(scope *Scope, input_dataset tf.Output, optional ...DatasetCardinalityAttr) (cardinality tf.Output)

Returns the cardinality of `input_dataset`.

Arguments:

input_dataset: A variant tensor representing the dataset to return cardinality for.

Returns The cardinality of `input_dataset`. Named constants are used to represent infinite and unknown cardinality.

func DatasetFingerprint ¶

func DatasetFingerprint(scope *Scope, input_dataset tf.Output) (fingerprint tf.Output)

Returns the fingerprint of `input_dataset`.

Arguments:

input_dataset: A variant tensor representing the dataset to return fingerprint for.

Returns The fingerprint of `input_dataset` in `uint64`

func DatasetFromGraph ¶

func DatasetFromGraph(scope *Scope, graph_def tf.Output) (handle tf.Output)

Creates a dataset from the given `graph_def`.

Creates a dataset from the provided `graph_def`.

Arguments:

graph_def: The graph representation of the dataset (as serialized GraphDef).

Returns A variant tensor representing the dataset.

func DatasetToGraph ¶ added in v1.11.0

func DatasetToGraph(scope *Scope, input_dataset tf.Output, optional ...DatasetToGraphAttr) (graph tf.Output)

Returns a serialized GraphDef representing `input_dataset`.

Returns a graph representation for `input_dataset`.

Arguments:

input_dataset: A variant tensor representing the dataset to return the graph representation for.

Returns The graph representation of the dataset (as serialized GraphDef).

func DatasetToGraphV2 ¶

func DatasetToGraphV2(scope *Scope, input_dataset tf.Output, optional ...DatasetToGraphV2Attr) (graph tf.Output)

Returns a serialized GraphDef representing `input_dataset`.

Returns a graph representation for `input_dataset`.

Arguments:

input_dataset: A variant tensor representing the dataset to return the graph representation for.

Returns The graph representation of the dataset (as serialized GraphDef).

func DatasetToSingleElement ¶ added in v1.5.0

func DatasetToSingleElement(scope *Scope, dataset tf.Output, output_types []tf.DataType, output_shapes []tf.Shape, optional ...DatasetToSingleElementAttr) (components []tf.Output)

Outputs the single element from the given dataset.

Arguments:

dataset: A handle to a dataset that contains a single element.

Returns The components of the single element of `input`.

func DatasetToTFRecord ¶ added in v1.11.0

func DatasetToTFRecord(scope *Scope, input_dataset tf.Output, filename tf.Output, compression_type tf.Output) (o *tf.Operation)

Writes the given dataset to the given file using the TFRecord format.

Arguments:

input_dataset: A variant tensor representing the dataset to write.
filename: A scalar string tensor representing the filename to use.
compression_type: A scalar string tensor containing either (i) the empty string (no

compression), (ii) "ZLIB", or (iii) "GZIP".

Returns the created operation.

func DebugGradientIdentity ¶ added in v1.4.0

func DebugGradientIdentity(scope *Scope, input tf.Output) (output tf.Output)

Identity op for gradient debugging.

This op is hidden from public in Python. It is used by TensorFlow Debugger to register gradient tensors for gradient debugging. This op operates on non-reference-type tensors.

func DebugIdentity ¶

func DebugIdentity(scope *Scope, input tf.Output, optional ...DebugIdentityAttr) (output tf.Output)

Provides an identity mapping of the non-Ref type input tensor for debugging.

Arguments:

input: Input tensor, non-Reference type

func DebugIdentityV2 ¶

func DebugIdentityV2(scope *Scope, input tf.Output, optional ...DebugIdentityV2Attr) (output tf.Output)

Debug Identity V2 Op.

Provides an identity mapping from input to output, while writing the content of the input tensor by calling DebugEventsWriter.

The semantics of the input tensor depends on tensor_debug_mode. In typical usage, the input tensor comes directly from the user computation only when graph_debug_mode is FULL_TENSOR (see protobuf/debug_event.proto for a list of all the possible values of graph_debug_mode). For the other debug modes, the input tensor should be produced by an additional op or subgraph that computes summary information about one or more tensors.

Arguments:

input: Input tensor, non-Reference type

func DebugIdentityV3 ¶

func DebugIdentityV3(scope *Scope, input tf.Output, optional ...DebugIdentityV3Attr) (output tf.Output)

Provides an identity mapping of the non-Ref type input tensor for debugging.

Arguments:

input: Input tensor, non-Reference type

func DebugNanCount ¶

func DebugNanCount(scope *Scope, input tf.Output, optional ...DebugNanCountAttr) (output tf.Output)

Debug NaN Value Counter Op.

Counts number of NaNs in the input tensor, for debugging.

Arguments:

input: Input tensor, non-Reference type.

func DebugNumericSummary ¶

func DebugNumericSummary(scope *Scope, input tf.Output, optional ...DebugNumericSummaryAttr) (output tf.Output)

Debug Numeric Summary Op.

Provide a basic summary of numeric value types, range and distribution.

output: A double tensor of shape [14 + nDimensions], where nDimensions is the

number of dimensions of the tensor's shape. The elements of output are:
[0]: is initialized (1.0) or not (0.0).
[1]: total number of elements
[2]: NaN element count
[3]: generalized -inf count: elements <= lower_bound. lower_bound is -inf by
  default.
[4]: negative element count (excluding -inf), if lower_bound is the default
  -inf. Otherwise, this is the count of elements > lower_bound and < 0.
[5]: zero element count
[6]: positive element count (excluding +inf), if upper_bound is the default
  +inf. Otherwise, this is the count of elements < upper_bound and > 0.
[7]: generalized +inf count, elements >= upper_bound. upper_bound is +inf by
  default.

Output elements [1:8] are all zero, if the tensor is uninitialized.

[8]: minimum of all non-inf and non-NaN elements.
     If uninitialized or no such element exists: +inf.
[9]: maximum of all non-inf and non-NaN elements.
     If uninitialized or no such element exists: -inf.
[10]: mean of all non-inf and non-NaN elements.
      If uninitialized or no such element exists: NaN.
[11]: variance of all non-inf and non-NaN elements.
      If uninitialized or no such element exists: NaN.
[12]: Data type of the tensor encoded as an enum integer. See the DataType
      proto for more details.
[13]: Number of dimensions of the tensor (ndims).
[14+]: Sizes of the dimensions.

Arguments:

input: Input tensor, non-Reference type.

func DebugNumericSummaryV2 ¶

func DebugNumericSummaryV2(scope *Scope, input tf.Output, optional ...DebugNumericSummaryV2Attr) (output tf.Output)

Debug Numeric Summary V2 Op.

Computes a numeric summary of the input tensor. The shape of the output depends on the tensor_debug_mode attribute. This op is used internally by TensorFlow Debugger (tfdbg) v2.

Arguments:

input: Input tensor, to be summarized by the op.

func DecodeAndCropJpeg ¶ added in v1.4.0

func DecodeAndCropJpeg(scope *Scope, contents tf.Output, crop_window tf.Output, optional ...DecodeAndCropJpegAttr) (image tf.Output)

Decode and Crop a JPEG-encoded image to a uint8 tensor.

The attr `channels` indicates the desired number of color channels for the decoded image.

Accepted values are:

* 0: Use the number of channels in the JPEG-encoded image. * 1: output a grayscale image. * 3: output an RGB image.

If needed, the JPEG-encoded image is transformed to match the requested number of color channels.

The attr `ratio` allows downscaling the image by an integer factor during decoding. Allowed values are: 1, 2, 4, and 8. This is much faster than downscaling the image later.

It is equivalent to a combination of decode and crop, but much faster by only decoding partial jpeg image.

Arguments:

contents: 0-D.  The JPEG-encoded image.
crop_window: 1-D.  The crop window: [crop_y, crop_x, crop_height, crop_width].

Returns 3-D with shape `[height, width, channels]`..

func DecodeBase64 ¶ added in v1.1.0

func DecodeBase64(scope *Scope, input tf.Output) (output tf.Output)

Decode web-safe base64-encoded strings.

Input may or may not have padding at the end. See EncodeBase64(https://www.tensorflow.org/api_docs/python/tf/io/encode_base64) for padding. Web-safe means that input must use - and _ instead of + and /.

Arguments:

input: Base64 strings to decode.

Returns Decoded strings.

func DecodeBmp ¶ added in v1.3.0

func DecodeBmp(scope *Scope, contents tf.Output, optional ...DecodeBmpAttr) (image tf.Output)

Decode the first frame of a BMP-encoded image to a uint8 tensor.

The attr `channels` indicates the desired number of color channels for the decoded image.

Accepted values are:

* 0: Use the number of channels in the BMP-encoded image. * 3: output an RGB image. * 4: output an RGBA image.

Arguments:

contents: 0-D.  The BMP-encoded image.

Returns 3-D with shape `[height, width, channels]`. RGB order

func DecodeCSV ¶ added in v1.1.0

func DecodeCSV(scope *Scope, records tf.Output, record_defaults []tf.Output, optional ...DecodeCSVAttr) (output []tf.Output)

Convert CSV records to tensors. Each column maps to one tensor.

RFC 4180 format is expected for the CSV records. (https://tools.ietf.org/html/rfc4180) Note that we allow leading and trailing spaces with int or float field.

Arguments:

records: Each string is a record/row in the csv and all records should have

the same format.

record_defaults: One tensor per column of the input record, with either a

scalar default value for that column or an empty vector if the column is required.

Returns Each tensor will have the same shape as records.

func DecodeCompressed ¶ added in v1.6.0

func DecodeCompressed(scope *Scope, bytes tf.Output, optional ...DecodeCompressedAttr) (output tf.Output)

Decompress strings.

This op decompresses each element of the `bytes` input `Tensor`, which is assumed to be compressed using the given `compression_type`.

The `output` is a string `Tensor` of the same shape as `bytes`, each element containing the decompressed data from the corresponding element in `bytes`.

Arguments:

bytes: A Tensor of string which is compressed.

Returns A Tensor with the same shape as input `bytes`, uncompressed from bytes.

func DecodeGif ¶ added in v1.1.0

func DecodeGif(scope *Scope, contents tf.Output) (image tf.Output)

Decode the frame(s) of a GIF-encoded image to a uint8 tensor.

GIF images with frame or transparency compression are not supported. On Linux and MacOS systems, convert animated GIFs from compressed to uncompressed by running:

convert $src.gif -coalesce $dst.gif

This op also supports decoding JPEGs and PNGs, though it is cleaner to use `tf.io.decode_image`.

Arguments:

contents: 0-D.  The GIF-encoded image.

Returns 4-D with shape `[num_frames, height, width, 3]`. RGB channel order.

func DecodeImage ¶

func DecodeImage(scope *Scope, contents tf.Output, optional ...DecodeImageAttr) (image tf.Output)

Function for decode_bmp, decode_gif, decode_jpeg, and decode_png.

Detects whether an image is a BMP, GIF, JPEG, or PNG, and performs the appropriate operation to convert the input bytes string into a Tensor of type dtype.

*NOTE*: decode_gif returns a 4-D array [num_frames, height, width, 3], as opposed to decode_bmp, decode_jpeg and decode_png, which return 3-D arrays [height, width, num_channels]. Make sure to take this into account when constructing your graph if you are intermixing GIF files with BMP, JPEG, and/or PNG files. Alternately, set the expand_animations argument of this function to False, in which case the op will return 3-dimensional tensors and will truncate animated GIF files to the first frame.

*NOTE*: If the first frame of an animated GIF does not occupy the entire canvas (maximum frame width x maximum frame height), then it fills the unoccupied areas (in the first frame) with zeros (black). For frames after the first frame that does not occupy the entire canvas, it uses the previous frame to fill the unoccupied areas.

Arguments:

contents: 0-D. The encoded image bytes.

Returns 3-D with shape `[height, width, channels]` or 4-D with shape `[frame, height, width, channels]`..

func DecodeJSONExample ¶ added in v1.1.0

func DecodeJSONExample(scope *Scope, json_examples tf.Output) (binary_examples tf.Output)

Convert JSON-encoded Example records to binary protocol buffer strings.

Note: This is **not** a general purpose JSON parsing op.

This op converts JSON-serialized `tf.train.Example` (created with `json_format.MessageToJson`, following the [standard JSON mapping](https://developers.google.com/protocol-buffers/docs/proto3#json)) to a binary-serialized `tf.train.Example` (equivalent to `Example.SerializeToString()`) suitable for conversion to tensors with `tf.io.parse_example`.

Arguments:

json_examples: Each string is a JSON object serialized according to the JSON

mapping of the Example proto.

Returns Each string is a binary Example protocol buffer corresponding to the respective element of `json_examples`.

func DecodeJpeg ¶ added in v1.1.0

func DecodeJpeg(scope *Scope, contents tf.Output, optional ...DecodeJpegAttr) (image tf.Output)

Decode a JPEG-encoded image to a uint8 tensor.

The attr `channels` indicates the desired number of color channels for the decoded image.

Accepted values are:

* 0: Use the number of channels in the JPEG-encoded image. * 1: output a grayscale image. * 3: output an RGB image.

If needed, the JPEG-encoded image is transformed to match the requested number of color channels.

The attr `ratio` allows downscaling the image by an integer factor during decoding. Allowed values are: 1, 2, 4, and 8. This is much faster than downscaling the image later.

This op also supports decoding PNGs and non-animated GIFs since the interface is the same, though it is cleaner to use `tf.io.decode_image`.

Arguments:

contents: 0-D.  The JPEG-encoded image.

Returns 3-D with shape `[height, width, channels]`..

func DecodePaddedRaw ¶ added in v1.14.0

func DecodePaddedRaw(scope *Scope, input_bytes tf.Output, fixed_length tf.Output, out_type tf.DataType, optional ...DecodePaddedRawAttr) (output tf.Output)

Reinterpret the bytes of a string as a vector of numbers.

Arguments:

input_bytes: Tensor of string to be decoded.
fixed_length: Length in bytes for each element of the decoded output. Must be a multiple

of the size of the output type.

Returns A Tensor with one more dimension than the input `bytes`. The added dimension will have size equal to the length of the elements of `bytes` divided by the number of bytes to represent `out_type`.

func DecodePng ¶ added in v1.1.0

func DecodePng(scope *Scope, contents tf.Output, optional ...DecodePngAttr) (image tf.Output)

Decode a PNG-encoded image to a uint8 or uint16 tensor.

The attr `channels` indicates the desired number of color channels for the decoded image.

Accepted values are:

* 0: Use the number of channels in the PNG-encoded image. * 1: output a grayscale image. * 3: output an RGB image. * 4: output an RGBA image.

If needed, the PNG-encoded image is transformed to match the requested number of color channels.

This op also supports decoding JPEGs and non-animated GIFs since the interface is the same, though it is cleaner to use `tf.io.decode_image`.

Arguments:

contents: 0-D.  The PNG-encoded image.

Returns 3-D with shape `[height, width, channels]`.

func DecodeProtoV2 ¶ added in v1.9.0

func DecodeProtoV2(scope *Scope, bytes tf.Output, message_type string, field_names []string, output_types []tf.DataType, optional ...DecodeProtoV2Attr) (sizes tf.Output, values []tf.Output)

The op extracts fields from a serialized protocol buffers message into tensors.

Note: This API is designed for orthogonality rather than human-friendliness. It can be used to parse input protos by hand, but it is intended for use in generated code.

The `decode_proto` op extracts fields from a serialized protocol buffers message into tensors. The fields in `field_names` are decoded and converted to the corresponding `output_types` if possible.

A `message_type` name must be provided to give context for the field names. The actual message descriptor can be looked up either in the linked-in descriptor pool or a filename provided by the caller using the `descriptor_source` attribute.

Each output tensor is a dense tensor. This means that it is padded to hold the largest number of repeated elements seen in the input minibatch. (The shape is also padded by one to prevent zero-sized dimensions). The actual repeat counts for each example in the minibatch can be found in the `sizes` output. In many cases the output of `decode_proto` is fed immediately into tf.squeeze if missing values are not a concern. When using tf.squeeze, always pass the squeeze dimension explicitly to avoid surprises.

For the most part, the mapping between Proto field types and TensorFlow dtypes is straightforward. However, there are a few special cases:

- A proto field that contains a submessage or group can only be converted to `DT_STRING` (the serialized submessage). This is to reduce the complexity of the API. The resulting string can be used as input to another instance of the decode_proto op.

- TensorFlow lacks support for unsigned integers. The ops represent uint64 types as a `DT_INT64` with the same twos-complement bit pattern (the obvious way). Unsigned int32 values can be represented exactly by specifying type `DT_INT64`, or using twos-complement if the caller specifies `DT_INT32` in the `output_types` attribute.

- `map` fields are not directly decoded. They are treated as `repeated` fields, of the appropriate entry type. The proto-compiler defines entry types for each map field. The type-name is the field name, converted to "CamelCase" with "Entry" appended. The `tf.train.Features.FeatureEntry` message is an example of one of these implicit `Entry` types.

- `enum` fields should be read as int32.

Both binary and text proto serializations are supported, and can be chosen using the `format` attribute.

The `descriptor_source` attribute selects the source of protocol descriptors to consult when looking up `message_type`. This may be:

- An empty string or "local://", in which case protocol descriptors are created for C++ (not Python) proto definitions linked to the binary.

- A file, in which case protocol descriptors are created from the file, which is expected to contain a `FileDescriptorSet` serialized as a string. NOTE: You can build a `descriptor_source` file using the `--descriptor_set_out` and `--include_imports` options to the protocol compiler `protoc`.

- A "bytes://<bytes>", in which protocol descriptors are created from `<bytes>`, which is expected to be a `FileDescriptorSet` serialized as a string.

Arguments:

bytes: Tensor of serialized protos with shape `batch_shape`.
message_type: Name of the proto message type to decode.
field_names: List of strings containing proto field names. An extension field can be decoded

by using its full name, e.g. EXT_PACKAGE.EXT_FIELD_NAME.

output_types: List of TF types to use for the respective field in field_names.

Returns:

sizes: Tensor of int32 with shape `[batch_shape, len(field_names)]`.

Each entry is the number of values found for the corresponding field. Optional fields may have 0 or 1 values.

values: List of tensors containing values for the corresponding field.

`values[i]` has datatype `output_types[i]` and shape `[batch_shape, max(sizes[...,i])]`.

func DecodeRaw ¶ added in v1.1.0

func DecodeRaw(scope *Scope, bytes tf.Output, out_type tf.DataType, optional ...DecodeRawAttr) (output tf.Output)

Reinterpret the bytes of a string as a vector of numbers.

Arguments:

bytes: All the elements must have the same length.

Returns A Tensor with one more dimension than the input `bytes`. The added dimension will have size equal to the length of the elements of `bytes` divided by the number of bytes to represent `out_type`.

func DecodeWav ¶ added in v1.2.0

func DecodeWav(scope *Scope, contents tf.Output, optional ...DecodeWavAttr) (audio tf.Output, sample_rate tf.Output)

Decode a 16-bit PCM WAV file to a float tensor.

The -32768 to 32767 signed 16-bit values will be scaled to -1.0 to 1.0 in float.

When desired_channels is set, if the input contains fewer channels than this then the last channel will be duplicated to give the requested number, else if the input has more channels than requested then the additional channels will be ignored.

If desired_samples is set, then the audio will be cropped or padded with zeroes to the requested length.

The first output contains a Tensor with the content of the audio samples. The lowest dimension will be the number of channels, and the second will be the number of samples. For example, a ten-sample-long stereo WAV file should give an output shape of [10, 2].

Arguments:

contents: The WAV-encoded audio, usually from a file.

Returns:

audio: 2-D with shape `[length, channels]`.
sample_rate: Scalar holding the sample rate found in the WAV header.

func DeepCopy ¶ added in v1.9.0

func DeepCopy(scope *Scope, x tf.Output) (y tf.Output)

Makes a copy of `x`.

Arguments:

x: The source tensor of type `T`.

Returns y: A `Tensor` of type `T`. A copy of `x`. Guaranteed that `y`

is not an alias of `x`.

func DeleteIterator ¶ added in v1.12.1

func DeleteIterator(scope *Scope, handle tf.Output, deleter tf.Output) (o *tf.Operation)

A container for an iterator resource.

Arguments:

handle: A handle to the iterator to delete.
deleter: A variant deleter.

Returns the created operation.

func DeleteMultiDeviceIterator ¶

func DeleteMultiDeviceIterator(scope *Scope, multi_device_iterator tf.Output, iterators []tf.Output, deleter tf.Output) (o *tf.Operation)

A container for an iterator resource.

Arguments:

multi_device_iterator: A handle to the multi device iterator to delete.
iterators: A list of iterator handles (unused). This is added so that automatic control dependencies get added during function tracing that ensure this op runs after all the dependent iterators are deleted.
deleter: A variant deleter.

Returns the created operation.

func DeleteSessionTensor ¶ added in v1.1.0

func DeleteSessionTensor(scope *Scope, handle tf.Output) (o *tf.Operation)

Delete the tensor specified by its handle in the session.

Arguments:

handle: The handle for a tensor stored in the session state.

Returns the created operation.

func DenseBincount ¶

func DenseBincount(scope *Scope, input tf.Output, size tf.Output, weights tf.Output, optional ...DenseBincountAttr) (output tf.Output)

Counts the number of occurrences of each value in an integer array.

Outputs a vector with length `size` and the same dtype as `weights`. If `weights` are empty, then index `i` stores the number of times the value `i` is counted in `arr`. If `weights` are non-empty, then index `i` stores the sum of the value in `weights` at each index where the corresponding value in `arr` is `i`.

Values in `arr` outside of the range [0, size) are ignored.

Arguments:

input: 1D or 2D int `Tensor`.
size: non-negative int scalar `Tensor`.
weights: is an int32, int64, float32, or float64 `Tensor` with the same

shape as `arr`, or a length-0 `Tensor`, in which case it acts as all weights equal to 1.

Returns 1D `Tensor` with length equal to `size` or 2D `Tensor` with [batch_size, `size`]. The counts or summed weights for each value in the range [0, size).

func DenseCountSparseOutput ¶

func DenseCountSparseOutput(scope *Scope, values tf.Output, weights tf.Output, binary_output bool, optional ...DenseCountSparseOutputAttr) (output_indices tf.Output, output_values tf.Output, output_dense_shape tf.Output)

Performs sparse-output bin counting for a tf.tensor input.

Counts the number of times each value occurs in the input.

Arguments:

values: Tensor containing data to count.
weights: A Tensor of the same shape as indices containing per-index weight values. May

also be the empty tensor if no weights are used.

binary_output: Whether to output the number of occurrences of each value or 1.

Returns:

output_indices: Indices tensor for the resulting sparse tensor object.
output_values: Values tensor for the resulting sparse tensor object.
output_dense_shape: Shape tensor for the resulting sparse tensor object.

func DenseToCSRSparseMatrix ¶

func DenseToCSRSparseMatrix(scope *Scope, dense_input tf.Output, indices tf.Output) (sparse_output tf.Output)

Converts a dense tensor to a (possibly batched) CSRSparseMatrix.

Arguments:

dense_input: A Dense tensor.
indices: Indices of nonzero elements.

Returns A (possibly batched) CSRSparseMatrix.

func DenseToDenseSetOperation ¶ added in v1.1.0

func DenseToDenseSetOperation(scope *Scope, set1 tf.Output, set2 tf.Output, set_operation string, optional ...DenseToDenseSetOperationAttr) (result_indices tf.Output, result_values tf.Output, result_shape tf.Output)

Applies set operation along last dimension of 2 `Tensor` inputs.

See SetOperationOp::SetOperationFromContext for values of `set_operation`.

Output `result` is a `SparseTensor` represented by `result_indices`, `result_values`, and `result_shape`. For `set1` and `set2` ranked `n`, this has rank `n` and the same 1st `n-1` dimensions as `set1` and `set2`. The `nth` dimension contains the result of `set_operation` applied to the corresponding `[0...n-1]` dimension of `set`.

Arguments:

set1: `Tensor` with rank `n`. 1st `n-1` dimensions must be the same as `set2`.

Dimension `n` contains values in a set, duplicates are allowed but ignored.

set2: `Tensor` with rank `n`. 1st `n-1` dimensions must be the same as `set1`.

Dimension `n` contains values in a set, duplicates are allowed but ignored.

Returns:

result_indices: 2D indices of a `SparseTensor`.
result_values: 1D values of a `SparseTensor`.
result_shape: 1D `Tensor` shape of a `SparseTensor`. `result_shape[0...n-1]` is

the same as the 1st `n-1` dimensions of `set1` and `set2`, `result_shape[n]` is the max result set size across all `0...n-1` dimensions.

func DenseToSparseBatchDataset ¶ added in v1.2.0

func DenseToSparseBatchDataset(scope *Scope, input_dataset tf.Output, batch_size tf.Output, row_shape tf.Output, output_types []tf.DataType, output_shapes []tf.Shape) (handle tf.Output)

Creates a dataset that batches input elements into a SparseTensor.

Arguments:

input_dataset: A handle to an input dataset. Must have a single component.
batch_size: A scalar representing the number of elements to accumulate in a

batch.

row_shape: A vector representing the dense shape of each row in the produced

SparseTensor. The shape may be partially specified, using `-1` to indicate that a particular dimension should use the maximum size of all batch elements.

func DenseToSparseSetOperation ¶ added in v1.1.0

func DenseToSparseSetOperation(scope *Scope, set1 tf.Output, set2_indices tf.Output, set2_values tf.Output, set2_shape tf.Output, set_operation string, optional ...DenseToSparseSetOperationAttr) (result_indices tf.Output, result_values tf.Output, result_shape tf.Output)

Applies set operation along last dimension of `Tensor` and `SparseTensor`.

See SetOperationOp::SetOperationFromContext for values of `set_operation`.

Input `set2` is a `SparseTensor` represented by `set2_indices`, `set2_values`, and `set2_shape`. For `set2` ranked `n`, 1st `n-1` dimensions must be the same as `set1`. Dimension `n` contains values in a set, duplicates are allowed but ignored.

If `validate_indices` is `True`, this op validates the order and range of `set2` indices.

Output `result` is a `SparseTensor` represented by `result_indices`, `result_values`, and `result_shape`. For `set1` and `set2` ranked `n`, this has rank `n` and the same 1st `n-1` dimensions as `set1` and `set2`. The `nth` dimension contains the result of `set_operation` applied to the corresponding `[0...n-1]` dimension of `set`.

Arguments:

set1: `Tensor` with rank `n`. 1st `n-1` dimensions must be the same as `set2`.

Dimension `n` contains values in a set, duplicates are allowed but ignored.

set2_indices: 2D `Tensor`, indices of a `SparseTensor`. Must be in row-major

order.

set2_values: 1D `Tensor`, values of a `SparseTensor`. Must be in row-major

order.

set2_shape: 1D `Tensor`, shape of a `SparseTensor`. `set2_shape[0...n-1]` must

be the same as the 1st `n-1` dimensions of `set1`, `result_shape[n]` is the max set size across `n-1` dimensions.

Returns:

result_indices: 2D indices of a `SparseTensor`.
result_values: 1D values of a `SparseTensor`.
result_shape: 1D `Tensor` shape of a `SparseTensor`. `result_shape[0...n-1]` is

the same as the 1st `n-1` dimensions of `set1` and `set2`, `result_shape[n]` is the max result set size across all `0...n-1` dimensions.

func DepthToSpace ¶ added in v1.1.0

func DepthToSpace(scope *Scope, input tf.Output, block_size int64, optional ...DepthToSpaceAttr) (output tf.Output)

DepthToSpace for tensors of type T.

Rearranges data from depth into blocks of spatial data. This is the reverse transformation of SpaceToDepth. More specifically, this op outputs a copy of the input tensor where values from the `depth` dimension are moved in spatial blocks to the `height` and `width` dimensions. The attr `block_size` indicates the input block size and how the data is moved.

Chunks of data of size `block_size * block_size` from depth are rearranged into non-overlapping blocks of size `block_size x block_size`
The width of the output tensor is `input_depth * block_size`, whereas the height is `input_height * block_size`.
The Y, X coordinates within each block of the output image are determined by the high order component of the input channel index.
The depth of the input tensor must be divisible by `block_size * block_size`.

The `data_format` attr specifies the layout of the input and output tensors with the following options:

"NHWC": `[ batch, height, width, channels ]`
"NCHW": `[ batch, channels, height, width ]`
"NCHW_VECT_C":
    `qint8 [ batch, channels / 4, height, width, 4 ]`

It is useful to consider the operation as transforming a 6-D Tensor. e.g. for data_format = NHWC,

Each element in the input tensor can be specified via 6 coordinates,
ordered by decreasing memory layout significance as:
n,iY,iX,bY,bX,oC  (where n=batch index, iX, iY means X or Y coordinates
                   within the input image, bX, bY means coordinates
                   within the output block, oC means output channels).
The output would be the input transposed to the following layout:
n,iY,bY,iX,bX,oC

This operation is useful for resizing the activations between convolutions (but keeping all data), e.g. instead of pooling. It is also useful for training purely convolutional models.

For example, given an input of shape `[1, 1, 1, 4]`, data_format = "NHWC" and block_size = 2:

``` x = [[[[1, 2, 3, 4]]]]

```

This operation will output a tensor of shape `[1, 2, 2, 1]`:

```

[[[[1], [2]],
  [[3], [4]]]]

```

Here, the input has a batch of 1 and each batch element has shape `[1, 1, 4]`, the corresponding output will have 2x2 elements and will have a depth of 1 channel (1 = `4 / (block_size * block_size)`). The output element shape is `[2, 2, 1]`.

For an input tensor with larger depth, here of shape `[1, 1, 1, 12]`, e.g.

``` x = [[[[1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12]]]] ```

This operation, for block size of 2, will return the following tensor of shape `[1, 2, 2, 3]`

```

[[[[1, 2, 3], [4, 5, 6]],
  [[7, 8, 9], [10, 11, 12]]]]

```

Similarly, for the following input of shape `[1 2 2 4]`, and a block size of 2:

``` x = [[[[1, 2, 3, 4],

 [5, 6, 7, 8]],
[[9, 10, 11, 12],
 [13, 14, 15, 16]]]]

```

the operator will return the following tensor of shape `[1 4 4 1]`:

``` x = [[[ [1], [2], [5], [6]],

[ [3],   [4],  [7],  [8]],
[ [9],  [10], [13],  [14]],
[ [11], [12], [15],  [16]]]]

```

Arguments:

block_size: The size of the spatial block, same as in Space2Depth.

func DepthwiseConv2dNative ¶ added in v1.1.0

func DepthwiseConv2dNative(scope *Scope, input tf.Output, filter tf.Output, strides []int64, padding string, optional ...DepthwiseConv2dNativeAttr) (output tf.Output)

Computes a 2-D depthwise convolution given 4-D `input` and `filter` tensors.

Given an input tensor of shape `[batch, in_height, in_width, in_channels]` and a filter / kernel tensor of shape `[filter_height, filter_width, in_channels, channel_multiplier]`, containing `in_channels` convolutional filters of depth 1, `depthwise_conv2d` applies a different filter to each input channel (expanding from 1 channel to `channel_multiplier` channels for each), then concatenates the results together. Thus, the output has `in_channels * channel_multiplier` channels.

``` for k in 0..in_channels-1

for q in 0..channel_multiplier-1
  output[b, i, j, k * channel_multiplier + q] =
    sum_{di, dj} input[b, strides[1] * i + di, strides[2] * j + dj, k] *
                      filter[di, dj, k, q]

```

Must have `strides[0] = strides[3] = 1`. For the most common case of the same horizontal and vertices strides, `strides = [1, stride, stride, 1]`.

Arguments:

strides: 1-D of length 4.  The stride of the sliding window for each dimension

of `input`.

padding: The type of padding algorithm to use.

func DepthwiseConv2dNativeBackpropFilter ¶ added in v1.1.0

func DepthwiseConv2dNativeBackpropFilter(scope *Scope, input tf.Output, filter_sizes tf.Output, out_backprop tf.Output, strides []int64, padding string, optional ...DepthwiseConv2dNativeBackpropFilterAttr) (output tf.Output)

Computes the gradients of depthwise convolution with respect to the filter.

Arguments:

input: 4-D with shape based on `data_format`.  For example, if

`data_format` is 'NHWC' then `input` is a 4-D `[batch, in_height, in_width, in_channels]` tensor.

filter_sizes: An integer vector representing the tensor shape of `filter`,

where `filter` is a 4-D `[filter_height, filter_width, in_channels, depthwise_multiplier]` tensor.

out_backprop: 4-D with shape  based on `data_format`.

For example, if `data_format` is 'NHWC' then out_backprop shape is `[batch, out_height, out_width, out_channels]`. Gradients w.r.t. the output of the convolution.

strides: The stride of the sliding window for each dimension of the input

of the convolution.

padding: The type of padding algorithm to use.

Returns 4-D with shape `[filter_height, filter_width, in_channels, out_channels]`. Gradient w.r.t. the `filter` input of the convolution.

func DepthwiseConv2dNativeBackpropInput ¶ added in v1.1.0

func DepthwiseConv2dNativeBackpropInput(scope *Scope, input_sizes tf.Output, filter tf.Output, out_backprop tf.Output, strides []int64, padding string, optional ...DepthwiseConv2dNativeBackpropInputAttr) (output tf.Output)

Computes the gradients of depthwise convolution with respect to the input.

Arguments:

input_sizes: An integer vector representing the shape of `input`, based

on `data_format`. For example, if `data_format` is 'NHWC' then

 `input` is a 4-D `[batch, height, width, channels]` tensor.
	filter: 4-D with shape

`[filter_height, filter_width, in_channels, depthwise_multiplier]`.

out_backprop: 4-D with shape  based on `data_format`.

For example, if `data_format` is 'NHWC' then out_backprop shape is `[batch, out_height, out_width, out_channels]`. Gradients w.r.t. the output of the convolution.

strides: The stride of the sliding window for each dimension of the input

of the convolution.

padding: The type of padding algorithm to use.

Returns 4-D with shape according to `data_format`. For example, if `data_format` is 'NHWC', output shape is `[batch, in_height, in_width, in_channels]`. Gradient w.r.t. the input of the convolution.

func Dequantize ¶ added in v1.1.0

func Dequantize(scope *Scope, input tf.Output, min_range tf.Output, max_range tf.Output, optional ...DequantizeAttr) (output tf.Output)

Dequantize the 'input' tensor into a float or bfloat16 Tensor.

[min_range, max_range] are scalar floats that specify the range for the output. The 'mode' attribute controls exactly which calculations are used to convert the float values to their quantized equivalents.

In 'MIN_COMBINED' mode, each value of the tensor will undergo the following:

``` if T == qint8: in[i] += (range(T) + 1)/ 2.0 out[i] = min_range + (in[i]* (max_range - min_range) / range(T)) ``` here `range(T) = numeric_limits<T>::max() - numeric_limits<T>::min()`

*MIN_COMBINED Mode Example*

If the input comes from a QuantizedRelu6, the output type is quint8 (range of 0-255) but the possible range of QuantizedRelu6 is 0-6. The min_range and max_range values are therefore 0.0 and 6.0. Dequantize on quint8 will take each value, cast to float, and multiply by 6 / 255. Note that if quantizedtype is qint8, the operation will additionally add each value by 128 prior to casting.

If the mode is 'MIN_FIRST', then this approach is used:

```c++ num_discrete_values = 1 << (# of bits in T) range_adjust = num_discrete_values / (num_discrete_values - 1) range = (range_max - range_min) * range_adjust range_scale = range / num_discrete_values const double offset_input = static_cast<double>(input) - lowest_quantized; result = range_min + ((input - numeric_limits<T>::min()) * range_scale) ```

If the mode is `SCALED`, dequantization is performed by multiplying each input value by a scaling_factor. (Thus an input of 0 always maps to 0.0).

The scaling_factor is determined from `min_range`, `max_range`, and `narrow_range` in a way that is compatible with `QuantizeAndDequantize{V2|V3}` and `QuantizeV2`, using the following algorithm:

```c++

const int min_expected_T = std::numeric_limits<T>::min() +
  (narrow_range ? 1 : 0);
const int max_expected_T = std::numeric_limits<T>::max();
const float max_expected_T = std::numeric_limits<float>::max();

const float scale_factor =
  (std::numeric_limits<T>::min() == 0) ? (max_range / max_expected_T)
                                       : std::max(min_range / min_expected_T,
                                                  max_range / max_expected_T);

```

Arguments:

min_range: The minimum scalar value possibly produced for the input.
max_range: The maximum scalar value possibly produced for the input.

func DeserializeIterator ¶ added in v1.5.0

func DeserializeIterator(scope *Scope, resource_handle tf.Output, serialized tf.Output) (o *tf.Operation)

Converts the given variant tensor to an iterator and stores it in the given resource.

Arguments:

resource_handle: A handle to an iterator resource.
serialized: A variant tensor storing the state of the iterator contained in the

resource.

Returns the created operation.

func DeserializeManySparse ¶ added in v1.1.0

func DeserializeManySparse(scope *Scope, serialized_sparse tf.Output, dtype tf.DataType) (sparse_indices tf.Output, sparse_values tf.Output, sparse_shape tf.Output)

Deserialize and concatenate `SparseTensors` from a serialized minibatch.

The input `serialized_sparse` must be a string matrix of shape `[N x 3]` where `N` is the minibatch size and the rows correspond to packed outputs of `SerializeSparse`. The ranks of the original `SparseTensor` objects must all match. When the final `SparseTensor` is created, it has rank one higher than the ranks of the incoming `SparseTensor` objects (they have been concatenated along a new row dimension).

The output `SparseTensor` object's shape values for all dimensions but the first are the max across the input `SparseTensor` objects' shape values for the corresponding dimensions. Its first shape value is `N`, the minibatch size.

The input `SparseTensor` objects' indices are assumed ordered in standard lexicographic order. If this is not the case, after this step run `SparseReorder` to restore index ordering.

For example, if the serialized input is a `[2 x 3]` matrix representing two original `SparseTensor` objects:

index = [ 0]
        [10]
        [20]
values = [1, 2, 3]
shape = [50]

and

index = [ 2]
        [10]
values = [4, 5]
shape = [30]

then the final deserialized `SparseTensor` will be:

index = [0  0]
        [0 10]
        [0 20]
        [1  2]
        [1 10]
values = [1, 2, 3, 4, 5]
shape = [2 50]

Arguments:

serialized_sparse: 2-D, The `N` serialized `SparseTensor` objects.

Must have 3 columns.

dtype: The `dtype` of the serialized `SparseTensor` objects.

func DeserializeSparse ¶ added in v1.5.0

func DeserializeSparse(scope *Scope, serialized_sparse tf.Output, dtype tf.DataType) (sparse_indices tf.Output, sparse_values tf.Output, sparse_shape tf.Output)

Deserialize `SparseTensor` objects.

The input `serialized_sparse` must have the shape `[?, ?, ..., ?, 3]` where the last dimension stores serialized `SparseTensor` objects and the other N dimensions (N >= 0) correspond to a batch. The ranks of the original `SparseTensor` objects must all match. When the final `SparseTensor` is created, its rank is the rank of the incoming `SparseTensor` objects plus N; the sparse tensors have been concatenated along new dimensions, one for each batch.

The output `SparseTensor` object's shape values for the original dimensions are the max across the input `SparseTensor` objects' shape values for the corresponding dimensions. The new dimensions match the size of the batch.

The input `SparseTensor` objects' indices are assumed ordered in standard lexicographic order. If this is not the case, after this step run `SparseReorder` to restore index ordering.

For example, if the serialized input is a `[2 x 3]` matrix representing two original `SparseTensor` objects:

index = [ 0]
        [10]
        [20]
values = [1, 2, 3]
shape = [50]

and

index = [ 2]
        [10]
values = [4, 5]
shape = [30]

then the final deserialized `SparseTensor` will be:

index = [0  0]
        [0 10]
        [0 20]
        [1  2]
        [1 10]
values = [1, 2, 3, 4, 5]
shape = [2 50]

Arguments:

serialized_sparse: The serialized `SparseTensor` objects. The last dimension

must have 3 columns.

dtype: The `dtype` of the serialized `SparseTensor` objects.

func DestroyResourceOp ¶ added in v1.1.0

func DestroyResourceOp(scope *Scope, resource tf.Output, optional ...DestroyResourceOpAttr) (o *tf.Operation)

Deletes the resource specified by the handle.

All subsequent operations using the resource will result in a NotFound error status.

Arguments:

resource: handle to the resource to delete.

Returns the created operation.

func DeviceIndex ¶

func DeviceIndex(scope *Scope, device_names []string) (index tf.Output)

Return the index of device the op runs.

Given a list of device names, this operation returns the index of the device this op runs. The length of the list is returned in two cases: (1) Device does not exist in the given device list. (2) It is in XLA compilation.

func Diag ¶ added in v1.1.0

func Diag(scope *Scope, diagonal tf.Output) (output tf.Output)

Returns a diagonal tensor with a given diagonal values.

Given a `diagonal`, this operation returns a tensor with the `diagonal` and everything else padded with zeros. The diagonal is computed as follows:

Assume `diagonal` has dimensions [D1,..., Dk], then the output is a tensor of rank 2k with dimensions [D1,..., Dk, D1,..., Dk] where:

`output[i1,..., ik, i1,..., ik] = diagonal[i1, ..., ik]` and 0 everywhere else.

For example:

``` # 'diagonal' is [1, 2, 3, 4] tf.diag(diagonal) ==> [[1, 0, 0, 0]

[0, 2, 0, 0]
[0, 0, 3, 0]
[0, 0, 0, 4]]

```

Arguments:

diagonal: Rank k tensor where k is at most 1.

func DiagPart ¶ added in v1.1.0

func DiagPart(scope *Scope, input tf.Output) (diagonal tf.Output)

Returns the diagonal part of the tensor.

This operation returns a tensor with the `diagonal` part of the `input`. The `diagonal` part is computed as follows:

Assume `input` has dimensions `[D1,..., Dk, D1,..., Dk]`, then the output is a tensor of rank `k` with dimensions `[D1,..., Dk]` where:

`diagonal[i1,..., ik] = input[i1, ..., ik, i1,..., ik]`.

For example:

``` # 'input' is [[1, 0, 0, 0]

[0, 2, 0, 0]
[0, 0, 3, 0]
[0, 0, 0, 4]]

tf.diag_part(input) ==> [1, 2, 3, 4] ```

Arguments:

input: Rank k tensor where k is even and not zero.

Returns The extracted diagonal.

func Digamma ¶ added in v1.1.0

func Digamma(scope *Scope, x tf.Output) (y tf.Output)

Computes Psi, the derivative of Lgamma (the log of the absolute value of

`Gamma(x)`), element-wise.

func Dilation2D ¶ added in v1.1.0

func Dilation2D(scope *Scope, input tf.Output, filter tf.Output, strides []int64, rates []int64, padding string) (output tf.Output)

Computes the grayscale dilation of 4-D `input` and 3-D `filter` tensors.

The `input` tensor has shape `[batch, in_height, in_width, depth]` and the `filter` tensor has shape `[filter_height, filter_width, depth]`, i.e., each input channel is processed independently of the others with its own structuring function. The `output` tensor has shape `[batch, out_height, out_width, depth]`. The spatial dimensions of the output tensor depend on the `padding` algorithm. We currently only support the default "NHWC" `data_format`.

In detail, the grayscale morphological 2-D dilation is the max-sum correlation (for consistency with `conv2d`, we use unmirrored filters):

output[b, y, x, c] =
   max_{dy, dx} input[b,
                      strides[1] * y + rates[1] * dy,
                      strides[2] * x + rates[2] * dx,
                      c] +
                filter[dy, dx, c]

Max-pooling is a special case when the filter has size equal to the pooling kernel size and contains all zeros.

Note on duality: The dilation of `input` by the `filter` is equal to the negation of the erosion of `-input` by the reflected `filter`.

Arguments:

input: 4-D with shape `[batch, in_height, in_width, depth]`.
filter: 3-D with shape `[filter_height, filter_width, depth]`.
strides: The stride of the sliding window for each dimension of the input

tensor. Must be: `[1, stride_height, stride_width, 1]`.

rates: The input stride for atrous morphological dilation. Must be:

`[1, rate_height, rate_width, 1]`.

padding: The type of padding algorithm to use.

Returns 4-D with shape `[batch, out_height, out_width, depth]`.

func Dilation2DBackpropFilter ¶ added in v1.1.0

func Dilation2DBackpropFilter(scope *Scope, input tf.Output, filter tf.Output, out_backprop tf.Output, strides []int64, rates []int64, padding string) (filter_backprop tf.Output)

Computes the gradient of morphological 2-D dilation with respect to the filter.

Arguments:

input: 4-D with shape `[batch, in_height, in_width, depth]`.
filter: 3-D with shape `[filter_height, filter_width, depth]`.
out_backprop: 4-D with shape `[batch, out_height, out_width, depth]`.
strides: 1-D of length 4. The stride of the sliding window for each dimension of

the input tensor. Must be: `[1, stride_height, stride_width, 1]`.

rates: 1-D of length 4. The input stride for atrous morphological dilation.

Must be: `[1, rate_height, rate_width, 1]`.

padding: The type of padding algorithm to use.

Returns 3-D with shape `[filter_height, filter_width, depth]`.

func Dilation2DBackpropInput ¶ added in v1.1.0

func Dilation2DBackpropInput(scope *Scope, input tf.Output, filter tf.Output, out_backprop tf.Output, strides []int64, rates []int64, padding string) (in_backprop tf.Output)

Computes the gradient of morphological 2-D dilation with respect to the input.

Arguments:

input: 4-D with shape `[batch, in_height, in_width, depth]`.
filter: 3-D with shape `[filter_height, filter_width, depth]`.
out_backprop: 4-D with shape `[batch, out_height, out_width, depth]`.
strides: 1-D of length 4. The stride of the sliding window for each dimension of

the input tensor. Must be: `[1, stride_height, stride_width, 1]`.

rates: 1-D of length 4. The input stride for atrous morphological dilation.

Must be: `[1, rate_height, rate_width, 1]`.

padding: The type of padding algorithm to use.

Returns 4-D with shape `[batch, in_height, in_width, depth]`.

func DirectedInterleaveDataset ¶ added in v1.15.0

func DirectedInterleaveDataset(scope *Scope, selector_input_dataset tf.Output, data_input_datasets []tf.Output, output_types []tf.DataType, output_shapes []tf.Shape, optional ...DirectedInterleaveDatasetAttr) (handle tf.Output)

A substitute for `InterleaveDataset` on a fixed list of `N` datasets.

Arguments:

selector_input_dataset: A dataset of scalar `DT_INT64` elements that determines which of the

`N` data inputs should produce the next output element.

data_input_datasets: `N` datasets with the same type that will be interleaved according to

the values of `selector_input_dataset`.

func DisableCopyOnRead ¶

func DisableCopyOnRead(scope *Scope, resource tf.Output) (o *tf.Operation)

Turns off the copy-on-read mode.

Turns off the copy-on-read mode of a resource variable. If the variable is not in copy-on-read mode, this op has no effect.

Arguments:

resource: The resource handle of the resource variable.

Returns the created operation.

func Div ¶ added in v1.1.0

func Div(scope *Scope, x tf.Output, y tf.Output) (z tf.Output)

Returns x / y element-wise.

*NOTE*: `Div` supports broadcasting. More about broadcasting [here](http://docs.scipy.org/doc/numpy/user/basics.broadcasting.html)

func DivNoNan ¶ added in v1.12.0

func DivNoNan(scope *Scope, x tf.Output, y tf.Output) (z tf.Output)

Returns 0 if the denominator is zero.

*NOTE*: `DivNoNan` supports broadcasting. More about broadcasting [here](http://docs.scipy.org/doc/numpy/user/basics.broadcasting.html)

func DrawBoundingBoxes ¶ added in v1.1.0

func DrawBoundingBoxes(scope *Scope, images tf.Output, boxes tf.Output) (output tf.Output)

Draw bounding boxes on a batch of images.

Outputs a copy of `images` but draws on top of the pixels zero or more bounding boxes specified by the locations in `boxes`. The coordinates of the each bounding box in `boxes` are encoded as `[y_min, x_min, y_max, x_max]`. The bounding box coordinates are floats in `[0.0, 1.0]` relative to the width and height of the underlying image.

For example, if an image is 100 x 200 pixels (height x width) and the bounding box is `[0.1, 0.2, 0.5, 0.9]`, the upper-left and bottom-right coordinates of the bounding box will be `(40, 10)` to `(180, 50)` (in (x,y) coordinates).

Parts of the bounding box may fall outside the image.

Arguments:

images: 4-D with shape `[batch, height, width, depth]`. A batch of images.
boxes: 3-D with shape `[batch, num_bounding_boxes, 4]` containing bounding

boxes.

Returns 4-D with the same shape as `images`. The batch of input images with bounding boxes drawn on the images.

func DrawBoundingBoxesV2 ¶ added in v1.12.1

func DrawBoundingBoxesV2(scope *Scope, images tf.Output, boxes tf.Output, colors tf.Output) (output tf.Output)

Draw bounding boxes on a batch of images.

Outputs a copy of `images` but draws on top of the pixels zero or more bounding boxes specified by the locations in `boxes`. The coordinates of the each bounding box in `boxes` are encoded as `[y_min, x_min, y_max, x_max]`. The bounding box coordinates are floats in `[0.0, 1.0]` relative to the width and height of the underlying image.

For example, if an image is 100 x 200 pixels (height x width) and the bounding box is `[0.1, 0.2, 0.5, 0.9]`, the upper-left and bottom-right coordinates of the bounding box will be `(40, 10)` to `(100, 50)` (in (x,y) coordinates).

Parts of the bounding box may fall outside the image.

Arguments:

images: 4-D with shape `[batch, height, width, depth]`. A batch of images.
boxes: 3-D with shape `[batch, num_bounding_boxes, 4]` containing bounding

boxes.

colors: 2-D. A list of RGBA colors to cycle through for the boxes.

Returns 4-D with the same shape as `images`. The batch of input images with bounding boxes drawn on the images.

func DynamicEnqueueTPUEmbeddingArbitraryTensorBatch ¶

func DynamicEnqueueTPUEmbeddingArbitraryTensorBatch(scope *Scope, sample_indices_or_row_splits []tf.Output, embedding_indices []tf.Output, aggregation_weights []tf.Output, mode_override tf.Output, device_ordinal tf.Output, optional ...DynamicEnqueueTPUEmbeddingArbitraryTensorBatchAttr) (o *tf.Operation)

Eases the porting of code that uses tf.nn.embedding_lookup_sparse().

embedding_indices[i] and aggregation_weights[i] correspond to the ith feature.

The tensors at corresponding positions in the three input lists (sample_indices, embedding_indices and aggregation_weights) must have the same shape, i.e. rank 1 with dim_size() equal to the total number of lookups into the table described by the corresponding feature.

Arguments:

sample_indices_or_row_splits: A list of rank 2 Tensors specifying the training example to which the

corresponding embedding_indices and aggregation_weights values belong. If the size of its first dimension is 0, we assume each embedding_indices belongs to a different sample. Both int32 and int64 are allowed and will be converted to int32 internally.

Or a list of rank 1 Tensors specifying the row splits for splitting embedding_indices and aggregation_weights into rows. It corresponds to ids.row_splits in embedding_lookup(), when ids is a RaggedTensor. When enqueuing N-D ragged tensor, only the last dimension is allowed to be ragged. the row splits is 1-D dense tensor. When empty, we assume a dense tensor is passed to the op Both int32 and int64 are allowed and will be converted to int32 internally.

embedding_indices: A list of rank 1 Tensors, indices into the embedding

tables. Both int32 and int64 are allowed and will be converted to int32 internally.

aggregation_weights: A list of rank 1 Tensors containing per training

example aggregation weights. Both float32 and float64 are allowed and will be converted to float32 internally.

mode_override: A string input that overrides the mode specified in the

TPUEmbeddingConfiguration. Supported values are {'unspecified', 'inference', 'training', 'backward_pass_only'}. When set to 'unspecified', the mode set in TPUEmbeddingConfiguration is used, otherwise mode_override is used.

device_ordinal: The TPU device to use. Should be >= 0 and less than the number

of TPU cores in the task on which the node is placed.

Returns the created operation.

func DynamicPartition ¶ added in v1.1.0

func DynamicPartition(scope *Scope, data tf.Output, partitions tf.Output, num_partitions int64) (outputs []tf.Output)

Partitions `data` into `num_partitions` tensors using indices from `partitions`.

For each index tuple `js` of size `partitions.ndim`, the slice `data[js, ...]` becomes part of `outputs[partitions[js]]`. The slices with `partitions[js] = i` are placed in `outputs[i]` in lexicographic order of `js`, and the first dimension of `outputs[i]` is the number of entries in `partitions` equal to `i`. In detail,

```python

outputs[i].shape = [sum(partitions == i)] + data.shape[partitions.ndim:]

outputs[i] = pack([data[js, ...] for js if partitions[js] == i])

```

`data.shape` must start with `partitions.shape`.

For example:

```python

# Scalar partitions.
partitions = 1
num_partitions = 2
data = [10, 20]
outputs[0] = []  # Empty with shape [0, 2]
outputs[1] = [[10, 20]]

# Vector partitions.
partitions = [0, 0, 1, 1, 0]
num_partitions = 2
data = [10, 20, 30, 40, 50]
outputs[0] = [10, 20, 50]
outputs[1] = [30, 40]

```

See `dynamic_stitch` for an example on how to merge partitions back.

Raises:

`InvalidArgumentError` in following cases:
If partitions is not in range `[0, num_partiions)`
If `partitions.shape` does not match prefix of `data.shape` argument.

Arguments:

partitions: Any shape.  Indices in the range `[0, num_partitions)`.
num_partitions: The number of partitions to output.

func DynamicStitch ¶ added in v1.1.0

func DynamicStitch(scope *Scope, indices []tf.Output, data []tf.Output) (merged tf.Output)

Interleave the values from the `data` tensors into a single tensor.

Builds a merged tensor such that ¶

```python

merged[indices[m][i, ..., j], ...] = data[m][i, ..., j, ...]

```

For example, if each `indices[m]` is scalar or vector, we have

```python

# Scalar indices:
merged[indices[m], ...] = data[m][...]

# Vector indices:
merged[indices[m][i], ...] = data[m][i, ...]

```

Each `data[i].shape` must start with the corresponding `indices[i].shape`, and the rest of `data[i].shape` must be constant w.r.t. `i`. That is, we must have `data[i].shape = indices[i].shape + constant`. In terms of this `constant`, the output shape is

merged.shape = [max(indices) + 1] + constant

Values are merged in order, so if an index appears in both `indices[m][i]` and `indices[n][j]` for `(m,i) < (n,j)` the slice `data[n][j]` will appear in the merged result. If you do not need this guarantee, ParallelDynamicStitch might perform better on some devices.

For example:

```python

indices[0] = 6
indices[1] = [4, 1]
indices[2] = [[5, 2], [0, 3]]
data[0] = [61, 62]
data[1] = [[41, 42], [11, 12]]
data[2] = [[[51, 52], [21, 22]], [[1, 2], [31, 32]]]
merged = [[1, 2], [11, 12], [21, 22], [31, 32], [41, 42],
          [51, 52], [61, 62]]

```

This method can be used to merge partitions created by `dynamic_partition` as illustrated on the following example:

```python

# Apply function (increments x_i) on elements for which a certain condition
# apply (x_i != -1 in this example).
x=tf.constant([0.1, -1., 5.2, 4.3, -1., 7.4])
condition_mask=tf.not_equal(x,tf.constant(-1.))
partitioned_data = tf.dynamic_partition(
    x, tf.cast(condition_mask, tf.int32) , 2)
partitioned_data[1] = partitioned_data[1] + 1.0
condition_indices = tf.dynamic_partition(
    tf.range(tf.shape(x)[0]), tf.cast(condition_mask, tf.int32) , 2)
x = tf.dynamic_stitch(condition_indices, partitioned_data)
# Here x=[1.1, -1., 6.2, 5.3, -1, 8.4], the -1. values remain
# unchanged.

```

func EagerPyFunc ¶ added in v1.6.0

func EagerPyFunc(scope *Scope, input []tf.Output, token string, Tout []tf.DataType, optional ...EagerPyFuncAttr) (output []tf.Output)

Eagerly executes a python function to compute func(input)->output. The

semantics of the input, output, and attributes are the same as those for PyFunc.

func EditDistance ¶ added in v1.1.0

func EditDistance(scope *Scope, hypothesis_indices tf.Output, hypothesis_values tf.Output, hypothesis_shape tf.Output, truth_indices tf.Output, truth_values tf.Output, truth_shape tf.Output, optional ...EditDistanceAttr) (output tf.Output)

Computes the (possibly normalized) Levenshtein Edit Distance.

The inputs are variable-length sequences provided by SparseTensors

(hypothesis_indices, hypothesis_values, hypothesis_shape)

and

(truth_indices, truth_values, truth_shape).

The inputs are:

Arguments:

hypothesis_indices: The indices of the hypothesis list SparseTensor.

This is an N x R int64 matrix.

hypothesis_values: The values of the hypothesis list SparseTensor.

This is an N-length vector.

hypothesis_shape: The shape of the hypothesis list SparseTensor.

This is an R-length vector.

truth_indices: The indices of the truth list SparseTensor.

This is an M x R int64 matrix.

truth_values: The values of the truth list SparseTensor.

This is an M-length vector.

truth_shape: truth indices, vector.

Returns A dense float tensor with rank R - 1.

For the example input:

// hypothesis represents a 2x1 matrix with variable-length values:
//   (0,0) = ["a"]
//   (1,0) = ["b"]
hypothesis_indices = [[0, 0, 0],
                      [1, 0, 0]]
hypothesis_values = ["a", "b"]
hypothesis_shape = [2, 1, 1]

// truth represents a 2x2 matrix with variable-length values:
//   (0,0) = []
//   (0,1) = ["a"]
//   (1,0) = ["b", "c"]
//   (1,1) = ["a"]
truth_indices = [[0, 1, 0],
                 [1, 0, 0],
                 [1, 0, 1],
                 [1, 1, 0]]
truth_values = ["a", "b", "c", "a"]
truth_shape = [2, 2, 2]
normalize = true

The output will be:

// output is a 2x2 matrix with edit distances normalized by truth lengths.
output = [[inf, 1.0],  // (0,0): no truth, (0,1): no hypothesis
          [0.5, 1.0]]  // (1,0): addition, (1,1): no hypothesis

func Eig ¶

func Eig(scope *Scope, input tf.Output, Tout tf.DataType, optional ...EigAttr) (e tf.Output, v tf.Output)

Computes the eigen decomposition of one or more square matrices.

Computes the eigenvalues and (optionally) right eigenvectors of each inner matrix in `input` such that `input[..., :, :] = v[..., :, :] * diag(e[..., :])`. The eigenvalues are sorted in non-decreasing order.

```python # a is a tensor. # e is a tensor of eigenvalues. # v is a tensor of eigenvectors. e, v = eig(a) e = eig(a, compute_v=False) ```

Arguments:

input: `Tensor` input of shape `[N, N]`.

Returns:

e: Eigenvalues. Shape is `[N]`.
v: Eigenvectors. Shape is `[N, N]`.

func Einsum ¶ added in v1.15.0

func Einsum(scope *Scope, inputs []tf.Output, equation string) (output tf.Output)

Tensor contraction according to Einstein summation convention.

Implements generalized Tensor contraction and reduction. Each input Tensor must have a corresponding input subscript appearing in the comma-separated left-hand side of the equation. The right-hand side of the equation consists of the output subscript. The input subscripts and the output subscript should consist of zero or more named axis labels and at most one ellipsis (`...`).

The named axis labels may be any single character other than those having special meaning, namely `,.->`. The behavior of this Op is undefined if it receives an ill-formatted equation; since the validation is done at graph-building time, we omit format validation checks at runtime.

Note: This Op is *not* intended to be called by the user; instead users should call `tf.einsum` directly. It is a hidden Op used by `tf.einsum`.

Operations are applied to the input(s) according to the following rules:

(a) Generalized Diagonals: For input dimensions corresponding to axis labels
    appearing more than once in the same input subscript, we take the
    generalized (`k`-dimensional) diagonal.
    For example, in the equation `iii->i` with input shape `[3, 3, 3]`, the
    generalized diagonal would consist of `3` elements at indices `(0, 0, 0)`,
    `(1, 1, 1)` and `(2, 2, 2)` to create a Tensor of shape `[3]`.

(b) Reduction: Axes corresponding to labels appearing only in one input
    subscript but not in the output subscript are summed over prior to Tensor
    contraction.
    For example, in the equation `ab,bc->b`, the axis labels `a` and `c` are
    the reduction axis labels.

(c) Batch Dimensions: Axes corresponding to labels appearing in each of the
    input subscripts and also in the output subscript make up the batch
    dimensions in Tensor contraction. Unnamed axis labels corresponding to
    ellipsis (`...`) also correspond to batch dimensions.
    For example, for the equation denoting batch matrix multiplication,
    `bij,bjk->bik`, the axis label `b` corresponds to a batch dimension.

(d) Contraction: In case of binary einsum, axes corresponding to labels
    appearing in two different inputs (and not in the output) are contracted
    against each other.
    Considering the batch matrix multiplication equation again
    (`bij,bjk->bik`), the contracted axis label is `j`.

(e) Expand Diagonal: If the output subscripts contain repeated (explicit) axis
    labels, the opposite operation of (a) is applied. For example, in the
    equation `i->iii`, and input shape `[3]`, the output of shape `[3, 3, 3]`
    are all zeros, except for the (generalized) diagonal which is populated
    with values from the input.
    Note: This operation is not supported by `np.einsum` or `tf.einsum`; it is
    provided to enable computing the symbolic gradient of `tf.einsum`.

The output subscripts must contain only labels appearing in at least one of the input subscripts. Furthermore, all dimensions mapping to the same axis label must be equal.

Any of the input and output subscripts may contain at most a single ellipsis (`...`). These ellipsis are mapped against dimensions not corresponding to any named axis label. If two inputs contain ellipsis, then they are broadcasted according to standard NumPy broadcasting [rules](http://docs.scipy.org/doc/numpy/user/basics.broadcasting.html).

The broadcasted dimensions are placed in the corresponding location of the ellipsis in the output subscript. If the broadcasted dimensions are non-empty and the output subscripts do not contain ellipsis, then an InvalidArgument error is raised.

@compatibility(numpy) Similar to [`numpy.einsum`](https://docs.scipy.org/doc/numpy/reference/generated/numpy.einsum.html).

Comparison with `numpy.einsum`:

This Op only supports unary and binary forms of `numpy.einsum`.
This Op does not support implicit form. (i.e. equations without `->`).
This Op also supports repeated indices in the output subscript, which is not supported by `numpy.einsum`.

@end_compatibility

Arguments:

inputs: List of 1 or 2 Tensors.
equation: String describing the Einstein Summation operation; in the format of np.einsum.

Returns Output Tensor with shape depending upon `equation`.

func Elu ¶ added in v1.1.0

func Elu(scope *Scope, features tf.Output) (activations tf.Output)

Computes the exponential linear function.

The ELU function is defined as:

$ e ^ x - 1 $ if $ x < 0 $
$ x $ if $ x >= 0 $

Examples:

>>> tf.nn.elu(1.0) <tf.Tensor: shape=(), dtype=float32, numpy=1.0> >>> tf.nn.elu(0.0) <tf.Tensor: shape=(), dtype=float32, numpy=0.0> >>> tf.nn.elu(-1000.0) <tf.Tensor: shape=(), dtype=float32, numpy=-1.0>

See [Fast and Accurate Deep Network Learning by Exponential Linear Units (ELUs) ](http://arxiv.org/abs/1511.07289)

func EluGrad ¶ added in v1.1.0

func EluGrad(scope *Scope, gradients tf.Output, outputs tf.Output) (backprops tf.Output)

Computes gradients for the exponential linear (Elu) operation.

Arguments:

gradients: The backpropagated gradients to the corresponding Elu operation.
outputs: The outputs of the corresponding Elu operation.

Returns The gradients: `gradients * (outputs + 1)` if outputs < 0, `gradients` otherwise.

func Empty ¶ added in v1.9.0

func Empty(scope *Scope, shape tf.Output, dtype tf.DataType, optional ...EmptyAttr) (output tf.Output)

Creates a tensor with the given shape.

This operation creates a tensor of `shape` and `dtype`.

Arguments:

shape: 1-D. Represents the shape of the output tensor.

Returns A `Tensor` of type `T`.

func EmptyTensorList ¶ added in v1.9.0

func EmptyTensorList(scope *Scope, element_shape tf.Output, max_num_elements tf.Output, element_dtype tf.DataType) (handle tf.Output)

Creates and returns an empty tensor list.

All list elements must be tensors of dtype element_dtype and shape compatible with element_shape.

handle: an empty tensor list. element_dtype: the type of elements in the list. element_shape: a shape compatible with that of elements in the list.

func EmptyTensorMap ¶

func EmptyTensorMap(scope *Scope) (handle tf.Output)

Creates and returns an empty tensor map.

handle: an empty tensor map

func EncodeBase64 ¶ added in v1.1.0

func EncodeBase64(scope *Scope, input tf.Output, optional ...EncodeBase64Attr) (output tf.Output)

Encode strings into web-safe base64 format.

Refer to [this article](https://en.wikipedia.org/wiki/Base64) for more information on base64 format. Base64 strings may have padding with '=' at the end so that the encoded has length multiple of 4. See Padding section of the link above.

Web-safe means that the encoder uses - and _ instead of + and /.

Arguments:

input: Strings to be encoded.

Returns Input strings encoded in base64.

func EncodeJpeg ¶ added in v1.1.0

func EncodeJpeg(scope *Scope, image tf.Output, optional ...EncodeJpegAttr) (contents tf.Output)

JPEG-encode an image.

`image` is a 3-D uint8 Tensor of shape `[height, width, channels]`.

The attr `format` can be used to override the color format of the encoded output. Values can be:

`”`: Use a default format based on the number of channels in the image.
`grayscale`: Output a grayscale JPEG image. The `channels` dimension of `image` must be 1.
`rgb`: Output an RGB JPEG image. The `channels` dimension of `image` must be 3.

If `format` is not specified or is the empty string, a default format is picked in function of the number of channels in `image`:

* 1: Output a grayscale image. * 3: Output an RGB image.

Arguments:

image: 3-D with shape `[height, width, channels]`.

Returns 0-D. JPEG-encoded image.

func EncodeJpegVariableQuality ¶ added in v1.12.1

func EncodeJpegVariableQuality(scope *Scope, images tf.Output, quality tf.Output) (contents tf.Output)

JPEG encode input image with provided compression quality.

`image` is a 3-D uint8 Tensor of shape `[height, width, channels]`. `quality` is an int32 jpeg compression quality value between 0 and 100.

Arguments:

images: Images to adjust.  At least 3-D.
quality: An int quality to encode to.

Returns 0-D. JPEG-encoded image.

func EncodePng ¶ added in v1.1.0

func EncodePng(scope *Scope, image tf.Output, optional ...EncodePngAttr) (contents tf.Output)

PNG-encode an image.

`image` is a 3-D uint8 or uint16 Tensor of shape `[height, width, channels]` where `channels` is:

* 1: for grayscale. * 2: for grayscale + alpha. * 3: for RGB. * 4: for RGBA.

The ZLIB compression level, `compression`, can be -1 for the PNG-encoder default or a value from 0 to 9. 9 is the highest compression level, generating the smallest output, but is slower.

Arguments:

image: 3-D with shape `[height, width, channels]`.

Returns 0-D. PNG-encoded image.

func EncodeProto ¶ added in v1.9.0

func EncodeProto(scope *Scope, sizes tf.Output, values []tf.Output, field_names []string, message_type string, optional ...EncodeProtoAttr) (bytes tf.Output)

The op serializes protobuf messages provided in the input tensors.

The types of the tensors in `values` must match the schema for the fields specified in `field_names`. All the tensors in `values` must have a common shape prefix, *batch_shape*.

The `sizes` tensor specifies repeat counts for each field. The repeat count (last dimension) of a each tensor in `values` must be greater than or equal to corresponding repeat count in `sizes`.