symbolic_pymc.tensorflow package

Submodules

symbolic_pymc.tensorflow.dispatch module

symbolic_pymc.tensorflow.graph module

symbolic_pymc.tensorflow.graph.normalize_tf_graph(graph_output, new_graph=True, verbose=False)

Use grappler to normalize a graph.

graph_output: Tensor

A tensor we want to consider as “output” of a FuncGraph.

The simplified graph.

symbolic_pymc.tensorflow.meta module

class symbolic_pymc.tensorflow.meta.MetaOpDefLibrary

Bases: object

A singleton-like object that holds correspondences between TF Python API functions and the `OpDef`s they construct.

It provides a map of OpDef names (lower-cased) to the Python API functions in tensorflow.raw_ops, as well as inspect.Signature objects for said functions so that default values and lists of arguments (keywords included) can be more easily used.

classmethod get_op_info(opdef)

Return the TF Python API function signature for a given OpDef.

opdef: str or OpDef object (meta or base)

lower_op_name_to_raw = {'abort': 'Abort', 'abs': 'Abs', 'accumulatenv2': 'AccumulateNV2', 'accumulatorapplygradient': 'AccumulatorApplyGradient', 'accumulatornumaccumulated': 'AccumulatorNumAccumulated', 'accumulatorsetglobalstep': 'AccumulatorSetGlobalStep', 'accumulatortakegradient': 'AccumulatorTakeGradient', 'acos': 'Acos', 'acosh': 'Acosh', 'add': 'Add', 'addmanysparsetotensorsmap': 'AddManySparseToTensorsMap', 'addn': 'AddN', 'addsparsetotensorsmap': 'AddSparseToTensorsMap', 'addv2': 'AddV2', 'adjustcontrast': 'AdjustContrast', 'adjustcontrastv2': 'AdjustContrastv2', 'adjusthue': 'AdjustHue', 'adjustsaturation': 'AdjustSaturation', 'all': 'All', 'allcandidatesampler': 'AllCandidateSampler', 'alltoall': 'AllToAll', 'angle': 'Angle', 'anonymousiterator': 'AnonymousIterator', 'anonymousiteratorv2': 'AnonymousIteratorV2', 'anonymousmemorycache': 'AnonymousMemoryCache', 'anonymousmultideviceiterator': 'AnonymousMultiDeviceIterator', 'anonymousrandomseedgenerator': 'AnonymousRandomSeedGenerator', 'any': 'Any', 'applyadadelta': 'ApplyAdadelta', 'applyadagrad': 'ApplyAdagrad', 'applyadagradda': 'ApplyAdagradDA', 'applyadagradv2': 'ApplyAdagradV2', 'applyadam': 'ApplyAdam', 'applyadamax': 'ApplyAdaMax', 'applyaddsign': 'ApplyAddSign', 'applycenteredrmsprop': 'ApplyCenteredRMSProp', 'applyftrl': 'ApplyFtrl', 'applyftrlv2': 'ApplyFtrlV2', 'applygradientdescent': 'ApplyGradientDescent', 'applymomentum': 'ApplyMomentum', 'applypowersign': 'ApplyPowerSign', 'applyproximaladagrad': 'ApplyProximalAdagrad', 'applyproximalgradientdescent': 'ApplyProximalGradientDescent', 'applyrmsprop': 'ApplyRMSProp', 'approximateequal': 'ApproximateEqual', 'argmax': 'ArgMax', 'argmin': 'ArgMin', 'asin': 'Asin', 'asinh': 'Asinh', 'assert': 'Assert', 'assertnextdataset': 'AssertNextDataset', 'assign': 'Assign', 'assignadd': 'AssignAdd', 'assignaddvariableop': 'AssignAddVariableOp', 'assignsub': 'AssignSub', 'assignsubvariableop': 'AssignSubVariableOp', 'assignvariableop': 'AssignVariableOp', 'asstring': 'AsString', 'atan': 'Atan', 'atan2': 'Atan2', 'atanh': 'Atanh', 'audiospectrogram': 'AudioSpectrogram', 'audiosummary': 'AudioSummary', 'audiosummaryv2': 'AudioSummaryV2', 'autosharddataset': 'AutoShardDataset', 'avgpool': 'AvgPool', 'avgpool3d': 'AvgPool3D', 'avgpool3dgrad': 'AvgPool3DGrad', 'avgpoolgrad': 'AvgPoolGrad', 'barrier': 'Barrier', 'barrierclose': 'BarrierClose', 'barrierincompletesize': 'BarrierIncompleteSize', 'barrierinsertmany': 'BarrierInsertMany', 'barrierreadysize': 'BarrierReadySize', 'barriertakemany': 'BarrierTakeMany', 'batch': 'Batch', 'batchcholesky': 'BatchCholesky', 'batchcholeskygrad': 'BatchCholeskyGrad', 'batchdataset': 'BatchDataset', 'batchdatasetv2': 'BatchDatasetV2', 'batchfft': 'BatchFFT', 'batchfft2d': 'BatchFFT2D', 'batchfft3d': 'BatchFFT3D', 'batchfunction': 'BatchFunction', 'batchifft': 'BatchIFFT', 'batchifft2d': 'BatchIFFT2D', 'batchifft3d': 'BatchIFFT3D', 'batchmatmul': 'BatchMatMul', 'batchmatmulv2': 'BatchMatMulV2', 'batchmatrixbandpart': 'BatchMatrixBandPart', 'batchmatrixdeterminant': 'BatchMatrixDeterminant', 'batchmatrixdiag': 'BatchMatrixDiag', 'batchmatrixdiagpart': 'BatchMatrixDiagPart', 'batchmatrixinverse': 'BatchMatrixInverse', 'batchmatrixsetdiag': 'BatchMatrixSetDiag', 'batchmatrixsolve': 'BatchMatrixSolve', 'batchmatrixsolvels': 'BatchMatrixSolveLs', 'batchmatrixtriangularsolve': 'BatchMatrixTriangularSolve', 'batchnormwithglobalnormalization': 'BatchNormWithGlobalNormalization', 'batchnormwithglobalnormalizationgrad': 'BatchNormWithGlobalNormalizationGrad', 'batchselfadjointeig': 'BatchSelfAdjointEig', 'batchselfadjointeigv2': 'BatchSelfAdjointEigV2', 'batchsvd': 'BatchSvd', 'batchtospace': 'BatchToSpace', 'batchtospacend': 'BatchToSpaceND', 'besseli0e': 'BesselI0e', 'besseli1e': 'BesselI1e', 'betainc': 'Betainc', 'biasadd': 'BiasAdd', 'biasaddgrad': 'BiasAddGrad', 'biasaddv1': 'BiasAddV1', 'bincount': 'Bincount', 'bitcast': 'Bitcast', 'bitwiseand': 'BitwiseAnd', 'bitwiseor': 'BitwiseOr', 'bitwisexor': 'BitwiseXor', 'blocklstm': 'BlockLSTM', 'blocklstmgrad': 'BlockLSTMGrad', 'blocklstmgradv2': 'BlockLSTMGradV2', 'blocklstmv2': 'BlockLSTMV2', 'boostedtreesaggregatestats': 'BoostedTreesAggregateStats', 'boostedtreesbucketize': 'BoostedTreesBucketize', 'boostedtreescalculatebestfeaturesplit': 'BoostedTreesCalculateBestFeatureSplit', 'boostedtreescalculatebestfeaturesplitv2': 'BoostedTreesCalculateBestFeatureSplitV2', 'boostedtreescalculatebestgainsperfeature': 'BoostedTreesCalculateBestGainsPerFeature', 'boostedtreescenterbias': 'BoostedTreesCenterBias', 'boostedtreescreateensemble': 'BoostedTreesCreateEnsemble', 'boostedtreescreatequantilestreamresource': 'BoostedTreesCreateQuantileStreamResource', 'boostedtreesdeserializeensemble': 'BoostedTreesDeserializeEnsemble', 'boostedtreesensembleresourcehandleop': 'BoostedTreesEnsembleResourceHandleOp', 'boostedtreesexampledebugoutputs': 'BoostedTreesExampleDebugOutputs', 'boostedtreesflushquantilesummaries': 'BoostedTreesFlushQuantileSummaries', 'boostedtreesgetensemblestates': 'BoostedTreesGetEnsembleStates', 'boostedtreesmakequantilesummaries': 'BoostedTreesMakeQuantileSummaries', 'boostedtreesmakestatssummary': 'BoostedTreesMakeStatsSummary', 'boostedtreespredict': 'BoostedTreesPredict', 'boostedtreesquantilestreamresourceaddsummaries': 'BoostedTreesQuantileStreamResourceAddSummaries', 'boostedtreesquantilestreamresourcedeserialize': 'BoostedTreesQuantileStreamResourceDeserialize', 'boostedtreesquantilestreamresourceflush': 'BoostedTreesQuantileStreamResourceFlush', 'boostedtreesquantilestreamresourcegetbucketboundaries': 'BoostedTreesQuantileStreamResourceGetBucketBoundaries', 'boostedtreesquantilestreamresourcehandleop': 'BoostedTreesQuantileStreamResourceHandleOp', 'boostedtreesserializeensemble': 'BoostedTreesSerializeEnsemble', 'boostedtreessparseaggregatestats': 'BoostedTreesSparseAggregateStats', 'boostedtreessparsecalculatebestfeaturesplit': 'BoostedTreesSparseCalculateBestFeatureSplit', 'boostedtreestrainingpredict': 'BoostedTreesTrainingPredict', 'boostedtreesupdateensemble': 'BoostedTreesUpdateEnsemble', 'boostedtreesupdateensemblev2': 'BoostedTreesUpdateEnsembleV2', 'broadcastargs': 'BroadcastArgs', 'broadcastgradientargs': 'BroadcastGradientArgs', 'broadcastto': 'BroadcastTo', 'bucketize': 'Bucketize', 'bytesproducedstatsdataset': 'BytesProducedStatsDataset', 'cachedataset': 'CacheDataset', 'cachedatasetv2': 'CacheDatasetV2', 'case': 'Case', 'cast': 'Cast', 'ceil': 'Ceil', 'checknumerics': 'CheckNumerics', 'checknumericsv2': 'CheckNumericsV2', 'cholesky': 'Cholesky', 'choleskygrad': 'CholeskyGrad', 'choosefastestbranchdataset': 'ChooseFastestBranchDataset', 'choosefastestdataset': 'ChooseFastestDataset', 'clipbyvalue': 'ClipByValue', 'closesummarywriter': 'CloseSummaryWriter', 'collectivebcastrecv': 'CollectiveBcastRecv', 'collectivebcastsend': 'CollectiveBcastSend', 'collectivegather': 'CollectiveGather', 'collectivepermute': 'CollectivePermute', 'collectivereduce': 'CollectiveReduce', 'combinednonmaxsuppression': 'CombinedNonMaxSuppression', 'compareandbitpack': 'CompareAndBitpack', 'complex': 'Complex', 'complexabs': 'ComplexAbs', 'computeaccidentalhits': 'ComputeAccidentalHits', 'concat': 'Concat', 'concatenatedataset': 'ConcatenateDataset', 'concatoffset': 'ConcatOffset', 'concatv2': 'ConcatV2', 'conditionalaccumulator': 'ConditionalAccumulator', 'configuredistributedtpu': 'ConfigureDistributedTPU', 'configuretpuembedding': 'ConfigureTPUEmbedding', 'conj': 'Conj', 'conjugatetranspose': 'ConjugateTranspose', 'const': 'Const', 'consumemutexlock': 'ConsumeMutexLock', 'controltrigger': 'ControlTrigger', 'conv2d': 'Conv2D', 'conv2dbackpropfilter': 'Conv2DBackpropFilter', 'conv2dbackpropinput': 'Conv2DBackpropInput', 'conv3d': 'Conv3D', 'conv3dbackpropfilter': 'Conv3DBackpropFilter', 'conv3dbackpropfilterv2': 'Conv3DBackpropFilterV2', 'conv3dbackpropinput': 'Conv3DBackpropInput', 'conv3dbackpropinputv2': 'Conv3DBackpropInputV2', 'copy': 'Copy', 'copyhost': 'CopyHost', 'cos': 'Cos', 'cosh': 'Cosh', 'countupto': 'CountUpTo', 'createsummarydbwriter': 'CreateSummaryDbWriter', 'createsummaryfilewriter': 'CreateSummaryFileWriter', 'cropandresize': 'CropAndResize', 'cropandresizegradboxes': 'CropAndResizeGradBoxes', 'cropandresizegradimage': 'CropAndResizeGradImage', 'cross': 'Cross', 'crossreplicasum': 'CrossReplicaSum', 'csrsparsematrixcomponents': 'CSRSparseMatrixComponents', 'csrsparsematrixtodense': 'CSRSparseMatrixToDense', 'csrsparsematrixtosparsetensor': 'CSRSparseMatrixToSparseTensor', 'csvdataset': 'CSVDataset', 'ctcbeamsearchdecoder': 'CTCBeamSearchDecoder', 'ctcgreedydecoder': 'CTCGreedyDecoder', 'ctcloss': 'CTCLoss', 'ctclossv2': 'CTCLossV2', 'cudnnrnn': 'CudnnRNN', 'cudnnrnnbackprop': 'CudnnRNNBackprop', 'cudnnrnnbackpropv2': 'CudnnRNNBackpropV2', 'cudnnrnnbackpropv3': 'CudnnRNNBackpropV3', 'cudnnrnncanonicaltoparams': 'CudnnRNNCanonicalToParams', 'cudnnrnncanonicaltoparamsv2': 'CudnnRNNCanonicalToParamsV2', 'cudnnrnnparamssize': 'CudnnRNNParamsSize', 'cudnnrnnparamstocanonical': 'CudnnRNNParamsToCanonical', 'cudnnrnnparamstocanonicalv2': 'CudnnRNNParamsToCanonicalV2', 'cudnnrnnv2': 'CudnnRNNV2', 'cudnnrnnv3': 'CudnnRNNV3', 'cumprod': 'Cumprod', 'cumsum': 'Cumsum', 'cumulativelogsumexp': 'CumulativeLogsumexp', 'dataformatdimmap': 'DataFormatDimMap', 'dataformatvecpermute': 'DataFormatVecPermute', 'datasetcardinality': 'DatasetCardinality', 'datasetfromgraph': 'DatasetFromGraph', 'datasettograph': 'DatasetToGraph', 'datasettographv2': 'DatasetToGraphV2', 'datasettosingleelement': 'DatasetToSingleElement', 'datasettotfrecord': 'DatasetToTFRecord', 'dawsn': 'Dawsn', 'debuggradientidentity': 'DebugGradientIdentity', 'debuggradientrefidentity': 'DebugGradientRefIdentity', 'debugidentity': 'DebugIdentity', 'debugidentityv2': 'DebugIdentityV2', 'debugnancount': 'DebugNanCount', 'debugnumericsummary': 'DebugNumericSummary', 'debugnumericsummaryv2': 'DebugNumericSummaryV2', 'decodeandcropjpeg': 'DecodeAndCropJpeg', 'decodebase64': 'DecodeBase64', 'decodebmp': 'DecodeBmp', 'decodecompressed': 'DecodeCompressed', 'decodecsv': 'DecodeCSV', 'decodegif': 'DecodeGif', 'decodejpeg': 'DecodeJpeg', 'decodejsonexample': 'DecodeJSONExample', 'decodepaddedraw': 'DecodePaddedRaw', 'decodepng': 'DecodePng', 'decodeprotov2': 'DecodeProtoV2', 'decoderaw': 'DecodeRaw', 'decodewav': 'DecodeWav', 'deepcopy': 'DeepCopy', 'deleteiterator': 'DeleteIterator', 'deletememorycache': 'DeleteMemoryCache', 'deletemultideviceiterator': 'DeleteMultiDeviceIterator', 'deleterandomseedgenerator': 'DeleteRandomSeedGenerator', 'deletesessiontensor': 'DeleteSessionTensor', 'densetocsrsparsematrix': 'DenseToCSRSparseMatrix', 'densetodensesetoperation': 'DenseToDenseSetOperation', 'densetosparsebatchdataset': 'DenseToSparseBatchDataset', 'densetosparsesetoperation': 'DenseToSparseSetOperation', 'depthtospace': 'DepthToSpace', 'depthwiseconv2dnative': 'DepthwiseConv2dNative', 'depthwiseconv2dnativebackpropfilter': 'DepthwiseConv2dNativeBackpropFilter', 'depthwiseconv2dnativebackpropinput': 'DepthwiseConv2dNativeBackpropInput', 'dequantize': 'Dequantize', 'deserializeiterator': 'DeserializeIterator', 'deserializemanysparse': 'DeserializeManySparse', 'deserializesparse': 'DeserializeSparse', 'destroyresourceop': 'DestroyResourceOp', 'destroytemporaryvariable': 'DestroyTemporaryVariable', 'diag': 'Diag', 'diagpart': 'DiagPart', 'digamma': 'Digamma', 'dilation2d': 'Dilation2D', 'dilation2dbackpropfilter': 'Dilation2DBackpropFilter', 'dilation2dbackpropinput': 'Dilation2DBackpropInput', 'directedinterleavedataset': 'DirectedInterleaveDataset', 'div': 'Div', 'divnonan': 'DivNoNan', 'drawboundingboxes': 'DrawBoundingBoxes', 'drawboundingboxesv2': 'DrawBoundingBoxesV2', 'dynamicpartition': 'DynamicPartition', 'dynamicstitch': 'DynamicStitch', 'eagerpyfunc': 'EagerPyFunc', 'editdistance': 'EditDistance', 'eig': 'Eig', 'einsum': 'Einsum', 'elu': 'Elu', 'elugrad': 'EluGrad', 'empty': 'Empty', 'emptytensorlist': 'EmptyTensorList', 'encodebase64': 'EncodeBase64', 'encodejpeg': 'EncodeJpeg', 'encodejpegvariablequality': 'EncodeJpegVariableQuality', 'encodepng': 'EncodePng', 'encodeproto': 'EncodeProto', 'encodewav': 'EncodeWav', 'enqueuetpuembeddingintegerbatch': 'EnqueueTPUEmbeddingIntegerBatch', 'enqueuetpuembeddingsparsebatch': 'EnqueueTPUEmbeddingSparseBatch', 'enqueuetpuembeddingsparsetensorbatch': 'EnqueueTPUEmbeddingSparseTensorBatch', 'ensureshape': 'EnsureShape', 'enter': 'Enter', 'equal': 'Equal', 'erf': 'Erf', 'erfc': 'Erfc', 'erfinv': 'Erfinv', 'euclideannorm': 'EuclideanNorm', 'exit': 'Exit', 'exp': 'Exp', 'expanddims': 'ExpandDims', 'experimentalassertnextdataset': 'ExperimentalAssertNextDataset', 'experimentalautosharddataset': 'ExperimentalAutoShardDataset', 'experimentalbytesproducedstatsdataset': 'ExperimentalBytesProducedStatsDataset', 'experimentalchoosefastestdataset': 'ExperimentalChooseFastestDataset', 'experimentalcsvdataset': 'ExperimentalCSVDataset', 'experimentaldatasetcardinality': 'ExperimentalDatasetCardinality', 'experimentaldatasettotfrecord': 'ExperimentalDatasetToTFRecord', 'experimentaldensetosparsebatchdataset': 'ExperimentalDenseToSparseBatchDataset', 'experimentaldirectedinterleavedataset': 'ExperimentalDirectedInterleaveDataset', 'experimentalgroupbyreducerdataset': 'ExperimentalGroupByReducerDataset', 'experimentalgroupbywindowdataset': 'ExperimentalGroupByWindowDataset', 'experimentalignoreerrorsdataset': 'ExperimentalIgnoreErrorsDataset', 'experimentaliteratorgetdevice': 'ExperimentalIteratorGetDevice', 'experimentallatencystatsdataset': 'ExperimentalLatencyStatsDataset', 'experimentallmdbdataset': 'ExperimentalLMDBDataset', 'experimentalmapandbatchdataset': 'ExperimentalMapAndBatchDataset', 'experimentalmapdataset': 'ExperimentalMapDataset', 'experimentalmatchingfilesdataset': 'ExperimentalMatchingFilesDataset', 'experimentalmaxintraopparallelismdataset': 'ExperimentalMaxIntraOpParallelismDataset', 'experimentalnonserializabledataset': 'ExperimentalNonSerializableDataset', 'experimentalparallelinterleavedataset': 'ExperimentalParallelInterleaveDataset', 'experimentalparseexampledataset': 'ExperimentalParseExampleDataset', 'experimentalprivatethreadpooldataset': 'ExperimentalPrivateThreadPoolDataset', 'experimentalrandomdataset': 'ExperimentalRandomDataset', 'experimentalrebatchdataset': 'ExperimentalRebatchDataset', 'experimentalscandataset': 'ExperimentalScanDataset', 'experimentalsetstatsaggregatordataset': 'ExperimentalSetStatsAggregatorDataset', 'experimentalsleepdataset': 'ExperimentalSleepDataset', 'experimentalslidingwindowdataset': 'ExperimentalSlidingWindowDataset', 'experimentalsqldataset': 'ExperimentalSqlDataset', 'experimentalstatsaggregatorhandle': 'ExperimentalStatsAggregatorHandle', 'experimentalstatsaggregatorsummary': 'ExperimentalStatsAggregatorSummary', 'experimentaltakewhiledataset': 'ExperimentalTakeWhileDataset', 'experimentalthreadpooldataset': 'ExperimentalThreadPoolDataset', 'experimentalthreadpoolhandle': 'ExperimentalThreadPoolHandle', 'experimentalunbatchdataset': 'ExperimentalUnbatchDataset', 'experimentaluniquedataset': 'ExperimentalUniqueDataset', 'expint': 'Expint', 'expm1': 'Expm1', 'extractglimpse': 'ExtractGlimpse', 'extractimagepatches': 'ExtractImagePatches', 'extractjpegshape': 'ExtractJpegShape', 'extractvolumepatches': 'ExtractVolumePatches', 'fact': 'Fact', 'fakeparam': 'FakeParam', 'fakequantwithminmaxargs': 'FakeQuantWithMinMaxArgs', 'fakequantwithminmaxargsgradient': 'FakeQuantWithMinMaxArgsGradient', 'fakequantwithminmaxvars': 'FakeQuantWithMinMaxVars', 'fakequantwithminmaxvarsgradient': 'FakeQuantWithMinMaxVarsGradient', 'fakequantwithminmaxvarsperchannel': 'FakeQuantWithMinMaxVarsPerChannel', 'fakequantwithminmaxvarsperchannelgradient': 'FakeQuantWithMinMaxVarsPerChannelGradient', 'fakequeue': 'FakeQueue', 'fft': 'FFT', 'fft2d': 'FFT2D', 'fft3d': 'FFT3D', 'fifoqueue': 'FIFOQueue', 'fifoqueuev2': 'FIFOQueueV2', 'fill': 'Fill', 'filterbylastcomponentdataset': 'FilterByLastComponentDataset', 'filterdataset': 'FilterDataset', 'fingerprint': 'Fingerprint', 'fixedlengthrecorddataset': 'FixedLengthRecordDataset', 'fixedlengthrecorddatasetv2': 'FixedLengthRecordDatasetV2', 'fixedlengthrecordreader': 'FixedLengthRecordReader', 'fixedlengthrecordreaderv2': 'FixedLengthRecordReaderV2', 'fixedunigramcandidatesampler': 'FixedUnigramCandidateSampler', 'flatmapdataset': 'FlatMapDataset', 'floor': 'Floor', 'floordiv': 'FloorDiv', 'floormod': 'FloorMod', 'flushsummarywriter': 'FlushSummaryWriter', 'for': 'For', 'fractionalavgpool': 'FractionalAvgPool', 'fractionalavgpoolgrad': 'FractionalAvgPoolGrad', 'fractionalmaxpool': 'FractionalMaxPool', 'fractionalmaxpoolgrad': 'FractionalMaxPoolGrad', 'fresnelcos': 'FresnelCos', 'fresnelsin': 'FresnelSin', 'fusedbatchnorm': 'FusedBatchNorm', 'fusedbatchnormgrad': 'FusedBatchNormGrad', 'fusedbatchnormgradv2': 'FusedBatchNormGradV2', 'fusedbatchnormgradv3': 'FusedBatchNormGradV3', 'fusedbatchnormv2': 'FusedBatchNormV2', 'fusedbatchnormv3': 'FusedBatchNormV3', 'fusedpadconv2d': 'FusedPadConv2D', 'fusedresizeandpadconv2d': 'FusedResizeAndPadConv2D', 'gather': 'Gather', 'gathernd': 'GatherNd', 'gatherv2': 'GatherV2', 'generateboundingboxproposals': 'GenerateBoundingBoxProposals', 'generatevocabremapping': 'GenerateVocabRemapping', 'generatordataset': 'GeneratorDataset', 'getsessionhandle': 'GetSessionHandle', 'getsessionhandlev2': 'GetSessionHandleV2', 'getsessiontensor': 'GetSessionTensor', 'greater': 'Greater', 'greaterequal': 'GreaterEqual', 'groupbyreducerdataset': 'GroupByReducerDataset', 'groupbywindowdataset': 'GroupByWindowDataset', 'grublockcell': 'GRUBlockCell', 'grublockcellgrad': 'GRUBlockCellGrad', 'guaranteeconst': 'GuaranteeConst', 'hashtable': 'HashTable', 'hashtablev2': 'HashTableV2', 'histogramfixedwidth': 'HistogramFixedWidth', 'histogramsummary': 'HistogramSummary', 'hsvtorgb': 'HSVToRGB', 'identity': 'Identity', 'identityn': 'IdentityN', 'identityreader': 'IdentityReader', 'identityreaderv2': 'IdentityReaderV2', 'if': 'If', 'ifft': 'IFFT', 'ifft2d': 'IFFT2D', 'ifft3d': 'IFFT3D', 'igamma': 'Igamma', 'igammac': 'Igammac', 'igammagrada': 'IgammaGradA', 'ignoreerrorsdataset': 'IgnoreErrorsDataset', 'imag': 'Imag', 'imageprojectivetransformv2': 'ImageProjectiveTransformV2', 'imagesummary': 'ImageSummary', 'immutableconst': 'ImmutableConst', 'importevent': 'ImportEvent', 'infeeddequeue': 'InfeedDequeue', 'infeeddequeuetuple': 'InfeedDequeueTuple', 'infeedenqueue': 'InfeedEnqueue', 'infeedenqueueprelinearizedbuffer': 'InfeedEnqueuePrelinearizedBuffer', 'infeedenqueuetuple': 'InfeedEnqueueTuple', 'initializetable': 'InitializeTable', 'initializetablefromtextfile': 'InitializeTableFromTextFile', 'initializetablefromtextfilev2': 'InitializeTableFromTextFileV2', 'initializetablev2': 'InitializeTableV2', 'inplaceadd': 'InplaceAdd', 'inplacesub': 'InplaceSub', 'inplaceupdate': 'InplaceUpdate', 'interleavedataset': 'InterleaveDataset', 'intopk': 'InTopK', 'intopkv2': 'InTopKV2', 'inv': 'Inv', 'invert': 'Invert', 'invertpermutation': 'InvertPermutation', 'invgrad': 'InvGrad', 'irfft': 'IRFFT', 'irfft2d': 'IRFFT2D', 'irfft3d': 'IRFFT3D', 'isboostedtreesensembleinitialized': 'IsBoostedTreesEnsembleInitialized', 'isboostedtreesquantilestreamresourceinitialized': 'IsBoostedTreesQuantileStreamResourceInitialized', 'isfinite': 'IsFinite', 'isinf': 'IsInf', 'isnan': 'IsNan', 'isvariableinitialized': 'IsVariableInitialized', 'iterator': 'Iterator', 'iteratorfromstringhandle': 'IteratorFromStringHandle', 'iteratorfromstringhandlev2': 'IteratorFromStringHandleV2', 'iteratorgetdevice': 'IteratorGetDevice', 'iteratorgetnext': 'IteratorGetNext', 'iteratorgetnextasoptional': 'IteratorGetNextAsOptional', 'iteratorgetnextsync': 'IteratorGetNextSync', 'iteratortostringhandle': 'IteratorToStringHandle', 'iteratorv2': 'IteratorV2', 'l2loss': 'L2Loss', 'latencystatsdataset': 'LatencyStatsDataset', 'leakyrelu': 'LeakyRelu', 'leakyrelugrad': 'LeakyReluGrad', 'learnedunigramcandidatesampler': 'LearnedUnigramCandidateSampler', 'leftshift': 'LeftShift', 'less': 'Less', 'lessequal': 'LessEqual', 'lgamma': 'Lgamma', 'linspace': 'LinSpace', 'listdiff': 'ListDiff', 'lmdbdataset': 'LMDBDataset', 'lmdbreader': 'LMDBReader', 'loadandremapmatrix': 'LoadAndRemapMatrix', 'loadtpuembeddingadadeltaparameters': 'LoadTPUEmbeddingAdadeltaParameters', 'loadtpuembeddingadadeltaparametersgradaccumdebug': 'LoadTPUEmbeddingAdadeltaParametersGradAccumDebug', 'loadtpuembeddingadagradparameters': 'LoadTPUEmbeddingAdagradParameters', 'loadtpuembeddingadagradparametersgradaccumdebug': 'LoadTPUEmbeddingAdagradParametersGradAccumDebug', 'loadtpuembeddingadamparameters': 'LoadTPUEmbeddingADAMParameters', 'loadtpuembeddingadamparametersgradaccumdebug': 'LoadTPUEmbeddingADAMParametersGradAccumDebug', 'loadtpuembeddingcenteredrmspropparameters': 'LoadTPUEmbeddingCenteredRMSPropParameters', 'loadtpuembeddingftrlparameters': 'LoadTPUEmbeddingFTRLParameters', 'loadtpuembeddingftrlparametersgradaccumdebug': 'LoadTPUEmbeddingFTRLParametersGradAccumDebug', 'loadtpuembeddingmdladagradlightparameters': 'LoadTPUEmbeddingMDLAdagradLightParameters', 'loadtpuembeddingmomentumparameters': 'LoadTPUEmbeddingMomentumParameters', 'loadtpuembeddingmomentumparametersgradaccumdebug': 'LoadTPUEmbeddingMomentumParametersGradAccumDebug', 'loadtpuembeddingproximaladagradparameters': 'LoadTPUEmbeddingProximalAdagradParameters', 'loadtpuembeddingproximaladagradparametersgradaccumdebug': 'LoadTPUEmbeddingProximalAdagradParametersGradAccumDebug', 'loadtpuembeddingrmspropparameters': 'LoadTPUEmbeddingRMSPropParameters', 'loadtpuembeddingrmspropparametersgradaccumdebug': 'LoadTPUEmbeddingRMSPropParametersGradAccumDebug', 'loadtpuembeddingstochasticgradientdescentparameters': 'LoadTPUEmbeddingStochasticGradientDescentParameters', 'log': 'Log', 'log1p': 'Log1p', 'logicaland': 'LogicalAnd', 'logicalnot': 'LogicalNot', 'logicalor': 'LogicalOr', 'logmatrixdeterminant': 'LogMatrixDeterminant', 'logsoftmax': 'LogSoftmax', 'loguniformcandidatesampler': 'LogUniformCandidateSampler', 'lookuptableexport': 'LookupTableExport', 'lookuptableexportv2': 'LookupTableExportV2', 'lookuptablefind': 'LookupTableFind', 'lookuptablefindv2': 'LookupTableFindV2', 'lookuptableimport': 'LookupTableImport', 'lookuptableimportv2': 'LookupTableImportV2', 'lookuptableinsert': 'LookupTableInsert', 'lookuptableinsertv2': 'LookupTableInsertV2', 'lookuptableremovev2': 'LookupTableRemoveV2', 'lookuptablesize': 'LookupTableSize', 'lookuptablesizev2': 'LookupTableSizeV2', 'loopcond': 'LoopCond', 'lowerbound': 'LowerBound', 'lrn': 'LRN', 'lrngrad': 'LRNGrad', 'lstmblockcell': 'LSTMBlockCell', 'lstmblockcellgrad': 'LSTMBlockCellGrad', 'lu': 'Lu', 'makeiterator': 'MakeIterator', 'mapandbatchdataset': 'MapAndBatchDataset', 'mapclear': 'MapClear', 'mapdataset': 'MapDataset', 'mapdefun': 'MapDefun', 'mapincompletesize': 'MapIncompleteSize', 'mappeek': 'MapPeek', 'mapsize': 'MapSize', 'mapstage': 'MapStage', 'mapunstage': 'MapUnstage', 'mapunstagenokey': 'MapUnstageNoKey', 'matchingfiles': 'MatchingFiles', 'matchingfilesdataset': 'MatchingFilesDataset', 'matmul': 'MatMul', 'matrixbandpart': 'MatrixBandPart', 'matrixdeterminant': 'MatrixDeterminant', 'matrixdiag': 'MatrixDiag', 'matrixdiagpart': 'MatrixDiagPart', 'matrixdiagpartv2': 'MatrixDiagPartV2', 'matrixdiagpartv3': 'MatrixDiagPartV3', 'matrixdiagv2': 'MatrixDiagV2', 'matrixdiagv3': 'MatrixDiagV3', 'matrixexponential': 'MatrixExponential', 'matrixinverse': 'MatrixInverse', 'matrixlogarithm': 'MatrixLogarithm', 'matrixsetdiag': 'MatrixSetDiag', 'matrixsetdiagv2': 'MatrixSetDiagV2', 'matrixsetdiagv3': 'MatrixSetDiagV3', 'matrixsolve': 'MatrixSolve', 'matrixsolvels': 'MatrixSolveLs', 'matrixsquareroot': 'MatrixSquareRoot', 'matrixtriangularsolve': 'MatrixTriangularSolve', 'max': 'Max', 'maximum': 'Maximum', 'maxintraopparallelismdataset': 'MaxIntraOpParallelismDataset', 'maxpool': 'MaxPool', 'maxpool3d': 'MaxPool3D', 'maxpool3dgrad': 'MaxPool3DGrad', 'maxpool3dgradgrad': 'MaxPool3DGradGrad', 'maxpoolgrad': 'MaxPoolGrad', 'maxpoolgradgrad': 'MaxPoolGradGrad', 'maxpoolgradgradv2': 'MaxPoolGradGradV2', 'maxpoolgradgradwithargmax': 'MaxPoolGradGradWithArgmax', 'maxpoolgradv2': 'MaxPoolGradV2', 'maxpoolgradwithargmax': 'MaxPoolGradWithArgmax', 'maxpoolv2': 'MaxPoolV2', 'maxpoolwithargmax': 'MaxPoolWithArgmax', 'mean': 'Mean', 'merge': 'Merge', 'mergesummary': 'MergeSummary', 'mergev2checkpoints': 'MergeV2Checkpoints', 'mfcc': 'Mfcc', 'min': 'Min', 'minimum': 'Minimum', 'mirrorpad': 'MirrorPad', 'mirrorpadgrad': 'MirrorPadGrad', 'mod': 'Mod', 'modeldataset': 'ModelDataset', 'mul': 'Mul', 'mulnonan': 'MulNoNan', 'multideviceiterator': 'MultiDeviceIterator', 'multideviceiteratorfromstringhandle': 'MultiDeviceIteratorFromStringHandle', 'multideviceiteratorgetnextfromshard': 'MultiDeviceIteratorGetNextFromShard', 'multideviceiteratorinit': 'MultiDeviceIteratorInit', 'multideviceiteratortostringhandle': 'MultiDeviceIteratorToStringHandle', 'multinomial': 'Multinomial', 'mutabledensehashtable': 'MutableDenseHashTable', 'mutabledensehashtablev2': 'MutableDenseHashTableV2', 'mutablehashtable': 'MutableHashTable', 'mutablehashtableoftensors': 'MutableHashTableOfTensors', 'mutablehashtableoftensorsv2': 'MutableHashTableOfTensorsV2', 'mutablehashtablev2': 'MutableHashTableV2', 'mutexlock': 'MutexLock', 'mutexv2': 'MutexV2', 'ncclallreduce': 'NcclAllReduce', 'ncclbroadcast': 'NcclBroadcast', 'ncclreduce': 'NcclReduce', 'ndtri': 'Ndtri', 'neg': 'Neg', 'nextafter': 'NextAfter', 'nextiteration': 'NextIteration', 'nondeterministicints': 'NonDeterministicInts', 'nonmaxsuppression': 'NonMaxSuppression', 'nonmaxsuppressionv2': 'NonMaxSuppressionV2', 'nonmaxsuppressionv3': 'NonMaxSuppressionV3', 'nonmaxsuppressionv4': 'NonMaxSuppressionV4', 'nonmaxsuppressionv5': 'NonMaxSuppressionV5', 'nonmaxsuppressionwithoverlaps': 'NonMaxSuppressionWithOverlaps', 'nonserializabledataset': 'NonSerializableDataset', 'noop': 'NoOp', 'notequal': 'NotEqual', 'nthelement': 'NthElement', 'onehot': 'OneHot', 'oneshotiterator': 'OneShotIterator', 'oneslike': 'OnesLike', 'optimizedataset': 'OptimizeDataset', 'optionalfromvalue': 'OptionalFromValue', 'optionalgetvalue': 'OptionalGetValue', 'optionalhasvalue': 'OptionalHasValue', 'optionalnone': 'OptionalNone', 'orderedmapclear': 'OrderedMapClear', 'orderedmapincompletesize': 'OrderedMapIncompleteSize', 'orderedmappeek': 'OrderedMapPeek', 'orderedmapsize': 'OrderedMapSize', 'orderedmapstage': 'OrderedMapStage', 'orderedmapunstage': 'OrderedMapUnstage', 'orderedmapunstagenokey': 'OrderedMapUnstageNoKey', 'outfeeddequeue': 'OutfeedDequeue', 'outfeeddequeuetuple': 'OutfeedDequeueTuple', 'outfeedenqueue': 'OutfeedEnqueue', 'outfeedenqueuetuple': 'OutfeedEnqueueTuple', 'pack': 'Pack', 'pad': 'Pad', 'paddedbatchdataset': 'PaddedBatchDataset', 'paddedbatchdatasetv2': 'PaddedBatchDatasetV2', 'paddingfifoqueue': 'PaddingFIFOQueue', 'paddingfifoqueuev2': 'PaddingFIFOQueueV2', 'padv2': 'PadV2', 'parallelconcat': 'ParallelConcat', 'paralleldynamicstitch': 'ParallelDynamicStitch', 'parallelinterleavedataset': 'ParallelInterleaveDataset', 'parallelinterleavedatasetv2': 'ParallelInterleaveDatasetV2', 'parallelmapdataset': 'ParallelMapDataset', 'parameterizedtruncatednormal': 'ParameterizedTruncatedNormal', 'parseexample': 'ParseExample', 'parseexampledataset': 'ParseExampleDataset', 'parseexamplev2': 'ParseExampleV2', 'parsesequenceexample': 'ParseSequenceExample', 'parsesequenceexamplev2': 'ParseSequenceExampleV2', 'parsesingleexample': 'ParseSingleExample', 'parsesinglesequenceexample': 'ParseSingleSequenceExample', 'parsetensor': 'ParseTensor', 'partitionedcall': 'PartitionedCall', 'placeholder': 'Placeholder', 'placeholderv2': 'PlaceholderV2', 'placeholderwithdefault': 'PlaceholderWithDefault', 'polygamma': 'Polygamma', 'populationcount': 'PopulationCount', 'pow': 'Pow', 'prefetchdataset': 'PrefetchDataset', 'prelinearize': 'Prelinearize', 'prelinearizetuple': 'PrelinearizeTuple', 'preventgradient': 'PreventGradient', 'print': 'Print', 'printv2': 'PrintV2', 'priorityqueue': 'PriorityQueue', 'priorityqueuev2': 'PriorityQueueV2', 'privatethreadpooldataset': 'PrivateThreadPoolDataset', 'prod': 'Prod', 'pyfunc': 'PyFunc', 'pyfuncstateless': 'PyFuncStateless', 'qr': 'Qr', 'quantizeanddequantize': 'QuantizeAndDequantize', 'quantizeanddequantizev2': 'QuantizeAndDequantizeV2', 'quantizeanddequantizev3': 'QuantizeAndDequantizeV3', 'quantizedadd': 'QuantizedAdd', 'quantizedavgpool': 'QuantizedAvgPool', 'quantizedbatchnormwithglobalnormalization': 'QuantizedBatchNormWithGlobalNormalization', 'quantizedbiasadd': 'QuantizedBiasAdd', 'quantizedconcat': 'QuantizedConcat', 'quantizedconv2d': 'QuantizedConv2D', 'quantizedconv2dandrelu': 'QuantizedConv2DAndRelu', 'quantizedconv2dandreluandrequantize': 'QuantizedConv2DAndReluAndRequantize', 'quantizedconv2dandrequantize': 'QuantizedConv2DAndRequantize', 'quantizedconv2dperchannel': 'QuantizedConv2DPerChannel', 'quantizedconv2dwithbias': 'QuantizedConv2DWithBias', 'quantizedconv2dwithbiasandrelu': 'QuantizedConv2DWithBiasAndRelu', 'quantizedconv2dwithbiasandreluandrequantize': 'QuantizedConv2DWithBiasAndReluAndRequantize', 'quantizedconv2dwithbiasandrequantize': 'QuantizedConv2DWithBiasAndRequantize', 'quantizedconv2dwithbiassignedsumandreluandrequantize': 'QuantizedConv2DWithBiasSignedSumAndReluAndRequantize', 'quantizedconv2dwithbiassumandrelu': 'QuantizedConv2DWithBiasSumAndRelu', 'quantizedconv2dwithbiassumandreluandrequantize': 'QuantizedConv2DWithBiasSumAndReluAndRequantize', 'quantizeddepthwiseconv2d': 'QuantizedDepthwiseConv2D', 'quantizeddepthwiseconv2dwithbias': 'QuantizedDepthwiseConv2DWithBias', 'quantizeddepthwiseconv2dwithbiasandrelu': 'QuantizedDepthwiseConv2DWithBiasAndRelu', 'quantizeddepthwiseconv2dwithbiasandreluandrequantize': 'QuantizedDepthwiseConv2DWithBiasAndReluAndRequantize', 'quantizedinstancenorm': 'QuantizedInstanceNorm', 'quantizedmatmul': 'QuantizedMatMul', 'quantizedmatmulwithbias': 'QuantizedMatMulWithBias', 'quantizedmatmulwithbiasandrelu': 'QuantizedMatMulWithBiasAndRelu', 'quantizedmatmulwithbiasandreluandrequantize': 'QuantizedMatMulWithBiasAndReluAndRequantize', 'quantizedmatmulwithbiasandrequantize': 'QuantizedMatMulWithBiasAndRequantize', 'quantizedmaxpool': 'QuantizedMaxPool', 'quantizedmul': 'QuantizedMul', 'quantizedownandshrinkrange': 'QuantizeDownAndShrinkRange', 'quantizedrelu': 'QuantizedRelu', 'quantizedrelu6': 'QuantizedRelu6', 'quantizedrelux': 'QuantizedReluX', 'quantizedreshape': 'QuantizedReshape', 'quantizedresizebilinear': 'QuantizedResizeBilinear', 'quantizev2': 'QuantizeV2', 'queueclose': 'QueueClose', 'queueclosev2': 'QueueCloseV2', 'queuedequeue': 'QueueDequeue', 'queuedequeuemany': 'QueueDequeueMany', 'queuedequeuemanyv2': 'QueueDequeueManyV2', 'queuedequeueupto': 'QueueDequeueUpTo', 'queuedequeueuptov2': 'QueueDequeueUpToV2', 'queuedequeuev2': 'QueueDequeueV2', 'queueenqueue': 'QueueEnqueue', 'queueenqueuemany': 'QueueEnqueueMany', 'queueenqueuemanyv2': 'QueueEnqueueManyV2', 'queueenqueuev2': 'QueueEnqueueV2', 'queueisclosed': 'QueueIsClosed', 'queueisclosedv2': 'QueueIsClosedV2', 'queuesize': 'QueueSize', 'queuesizev2': 'QueueSizeV2', 'raggedgather': 'RaggedGather', 'raggedrange': 'RaggedRange', 'raggedtensorfromvariant': 'RaggedTensorFromVariant', 'raggedtensortosparse': 'RaggedTensorToSparse', 'raggedtensortotensor': 'RaggedTensorToTensor', 'raggedtensortovariant': 'RaggedTensorToVariant', 'randomcrop': 'RandomCrop', 'randomdataset': 'RandomDataset', 'randomgamma': 'RandomGamma', 'randomgammagrad': 'RandomGammaGrad', 'randompoisson': 'RandomPoisson', 'randompoissonv2': 'RandomPoissonV2', 'randomshuffle': 'RandomShuffle', 'randomshufflequeue': 'RandomShuffleQueue', 'randomshufflequeuev2': 'RandomShuffleQueueV2', 'randomstandardnormal': 'RandomStandardNormal', 'randomuniform': 'RandomUniform', 'randomuniformint': 'RandomUniformInt', 'range': 'Range', 'rangedataset': 'RangeDataset', 'rank': 'Rank', 'readernumrecordsproduced': 'ReaderNumRecordsProduced', 'readernumrecordsproducedv2': 'ReaderNumRecordsProducedV2', 'readernumworkunitscompleted': 'ReaderNumWorkUnitsCompleted', 'readernumworkunitscompletedv2': 'ReaderNumWorkUnitsCompletedV2', 'readerread': 'ReaderRead', 'readerreadupto': 'ReaderReadUpTo', 'readerreaduptov2': 'ReaderReadUpToV2', 'readerreadv2': 'ReaderReadV2', 'readerreset': 'ReaderReset', 'readerresetv2': 'ReaderResetV2', 'readerrestorestate': 'ReaderRestoreState', 'readerrestorestatev2': 'ReaderRestoreStateV2', 'readerserializestate': 'ReaderSerializeState', 'readerserializestatev2': 'ReaderSerializeStateV2', 'readfile': 'ReadFile', 'readvariableop': 'ReadVariableOp', 'real': 'Real', 'realdiv': 'RealDiv', 'rebatchdataset': 'RebatchDataset', 'reciprocal': 'Reciprocal', 'reciprocalgrad': 'ReciprocalGrad', 'recordinput': 'RecordInput', 'recv': 'Recv', 'recvtpuembeddingactivations': 'RecvTPUEmbeddingActivations', 'reducedataset': 'ReduceDataset', 'reducejoin': 'ReduceJoin', 'refenter': 'RefEnter', 'refexit': 'RefExit', 'refidentity': 'RefIdentity', 'refmerge': 'RefMerge', 'refnextiteration': 'RefNextIteration', 'refselect': 'RefSelect', 'refswitch': 'RefSwitch', 'regexfullmatch': 'RegexFullMatch', 'regexreplace': 'RegexReplace', 'relu': 'Relu', 'relu6': 'Relu6', 'relu6grad': 'Relu6Grad', 'relugrad': 'ReluGrad', 'remotecall': 'RemoteCall', 'repeatdataset': 'RepeatDataset', 'requantizationrange': 'RequantizationRange', 'requantizationrangeperchannel': 'RequantizationRangePerChannel', 'requantize': 'Requantize', 'requantizeperchannel': 'RequantizePerChannel', 'reshape': 'Reshape', 'resizearea': 'ResizeArea', 'resizebicubic': 'ResizeBicubic', 'resizebicubicgrad': 'ResizeBicubicGrad', 'resizebilinear': 'ResizeBilinear', 'resizebilineargrad': 'ResizeBilinearGrad', 'resizenearestneighbor': 'ResizeNearestNeighbor', 'resizenearestneighborgrad': 'ResizeNearestNeighborGrad', 'resourceaccumulatorapplygradient': 'ResourceAccumulatorApplyGradient', 'resourceaccumulatornumaccumulated': 'ResourceAccumulatorNumAccumulated', 'resourceaccumulatorsetglobalstep': 'ResourceAccumulatorSetGlobalStep', 'resourceaccumulatortakegradient': 'ResourceAccumulatorTakeGradient', 'resourceapplyadadelta': 'ResourceApplyAdadelta', 'resourceapplyadagrad': 'ResourceApplyAdagrad', 'resourceapplyadagradda': 'ResourceApplyAdagradDA', 'resourceapplyadagradv2': 'ResourceApplyAdagradV2', 'resourceapplyadam': 'ResourceApplyAdam', 'resourceapplyadamax': 'ResourceApplyAdaMax', 'resourceapplyadamwithamsgrad': 'ResourceApplyAdamWithAmsgrad', 'resourceapplyaddsign': 'ResourceApplyAddSign', 'resourceapplycenteredrmsprop': 'ResourceApplyCenteredRMSProp', 'resourceapplyftrl': 'ResourceApplyFtrl', 'resourceapplyftrlv2': 'ResourceApplyFtrlV2', 'resourceapplygradientdescent': 'ResourceApplyGradientDescent', 'resourceapplykerasmomentum': 'ResourceApplyKerasMomentum', 'resourceapplymomentum': 'ResourceApplyMomentum', 'resourceapplypowersign': 'ResourceApplyPowerSign', 'resourceapplyproximaladagrad': 'ResourceApplyProximalAdagrad', 'resourceapplyproximalgradientdescent': 'ResourceApplyProximalGradientDescent', 'resourceapplyrmsprop': 'ResourceApplyRMSProp', 'resourceconditionalaccumulator': 'ResourceConditionalAccumulator', 'resourcecountupto': 'ResourceCountUpTo', 'resourcegather': 'ResourceGather', 'resourcegathernd': 'ResourceGatherNd', 'resourcescatteradd': 'ResourceScatterAdd', 'resourcescatterdiv': 'ResourceScatterDiv', 'resourcescattermax': 'ResourceScatterMax', 'resourcescattermin': 'ResourceScatterMin', 'resourcescattermul': 'ResourceScatterMul', 'resourcescatterndadd': 'ResourceScatterNdAdd', 'resourcescatterndsub': 'ResourceScatterNdSub', 'resourcescatterndupdate': 'ResourceScatterNdUpdate', 'resourcescattersub': 'ResourceScatterSub', 'resourcescatterupdate': 'ResourceScatterUpdate', 'resourcesparseapplyadadelta': 'ResourceSparseApplyAdadelta', 'resourcesparseapplyadagrad': 'ResourceSparseApplyAdagrad', 'resourcesparseapplyadagradda': 'ResourceSparseApplyAdagradDA', 'resourcesparseapplyadagradv2': 'ResourceSparseApplyAdagradV2', 'resourcesparseapplycenteredrmsprop': 'ResourceSparseApplyCenteredRMSProp', 'resourcesparseapplyftrl': 'ResourceSparseApplyFtrl', 'resourcesparseapplyftrlv2': 'ResourceSparseApplyFtrlV2', 'resourcesparseapplykerasmomentum': 'ResourceSparseApplyKerasMomentum', 'resourcesparseapplymomentum': 'ResourceSparseApplyMomentum', 'resourcesparseapplyproximaladagrad': 'ResourceSparseApplyProximalAdagrad', 'resourcesparseapplyproximalgradientdescent': 'ResourceSparseApplyProximalGradientDescent', 'resourcesparseapplyrmsprop': 'ResourceSparseApplyRMSProp', 'resourcestridedsliceassign': 'ResourceStridedSliceAssign', 'restore': 'Restore', 'restoreslice': 'RestoreSlice', 'restorev2': 'RestoreV2', 'retrievetpuembeddingadadeltaparameters': 'RetrieveTPUEmbeddingAdadeltaParameters', 'retrievetpuembeddingadadeltaparametersgradaccumdebug': 'RetrieveTPUEmbeddingAdadeltaParametersGradAccumDebug', 'retrievetpuembeddingadagradparameters': 'RetrieveTPUEmbeddingAdagradParameters', 'retrievetpuembeddingadagradparametersgradaccumdebug': 'RetrieveTPUEmbeddingAdagradParametersGradAccumDebug', 'retrievetpuembeddingadamparameters': 'RetrieveTPUEmbeddingADAMParameters', 'retrievetpuembeddingadamparametersgradaccumdebug': 'RetrieveTPUEmbeddingADAMParametersGradAccumDebug', 'retrievetpuembeddingcenteredrmspropparameters': 'RetrieveTPUEmbeddingCenteredRMSPropParameters', 'retrievetpuembeddingftrlparameters': 'RetrieveTPUEmbeddingFTRLParameters', 'retrievetpuembeddingftrlparametersgradaccumdebug': 'RetrieveTPUEmbeddingFTRLParametersGradAccumDebug', 'retrievetpuembeddingmdladagradlightparameters': 'RetrieveTPUEmbeddingMDLAdagradLightParameters', 'retrievetpuembeddingmomentumparameters': 'RetrieveTPUEmbeddingMomentumParameters', 'retrievetpuembeddingmomentumparametersgradaccumdebug': 'RetrieveTPUEmbeddingMomentumParametersGradAccumDebug', 'retrievetpuembeddingproximaladagradparameters': 'RetrieveTPUEmbeddingProximalAdagradParameters', 'retrievetpuembeddingproximaladagradparametersgradaccumdebug': 'RetrieveTPUEmbeddingProximalAdagradParametersGradAccumDebug', 'retrievetpuembeddingrmspropparameters': 'RetrieveTPUEmbeddingRMSPropParameters', 'retrievetpuembeddingrmspropparametersgradaccumdebug': 'RetrieveTPUEmbeddingRMSPropParametersGradAccumDebug', 'retrievetpuembeddingstochasticgradientdescentparameters': 'RetrieveTPUEmbeddingStochasticGradientDescentParameters', 'reverse': 'Reverse', 'reversesequence': 'ReverseSequence', 'reversev2': 'ReverseV2', 'rfft': 'RFFT', 'rfft2d': 'RFFT2D', 'rfft3d': 'RFFT3D', 'rgbtohsv': 'RGBToHSV', 'rightshift': 'RightShift', 'rint': 'Rint', 'rngskip': 'RngSkip', 'roll': 'Roll', 'round': 'Round', 'rsqrt': 'Rsqrt', 'rsqrtgrad': 'RsqrtGrad', 'sampledistortedboundingbox': 'SampleDistortedBoundingBox', 'sampledistortedboundingboxv2': 'SampleDistortedBoundingBoxV2', 'samplingdataset': 'SamplingDataset', 'save': 'Save', 'saveslices': 'SaveSlices', 'savev2': 'SaveV2', 'scalarsummary': 'ScalarSummary', 'scaleandtranslate': 'ScaleAndTranslate', 'scaleandtranslategrad': 'ScaleAndTranslateGrad', 'scandataset': 'ScanDataset', 'scatteradd': 'ScatterAdd', 'scatterdiv': 'ScatterDiv', 'scattermax': 'ScatterMax', 'scattermin': 'ScatterMin', 'scattermul': 'ScatterMul', 'scatternd': 'ScatterNd', 'scatterndadd': 'ScatterNdAdd', 'scatterndnonaliasingadd': 'ScatterNdNonAliasingAdd', 'scatterndsub': 'ScatterNdSub', 'scatterndupdate': 'ScatterNdUpdate', 'scattersub': 'ScatterSub', 'scatterupdate': 'ScatterUpdate', 'sdcafprint': 'SdcaFprint', 'sdcaoptimizer': 'SdcaOptimizer', 'sdcaoptimizerv2': 'SdcaOptimizerV2', 'sdcashrinkl1': 'SdcaShrinkL1', 'segmentmax': 'SegmentMax', 'segmentmean': 'SegmentMean', 'segmentmin': 'SegmentMin', 'segmentprod': 'SegmentProd', 'segmentsum': 'SegmentSum', 'select': 'Select', 'selectv2': 'SelectV2', 'selfadjointeig': 'SelfAdjointEig', 'selfadjointeigv2': 'SelfAdjointEigV2', 'selu': 'Selu', 'selugrad': 'SeluGrad', 'send': 'Send', 'sendtpuembeddinggradients': 'SendTPUEmbeddingGradients', 'serializeiterator': 'SerializeIterator', 'serializemanysparse': 'SerializeManySparse', 'serializesparse': 'SerializeSparse', 'serializetensor': 'SerializeTensor', 'setsize': 'SetSize', 'setstatsaggregatordataset': 'SetStatsAggregatorDataset', 'shape': 'Shape', 'shapen': 'ShapeN', 'sharddataset': 'ShardDataset', 'shardedfilename': 'ShardedFilename', 'shardedfilespec': 'ShardedFilespec', 'shuffleandrepeatdataset': 'ShuffleAndRepeatDataset', 'shuffledataset': 'ShuffleDataset', 'shuffledatasetv2': 'ShuffleDatasetV2', 'shutdowndistributedtpu': 'ShutdownDistributedTPU', 'sigmoid': 'Sigmoid', 'sigmoidgrad': 'SigmoidGrad', 'sign': 'Sign', 'sin': 'Sin', 'sinh': 'Sinh', 'size': 'Size', 'skipdataset': 'SkipDataset', 'sleepdataset': 'SleepDataset', 'slice': 'Slice', 'slidingwindowdataset': 'SlidingWindowDataset', 'snapshot': 'Snapshot', 'snapshotdataset': 'SnapshotDataset', 'sobolsample': 'SobolSample', 'softmax': 'Softmax', 'softmaxcrossentropywithlogits': 'SoftmaxCrossEntropyWithLogits', 'softplus': 'Softplus', 'softplusgrad': 'SoftplusGrad', 'softsign': 'Softsign', 'softsigngrad': 'SoftsignGrad', 'spacetobatch': 'SpaceToBatch', 'spacetobatchnd': 'SpaceToBatchND', 'spacetodepth': 'SpaceToDepth', 'sparseaccumulatorapplygradient': 'SparseAccumulatorApplyGradient', 'sparseaccumulatortakegradient': 'SparseAccumulatorTakeGradient', 'sparseadd': 'SparseAdd', 'sparseaddgrad': 'SparseAddGrad', 'sparseapplyadadelta': 'SparseApplyAdadelta', 'sparseapplyadagrad': 'SparseApplyAdagrad', 'sparseapplyadagradda': 'SparseApplyAdagradDA', 'sparseapplyadagradv2': 'SparseApplyAdagradV2', 'sparseapplycenteredrmsprop': 'SparseApplyCenteredRMSProp', 'sparseapplyftrl': 'SparseApplyFtrl', 'sparseapplyftrlv2': 'SparseApplyFtrlV2', 'sparseapplymomentum': 'SparseApplyMomentum', 'sparseapplyproximaladagrad': 'SparseApplyProximalAdagrad', 'sparseapplyproximalgradientdescent': 'SparseApplyProximalGradientDescent', 'sparseapplyrmsprop': 'SparseApplyRMSProp', 'sparseconcat': 'SparseConcat', 'sparseconditionalaccumulator': 'SparseConditionalAccumulator', 'sparsecross': 'SparseCross', 'sparsedensecwiseadd': 'SparseDenseCwiseAdd', 'sparsedensecwisediv': 'SparseDenseCwiseDiv', 'sparsedensecwisemul': 'SparseDenseCwiseMul', 'sparsefillemptyrows': 'SparseFillEmptyRows', 'sparsefillemptyrowsgrad': 'SparseFillEmptyRowsGrad', 'sparsematmul': 'SparseMatMul', 'sparsematrixadd': 'SparseMatrixAdd', 'sparsematrixmatmul': 'SparseMatrixMatMul', 'sparsematrixmul': 'SparseMatrixMul', 'sparsematrixnnz': 'SparseMatrixNNZ', 'sparsematrixorderingamd': 'SparseMatrixOrderingAMD', 'sparsematrixsoftmax': 'SparseMatrixSoftmax', 'sparsematrixsoftmaxgrad': 'SparseMatrixSoftmaxGrad', 'sparsematrixsparsecholesky': 'SparseMatrixSparseCholesky', 'sparsematrixsparsematmul': 'SparseMatrixSparseMatMul', 'sparsematrixtranspose': 'SparseMatrixTranspose', 'sparsematrixzeros': 'SparseMatrixZeros', 'sparsereducemax': 'SparseReduceMax', 'sparsereducemaxsparse': 'SparseReduceMaxSparse', 'sparsereducesum': 'SparseReduceSum', 'sparsereducesumsparse': 'SparseReduceSumSparse', 'sparsereorder': 'SparseReorder', 'sparsereshape': 'SparseReshape', 'sparsesegmentmean': 'SparseSegmentMean', 'sparsesegmentmeangrad': 'SparseSegmentMeanGrad', 'sparsesegmentmeanwithnumsegments': 'SparseSegmentMeanWithNumSegments', 'sparsesegmentsqrtn': 'SparseSegmentSqrtN', 'sparsesegmentsqrtngrad': 'SparseSegmentSqrtNGrad', 'sparsesegmentsqrtnwithnumsegments': 'SparseSegmentSqrtNWithNumSegments', 'sparsesegmentsum': 'SparseSegmentSum', 'sparsesegmentsumwithnumsegments': 'SparseSegmentSumWithNumSegments', 'sparseslice': 'SparseSlice', 'sparseslicegrad': 'SparseSliceGrad', 'sparsesoftmax': 'SparseSoftmax', 'sparsesoftmaxcrossentropywithlogits': 'SparseSoftmaxCrossEntropyWithLogits', 'sparsesparsemaximum': 'SparseSparseMaximum', 'sparsesparseminimum': 'SparseSparseMinimum', 'sparsesplit': 'SparseSplit', 'sparsetensordenseadd': 'SparseTensorDenseAdd', 'sparsetensordensematmul': 'SparseTensorDenseMatMul', 'sparsetensorslicedataset': 'SparseTensorSliceDataset', 'sparsetensortocsrsparsematrix': 'SparseTensorToCSRSparseMatrix', 'sparsetodense': 'SparseToDense', 'sparsetosparsesetoperation': 'SparseToSparseSetOperation', 'spence': 'Spence', 'split': 'Split', 'splitv': 'SplitV', 'sqldataset': 'SqlDataset', 'sqrt': 'Sqrt', 'sqrtgrad': 'SqrtGrad', 'square': 'Square', 'squareddifference': 'SquaredDifference', 'squeeze': 'Squeeze', 'stack': 'Stack', 'stackclose': 'StackClose', 'stackclosev2': 'StackCloseV2', 'stackpop': 'StackPop', 'stackpopv2': 'StackPopV2', 'stackpush': 'StackPush', 'stackpushv2': 'StackPushV2', 'stackv2': 'StackV2', 'stage': 'Stage', 'stageclear': 'StageClear', 'stagepeek': 'StagePeek', 'stagesize': 'StageSize', 'statefulpartitionedcall': 'StatefulPartitionedCall', 'statefulrandombinomial': 'StatefulRandomBinomial', 'statefulstandardnormal': 'StatefulStandardNormal', 'statefulstandardnormalv2': 'StatefulStandardNormalV2', 'statefultruncatednormal': 'StatefulTruncatedNormal', 'statefuluniform': 'StatefulUniform', 'statefuluniformfullint': 'StatefulUniformFullInt', 'statefuluniformint': 'StatefulUniformInt', 'statelessif': 'StatelessIf', 'statelessmultinomial': 'StatelessMultinomial', 'statelessrandomnormal': 'StatelessRandomNormal', 'statelessrandomuniform': 'StatelessRandomUniform', 'statelessrandomuniformint': 'StatelessRandomUniformInt', 'statelesstruncatednormal': 'StatelessTruncatedNormal', 'statelesswhile': 'StatelessWhile', 'staticregexfullmatch': 'StaticRegexFullMatch', 'staticregexreplace': 'StaticRegexReplace', 'statsaggregatorhandle': 'StatsAggregatorHandle', 'statsaggregatorhandlev2': 'StatsAggregatorHandleV2', 'statsaggregatorsetsummarywriter': 'StatsAggregatorSetSummaryWriter', 'statsaggregatorsummary': 'StatsAggregatorSummary', 'stopgradient': 'StopGradient', 'stridedslice': 'StridedSlice', 'stridedsliceassign': 'StridedSliceAssign', 'stridedslicegrad': 'StridedSliceGrad', 'stringformat': 'StringFormat', 'stringjoin': 'StringJoin', 'stringlength': 'StringLength', 'stringlower': 'StringLower', 'stringngrams': 'StringNGrams', 'stringsplit': 'StringSplit', 'stringsplitv2': 'StringSplitV2', 'stringstrip': 'StringStrip', 'stringtohashbucket': 'StringToHashBucket', 'stringtohashbucketfast': 'StringToHashBucketFast', 'stringtohashbucketstrong': 'StringToHashBucketStrong', 'stringtonumber': 'StringToNumber', 'stringupper': 'StringUpper', 'sub': 'Sub', 'substr': 'Substr', 'sum': 'Sum', 'summarywriter': 'SummaryWriter', 'svd': 'Svd', 'switch': 'Switch', 'symbolicgradient': 'SymbolicGradient', 'takedataset': 'TakeDataset', 'takemanysparsefromtensorsmap': 'TakeManySparseFromTensorsMap', 'takewhiledataset': 'TakeWhileDataset', 'tan': 'Tan', 'tanh': 'Tanh', 'tanhgrad': 'TanhGrad', 'temporaryvariable': 'TemporaryVariable', 'tensorarray': 'TensorArray', 'tensorarrayclose': 'TensorArrayClose', 'tensorarrayclosev2': 'TensorArrayCloseV2', 'tensorarrayclosev3': 'TensorArrayCloseV3', 'tensorarrayconcat': 'TensorArrayConcat', 'tensorarrayconcatv2': 'TensorArrayConcatV2', 'tensorarrayconcatv3': 'TensorArrayConcatV3', 'tensorarraygather': 'TensorArrayGather', 'tensorarraygatherv2': 'TensorArrayGatherV2', 'tensorarraygatherv3': 'TensorArrayGatherV3', 'tensorarraygrad': 'TensorArrayGrad', 'tensorarraygradv2': 'TensorArrayGradV2', 'tensorarraygradv3': 'TensorArrayGradV3', 'tensorarraygradwithshape': 'TensorArrayGradWithShape', 'tensorarraypack': 'TensorArrayPack', 'tensorarrayread': 'TensorArrayRead', 'tensorarrayreadv2': 'TensorArrayReadV2', 'tensorarrayreadv3': 'TensorArrayReadV3', 'tensorarrayscatter': 'TensorArrayScatter', 'tensorarrayscatterv2': 'TensorArrayScatterV2', 'tensorarrayscatterv3': 'TensorArrayScatterV3', 'tensorarraysize': 'TensorArraySize', 'tensorarraysizev2': 'TensorArraySizeV2', 'tensorarraysizev3': 'TensorArraySizeV3', 'tensorarraysplit': 'TensorArraySplit', 'tensorarraysplitv2': 'TensorArraySplitV2', 'tensorarraysplitv3': 'TensorArraySplitV3', 'tensorarrayunpack': 'TensorArrayUnpack', 'tensorarrayv2': 'TensorArrayV2', 'tensorarrayv3': 'TensorArrayV3', 'tensorarraywrite': 'TensorArrayWrite', 'tensorarraywritev2': 'TensorArrayWriteV2', 'tensorarraywritev3': 'TensorArrayWriteV3', 'tensordataset': 'TensorDataset', 'tensorlistconcat': 'TensorListConcat', 'tensorlistconcatlists': 'TensorListConcatLists', 'tensorlistconcatv2': 'TensorListConcatV2', 'tensorlistelementshape': 'TensorListElementShape', 'tensorlistfromtensor': 'TensorListFromTensor', 'tensorlistgather': 'TensorListGather', 'tensorlistgetitem': 'TensorListGetItem', 'tensorlistlength': 'TensorListLength', 'tensorlistpopback': 'TensorListPopBack', 'tensorlistpushback': 'TensorListPushBack', 'tensorlistpushbackbatch': 'TensorListPushBackBatch', 'tensorlistreserve': 'TensorListReserve', 'tensorlistresize': 'TensorListResize', 'tensorlistscatter': 'TensorListScatter', 'tensorlistscatterintoexistinglist': 'TensorListScatterIntoExistingList', 'tensorlistscatterv2': 'TensorListScatterV2', 'tensorlistsetitem': 'TensorListSetItem', 'tensorlistsplit': 'TensorListSplit', 'tensorliststack': 'TensorListStack', 'tensorscatteradd': 'TensorScatterAdd', 'tensorscattersub': 'TensorScatterSub', 'tensorscatterupdate': 'TensorScatterUpdate', 'tensorslicedataset': 'TensorSliceDataset', 'tensorstridedsliceupdate': 'TensorStridedSliceUpdate', 'tensorsummary': 'TensorSummary', 'tensorsummaryv2': 'TensorSummaryV2', 'textlinedataset': 'TextLineDataset', 'textlinereader': 'TextLineReader', 'textlinereaderv2': 'TextLineReaderV2', 'tfrecorddataset': 'TFRecordDataset', 'tfrecordreader': 'TFRecordReader', 'tfrecordreaderv2': 'TFRecordReaderV2', 'threadpooldataset': 'ThreadPoolDataset', 'threadpoolhandle': 'ThreadPoolHandle', 'threadunsafeunigramcandidatesampler': 'ThreadUnsafeUnigramCandidateSampler', 'tile': 'Tile', 'tilegrad': 'TileGrad', 'timestamp': 'Timestamp', 'topk': 'TopK', 'topkv2': 'TopKV2', 'tpucompilationresult': 'TPUCompilationResult', 'tpuembeddingactivations': 'TPUEmbeddingActivations', 'tpuordinalselector': 'TPUOrdinalSelector', 'tpupartitionedcall': 'TPUPartitionedCall', 'tpureplicatedinput': 'TPUReplicatedInput', 'tpureplicatedoutput': 'TPUReplicatedOutput', 'tpureplicatemetadata': 'TPUReplicateMetadata', 'transpose': 'Transpose', 'tridiagonalmatmul': 'TridiagonalMatMul', 'tridiagonalsolve': 'TridiagonalSolve', 'truncatediv': 'TruncateDiv', 'truncatednormal': 'TruncatedNormal', 'truncatemod': 'TruncateMod', 'unbatch': 'Unbatch', 'unbatchdataset': 'UnbatchDataset', 'unbatchgrad': 'UnbatchGrad', 'unicodedecode': 'UnicodeDecode', 'unicodedecodewithoffsets': 'UnicodeDecodeWithOffsets', 'unicodeencode': 'UnicodeEncode', 'unicodescript': 'UnicodeScript', 'unicodetranscode': 'UnicodeTranscode', 'uniformcandidatesampler': 'UniformCandidateSampler', 'unique': 'Unique', 'uniquedataset': 'UniqueDataset', 'uniquev2': 'UniqueV2', 'uniquewithcounts': 'UniqueWithCounts', 'uniquewithcountsv2': 'UniqueWithCountsV2', 'unpack': 'Unpack', 'unravelindex': 'UnravelIndex', 'unsortedsegmentjoin': 'UnsortedSegmentJoin', 'unsortedsegmentmax': 'UnsortedSegmentMax', 'unsortedsegmentmin': 'UnsortedSegmentMin', 'unsortedsegmentprod': 'UnsortedSegmentProd', 'unsortedsegmentsum': 'UnsortedSegmentSum', 'unstage': 'Unstage', 'unwrapdatasetvariant': 'UnwrapDatasetVariant', 'upperbound': 'UpperBound', 'varhandleop': 'VarHandleOp', 'variable': 'Variable', 'variableshape': 'VariableShape', 'variablev2': 'VariableV2', 'varisinitializedop': 'VarIsInitializedOp', 'where': 'Where', 'while': 'While', 'wholefilereader': 'WholeFileReader', 'wholefilereaderv2': 'WholeFileReaderV2', 'windowdataset': 'WindowDataset', 'workerheartbeat': 'WorkerHeartbeat', 'wrapdatasetvariant': 'WrapDatasetVariant', 'writeaudiosummary': 'WriteAudioSummary', 'writefile': 'WriteFile', 'writegraphsummary': 'WriteGraphSummary', 'writehistogramsummary': 'WriteHistogramSummary', 'writeimagesummary': 'WriteImageSummary', 'writerawprotosummary': 'WriteRawProtoSummary', 'writescalarsummary': 'WriteScalarSummary', 'writesummary': 'WriteSummary', 'xdivy': 'Xdivy', 'xlog1py': 'Xlog1py', 'xlogy': 'Xlogy', 'zeroslike': 'ZerosLike', 'zeta': 'Zeta', 'zipdataset': 'ZipDataset'}
classmethod make_opdef_sig(opdef, opdef_py_func=None)

Create a Signature object for an OpDef.

Annotations are include so that one can partially verify arguments.

opdef_signatures = {'AddN': (<Signature (inputs, name=None) -> [('sum', 'T')]>, <function add_n>), 'AddV2': (<Signature (x, y, name=None) -> [('z', 'T')]>, <function add_v2>), 'Const': (<Signature (value, dtype, name=None) -> [('output', 'dtype')]>, <function MetaOpDefLibrary.__init__.<locals>.mt_const>), 'Mul': (<Signature (x, y, name=None) -> [('z', 'T')]>, <function mul>)}
class symbolic_pymc.tensorflow.meta.OpDefFactoryType

Bases: symbolic_pymc.meta.MetaSymbolType

class symbolic_pymc.tensorflow.meta.TFlowMetaAccessor(namespace=None)

Bases: object

An accessor object that simplifies the use of meta objects.

Instances of this class can be used to implicitly convert TensorFlow functions and objects into meta objects.

namespaces = [<module 'tensorflow' from '/home/bwillard/apps/anaconda3/envs/symbolic-pymc/lib/python3.7/site-packages/tensorflow/__init__.py'>, <module 'tensorflow_core._api.v2.raw_ops' from '/home/bwillard/apps/anaconda3/envs/symbolic-pymc/lib/python3.7/site-packages/tensorflow_core/_api/v2/raw_ops/__init__.py'>, <module 'tensorflow_probability' from '/home/bwillard/apps/anaconda3/envs/symbolic-pymc/lib/python3.7/site-packages/tensorflow_probability/__init__.py'>, <module 'tensorflow_probability.python.distributions' from '/home/bwillard/apps/anaconda3/envs/symbolic-pymc/lib/python3.7/site-packages/tensorflow_probability/python/distributions/__init__.py'>, <module 'symbolic_pymc.tensorflow.meta' from '/home/bwillard/projects/code/python/symbolic-pymc/symbolic_pymc/tensorflow/meta.py'>]
class symbolic_pymc.tensorflow.meta.TFlowMetaNodeDef(op, name, attr, obj=None)

Bases: symbolic_pymc.tensorflow.meta.TFlowMetaSymbol

A meta NodeDef.

NOTE: We’re ignoring node_def.input; it’s just an unnecessary hassle.

attr
base

alias of tensorflow.core.framework.node_def_pb2.NodeDef

property frozen_attr
name
op
reset()
class symbolic_pymc.tensorflow.meta.TFlowMetaOp(op_def, node_def, inputs, outputs=None, obj=None)

Bases: symbolic_pymc.tensorflow.meta.TFlowMetaSymbol

A meta Operation.

This is like an Apply node in Theano.

TODO: This whole thing should probably be a “NodeDef” class?

Create a TensorFlow meta Operation.

The real signature of tf.Operation.__init__ includes the graph object, so we can’t really the signature directly. This is part of the reason why we have TFlowMetaOpFactory.__call__ and TFlowMetaTensor.operator + TFlowMetaTensor.inputs that do not directly use __all_props__/TFlowMetaTensor.rands and construct the objects directly.

base

alias of tensorflow.python.framework.ops.Operation

property default_output

Return the default output for this Operation.

TODO: It might be worth considering a direct approach, and not one that requires the generation of all meta outputs.

inputs
property name
node_def
op_def
property outputs

Compute outputs for this meta Operation.

reify()

Attempt to create a concrete base object from this meta object.

During the process, dependent objects will need to be reified, which may result in updates to the object(s) being reified.

For instance, if a meta tensor’s parent operator is fully reifiable to a base object, then the meta tensor’s dtype and shape may be fixed: e.g. a tensor corresponding to the output of a sum of two float64 scalars is necessarily a float64 scalar.

This function will set any unspecified properties (e.g. dtype and shape values for the previous example), mutating the object in-place when possible. It will return a [refined/partially reified] meta object when it can’t fully reify to a base object (in which case, it will return the base object) or when partial reification results in a meta object from a subclass.

reset()
property type
class symbolic_pymc.tensorflow.meta.TFlowMetaOpDef(obj=None)

Bases: symbolic_pymc.tensorflow.meta.TFlowMetaSymbol

A meta OpDef.

This is like an Op node in Theano.

Some useful info/links:
property attr
base

alias of tensorflow.core.framework.op_def_pb2.OpDef

reify()

Attempt to create a concrete base object from this meta object.

During the process, dependent objects will need to be reified, which may result in updates to the object(s) being reified.

For instance, if a meta tensor’s parent operator is fully reifiable to a base object, then the meta tensor’s dtype and shape may be fixed: e.g. a tensor corresponding to the output of a sum of two float64 scalars is necessarily a float64 scalar.

This function will set any unspecified properties (e.g. dtype and shape values for the previous example), mutating the object in-place when possible. It will return a [refined/partially reified] meta object when it can’t fully reify to a base object (in which case, it will return the base object) or when partial reification results in a meta object from a subclass.

reset()
class symbolic_pymc.tensorflow.meta.TFlowMetaOperator(op_def, node_def=None, obj=None)

Bases: symbolic_pymc.tensorflow.meta.TFlowMetaSymbol, symbolic_pymc.meta.MetaOp

A class that implements the notion of an operator on top of TensorFlow’s OpDef and NodeDef objects.

With this abstraction, we can better model operators by distinguishing parameterized operators and their respective parameter values from the operator’s inputs, which may have similar properties across the entire family of operators (i.e. across all parameter values).

For example, addition is commutative in its arguments, so modeling addition as an operator parameterized on dtypes and/or names, we may want to preserve the distinction of the operators inputs and its parameterized values so that we can implement commutativity exclusively on the non-dtype/name inputs.

base = None
classmethod get_metaopdef(name)

Obtain a MetaOpDef for a given string name.

This is more flexible because it ignores things like string case (when the non-raw_ops name differs from the TF user-level API).

input_args(*args, apply_defaults=True, **kwargs)

Make args and kwargs conform to an OpDef’s “apply function” arguments.

In order to do this, we effectively need to map OpDef and NodeDef values to tf.raw_ops.* function arguments (i.e. the reverse of what op_def_library._apply_op_helper does).

Returns an OrderedDict.

node_def
op_args_to_operation_inputs(apply_arguments)

Map an OpDef’s “apply function” arguments to Operation inputs and a meta NodeDef.

op_def
output_meta_types(inputs=None)

Return a list of tuples containing object types and corresponding dtypes for the outputs of this OpDef.

This work is done in https://github.com/tensorflow/tensorflow/blob/master/tensorflow/python/framework/op_def_library.py#L337. Would be very nice if the dtype inference and/or NodeDef generation was abstracted-out from that function.

reify()

Attempt to create a concrete base object from this meta object.

During the process, dependent objects will need to be reified, which may result in updates to the object(s) being reified.

For instance, if a meta tensor’s parent operator is fully reifiable to a base object, then the meta tensor’s dtype and shape may be fixed: e.g. a tensor corresponding to the output of a sum of two float64 scalars is necessarily a float64 scalar.

This function will set any unspecified properties (e.g. dtype and shape values for the previous example), mutating the object in-place when possible. It will return a [refined/partially reified] meta object when it can’t fully reify to a base object (in which case, it will return the base object) or when partial reification results in a meta object from a subclass.

reset()
class symbolic_pymc.tensorflow.meta.TFlowMetaSymbol(obj=None)

Bases: symbolic_pymc.meta.MetaSymbol

reset()
validate_objs()
class symbolic_pymc.tensorflow.meta.TFlowMetaTensor(op, value_index, dtype, obj=None)

Bases: symbolic_pymc.tensorflow.meta.TFlowMetaSymbol, symbolic_pymc.meta.MetaVariable

base

alias of tensorflow.python.framework.ops.Tensor

property base_arguments

Return the base-level arguments.

These arguments used in conjunction with the callable self.base_operator should re-produce this variable.

property base_operator

Return a meta object representing a base-level operator.

It should be callable with all inputs necessary to reproduce this tensor given by self.base_arguments.

dtype
property name
op
reify()

Attempt to create a concrete base object from this meta object.

During the process, dependent objects will need to be reified, which may result in updates to the object(s) being reified.

For instance, if a meta tensor’s parent operator is fully reifiable to a base object, then the meta tensor’s dtype and shape may be fixed: e.g. a tensor corresponding to the output of a sum of two float64 scalars is necessarily a float64 scalar.

This function will set any unspecified properties (e.g. dtype and shape values for the previous example), mutating the object in-place when possible. It will return a [refined/partially reified] meta object when it can’t fully reify to a base object (in which case, it will return the base object) or when partial reification results in a meta object from a subclass.

reset()
property shape
value_index
class symbolic_pymc.tensorflow.meta.TFlowMetaTensorShape(dims, obj=None)

Bases: symbolic_pymc.tensorflow.meta.TFlowMetaSymbol

as_list()
base

alias of tensorflow.python.framework.tensor_shape.TensorShape

dims
property ndims
property rank
reset()
symbolic_pymc.tensorflow.meta.load_dispatcher()

Set/override dispatcher to default to TF objects.

symbolic_pymc.tensorflow.printing module

exception symbolic_pymc.tensorflow.printing.DepthExceededException

Bases: Exception

class symbolic_pymc.tensorflow.printing.TFlowPrinter(formatter, buffer, depth_lower_idx=0, depth_upper_idx=9223372036854775807)

Bases: object

A printer that indents and keeps track of already printed subgraphs.

format(obj)
indented(indent)
print(obj, suffix='')
println(obj)
subgraph_add(obj)
symbolic_pymc.tensorflow.printing.tf_dprint(obj, depth_lower=0, depth_upper=10, printer=None)

Print a textual representation of a TF graph. The output roughly follows the format of theano.printing.debugprint.

objTensorflow object

Tensorflow graph object to be represented.

depth_lowerint

Used to index specific portions of the graph.

depth_upperint

Used to index specific portions of the graph.

printeroptional

Backend used to display the output.

Module contents