# BNNSGraph.Builder.Tensor A structure that represents an abstract handle to a tensor that you use within a `BNNSGraph.makeContext` closure. ``` struct Tensor where T : BNNSScalar ``` ## Topics ### Operators [`static func .! (BNNSGraph.Builder.Tensor) -> BNNSGraph.Builder.Tensor`](/documentation/Accelerate/BNNSGraph/Builder/Tensor/.!(_:)) Adds an element-wise logical not operation to the current graph. [`static func .!= (BNNSGraph.Builder.Tensor, BNNSGraph.Builder.Tensor) -> BNNSGraph.Builder.Tensor`](/documentation/Accelerate/BNNSGraph/Builder/Tensor/.!=(_:_:)) Adds an element-wise inequality operation to the current graph. [`static func .& (BNNSGraph.Builder.Tensor, BNNSGraph.Builder.Tensor) -> BNNSGraph.Builder.Tensor`](/documentation/Accelerate/BNNSGraph/Builder/Tensor/.&(_:_:)) Adds an element-wise logical and operation to the current graph. [`static func .== (BNNSGraph.Builder.Tensor, BNNSGraph.Builder.Tensor) -> BNNSGraph.Builder.Tensor`](/documentation/Accelerate/BNNSGraph/Builder/Tensor/.==(_:_:)) Adds an element-wise equality operation to the current graph. [`static func .| (BNNSGraph.Builder.Tensor, BNNSGraph.Builder.Tensor) -> BNNSGraph.Builder.Tensor`](/documentation/Accelerate/BNNSGraph/Builder/Tensor/._(_:_:)-171l2) Adds an element-wise logical or operation to the current graph. [`static func .< (BNNSGraph.Builder.Tensor, BNNSGraph.Builder.Tensor) -> BNNSGraph.Builder.Tensor`](/documentation/Accelerate/BNNSGraph/Builder/Tensor/._(_:_:)-6sgl6) Adds an element-wise less than comparison operation to the current graph. [`static func .^ (BNNSGraph.Builder.Tensor, BNNSGraph.Builder.Tensor) -> BNNSGraph.Builder.Tensor`](/documentation/Accelerate/BNNSGraph/Builder/Tensor/._(_:_:)-6soc4) Adds an element-wise logical xor operation to the current graph. [`static func .> (BNNSGraph.Builder.Tensor, BNNSGraph.Builder.Tensor) -> BNNSGraph.Builder.Tensor`](/documentation/Accelerate/BNNSGraph/Builder/Tensor/._(_:_:)-9vd51) Adds an element-wise greater than comparison operation to the current graph. [`static func .<= (BNNSGraph.Builder.Tensor, BNNSGraph.Builder.Tensor) -> BNNSGraph.Builder.Tensor`](/documentation/Accelerate/BNNSGraph/Builder/Tensor/._=(_:_:)-20vpj) Adds an element-wise less than or equal comparison operation to the current graph. [`static func .>= (BNNSGraph.Builder.Tensor, BNNSGraph.Builder.Tensor) -> BNNSGraph.Builder.Tensor`](/documentation/Accelerate/BNNSGraph/Builder/Tensor/._=(_:_:)-6gvvh) Adds an element-wise greater than or equal comparison operation to the current graph. ### Instance Properties [`var dataType: BNNSDataType?`](/documentation/Accelerate/BNNSGraph/Builder/Tensor/dataType) The data type of the tensor. [`var description: String`](/documentation/Accelerate/BNNSGraph/Builder/Tensor/description) A textual representation of this instance. [`var rank: Int?`](/documentation/Accelerate/BNNSGraph/Builder/Tensor/rank) The rank of the tensor. [`var shape: [Int]?`](/documentation/Accelerate/BNNSGraph/Builder/Tensor/shape) The shape of the tensor. [`var stride: [Int]?`](/documentation/Accelerate/BNNSGraph/Builder/Tensor/stride) The stride of the tensor. ### Instance Methods [`func abs() -> BNNSGraph.Builder.Tensor`](/documentation/Accelerate/BNNSGraph/Builder/Tensor/abs()) Adds an element-wise absolute operation to the current graph. [`func acos() -> BNNSGraph.Builder.Tensor`](/documentation/Accelerate/BNNSGraph/Builder/Tensor/acos()) Adds an element-wise acos operation to the current graph. [`func acosh() -> BNNSGraph.Builder.Tensor`](/documentation/Accelerate/BNNSGraph/Builder/Tensor/acosh()) Adds an element-wise atan operation to the current graph. [`func argMax(axis: Int, keepDimension: Bool) -> BNNSGraph.Builder.Tensor`](/documentation/Accelerate/BNNSGraph/Builder/Tensor/argMax(axis:keepDimension:)) Adds an argmax operation to the current graph. [`func argMin(axis: Int, keepDimension: Bool) -> BNNSGraph.Builder.Tensor`](/documentation/Accelerate/BNNSGraph/Builder/Tensor/argMin(axis:keepDimension:)) Adds an argmin operation to the current graph. [`func argSort(axis: Int, sortOrder: BNNSGraph.Builder.SortOrder) -> BNNSGraph.Builder.Tensor`](/documentation/Accelerate/BNNSGraph/Builder/Tensor/argSort(axis:sortOrder:)) Adds an argsort operation to the current graph. [`func asin() -> BNNSGraph.Builder.Tensor`](/documentation/Accelerate/BNNSGraph/Builder/Tensor/asin()) Adds an element-wise asin operation to the current graph. [`func asinh() -> BNNSGraph.Builder.Tensor`](/documentation/Accelerate/BNNSGraph/Builder/Tensor/asinh()) Adds an element-wise asinh operation to the current graph. [`func atan() -> BNNSGraph.Builder.Tensor`](/documentation/Accelerate/BNNSGraph/Builder/Tensor/atan()) Adds an element-wise atan operation to the current graph. [`func atanh() -> BNNSGraph.Builder.Tensor`](/documentation/Accelerate/BNNSGraph/Builder/Tensor/atanh()) Adds an element-wise atanh operation to the current graph. [`func batchNorm(mean: some BNNSGraph.Builder.OperationParameter, variance: some BNNSGraph.Builder.OperationParameter, epsilon: Float) -> BNNSGraph.Builder.Tensor`](/documentation/Accelerate/BNNSGraph/Builder/Tensor/batchNorm(mean:variance:epsilon:)) Adds a batch normalization operation to the current graph. [`func batchNorm(mean: some BNNSGraph.Builder.OperationParameter, variance: some BNNSGraph.Builder.OperationParameter, weight: some BNNSGraph.Builder.OperationParameter, bias: some BNNSGraph.Builder.OperationParameter, epsilon: Float) -> BNNSGraph.Builder.Tensor`](/documentation/Accelerate/BNNSGraph/Builder/Tensor/batchNorm(mean:variance:weight:bias:epsilon:)) Adds a batch normalization operation to the current graph. [`func bidirectionalLSTM(initialHiddenStates: BNNSGraph.Builder.Tensor, initialCellStates: BNNSGraph.Builder.Tensor, inputHiddenWeight: BNNSGraph.Builder.Tensor, hiddenHiddenWeight: BNNSGraph.Builder.Tensor, bias: BNNSGraph.Builder.Tensor, inputHiddenWeightBack: BNNSGraph.Builder.Tensor, hiddenHiddenWeightBack: BNNSGraph.Builder.Tensor, biasBack: BNNSGraph.Builder.Tensor, activation: BNNSGraph.Builder.Activation, recurrentActivation: BNNSGraph.Builder.Activation, cellActivation: BNNSGraph.Builder.Activation, outputSequence: Bool) -> (output: BNNSGraph.Builder.Tensor, hiddenStates: BNNSGraph.Builder.Tensor, memoryStates: BNNSGraph.Builder.Tensor)`](/documentation/Accelerate/BNNSGraph/Builder/Tensor/bidirectionalLSTM(initialHiddenStates:initialCellStates:inputHiddenWeight:hiddenHiddenWeight:bias:inputHiddenWeightBack:hiddenHiddenWeightBack:biasBack:activation:recurrentActivation:cellActivation:outputSequence:)) Adds a bidirectional LSTM operation to the current graph. [`func cast(to: U.Type) -> BNNSGraph.Builder.Tensor`](/documentation/Accelerate/BNNSGraph/Builder/Tensor/cast(to:)) Adds a cast operation to the current graph. [`func ceil() -> BNNSGraph.Builder.Tensor`](/documentation/Accelerate/BNNSGraph/Builder/Tensor/ceil()) Adds an element-wise ceiling operation to the current graph. [`func channelNorm(epsilon: Float) -> BNNSGraph.Builder.Tensor`](/documentation/Accelerate/BNNSGraph/Builder/Tensor/channelNorm(epsilon:)) Adds a channel normalization operation to the current graph. [`func channelNorm(weight: some BNNSGraph.Builder.OperationParameter, bias: some BNNSGraph.Builder.OperationParameter, epsilon: Float) -> BNNSGraph.Builder.Tensor`](/documentation/Accelerate/BNNSGraph/Builder/Tensor/channelNorm(weight:bias:epsilon:)) Adds a channel normalization operation to the current graph. [`func clampedReLU(alpha: Float, beta: Float) -> BNNSGraph.Builder.Tensor`](/documentation/Accelerate/BNNSGraph/Builder/Tensor/clampedReLU(alpha:beta:)) Adds a Clamped Rectified Linear Unit (ReLU) activation operation to the current graph. [`func clip(to: ClosedRange) -> BNNSGraph.Builder.Tensor`](/documentation/Accelerate/BNNSGraph/Builder/Tensor/clip(to:)) Adds an element-wise clip operation to the current graph. [`func conv(weight: some BNNSGraph.Builder.OperationParameter, strides: [Int], bias: some BNNSGraph.Builder.OperationParameter, padding: BNNSGraph.Builder.ConvolutionPadding, dilations: [Int]?, groupCount: Int) -> BNNSGraph.Builder.Tensor`](/documentation/Accelerate/BNNSGraph/Builder/Tensor/conv(weight:strides:bias:padding:dilations:groupCount:)) Adds a convolution operation to the current graph. [`func conv(weight: some BNNSGraph.Builder.OperationParameter, strides: [Int], padding: BNNSGraph.Builder.ConvolutionPadding, dilations: [Int]?, groupCount: Int) -> BNNSGraph.Builder.Tensor`](/documentation/Accelerate/BNNSGraph/Builder/Tensor/conv(weight:strides:padding:dilations:groupCount:)) Adds a convolution operation to the current graph. [`func convTranspose(weight: some BNNSGraph.Builder.OperationParameter, strides: [Int], bias: some BNNSGraph.Builder.OperationParameter, padding: BNNSGraph.Builder.ConvolutionPadding, outputPaddingValues: [Int]?, dilations: [Int]?, groupCount: Int) -> BNNSGraph.Builder.Tensor`](/documentation/Accelerate/BNNSGraph/Builder/Tensor/convTranspose(weight:strides:bias:padding:outputPaddingValues:dilations:groupCount:)) Adds a transposed convolution operation to the current graph. [`func convTranspose(weight: some BNNSGraph.Builder.OperationParameter, strides: [Int], padding: BNNSGraph.Builder.ConvolutionPadding, outputPaddingValues: [Int]?, dilations: [Int]?, groupCount: Int) -> BNNSGraph.Builder.Tensor`](/documentation/Accelerate/BNNSGraph/Builder/Tensor/convTranspose(weight:strides:padding:outputPaddingValues:dilations:groupCount:)) Adds a transposed convolution operation to the current graph. [`func cos() -> BNNSGraph.Builder.Tensor`](/documentation/Accelerate/BNNSGraph/Builder/Tensor/cos()) Adds an element-wise cos operation to the current graph. [`func cosh() -> BNNSGraph.Builder.Tensor`](/documentation/Accelerate/BNNSGraph/Builder/Tensor/cosh()) Adds an element-wise cosh operation to the current graph. [`func cumulativeSum(axis: Int) -> BNNSGraph.Builder.Tensor`](/documentation/Accelerate/BNNSGraph/Builder/Tensor/cumulativeSum(axis:)) Adds a cumulative sum operation to the the graph. [`func elu(alpha: Float) -> BNNSGraph.Builder.Tensor`](/documentation/Accelerate/BNNSGraph/Builder/Tensor/elu(alpha:)) Adds an Exponential Linear Unit (ELU) activation operation to the current graph. [`func erf() -> BNNSGraph.Builder.Tensor`](/documentation/Accelerate/BNNSGraph/Builder/Tensor/erf()) Adds an Error Function (erf) activation operation to the current graph. [`func exp() -> BNNSGraph.Builder.Tensor`](/documentation/Accelerate/BNNSGraph/Builder/Tensor/exp()) Adds an element-wise exp operation to the current graph. [`func exp2() -> BNNSGraph.Builder.Tensor`](/documentation/Accelerate/BNNSGraph/Builder/Tensor/exp2()) Adds an element-wise exp operation to the current graph. [`func floor() -> BNNSGraph.Builder.Tensor`](/documentation/Accelerate/BNNSGraph/Builder/Tensor/floor()) Adds an element-wise floor operation to the current graph. [`func fma(y: some BNNSGraph.Builder.OperationParameter, z: some BNNSGraph.Builder.OperationParameter) -> BNNSGraph.Builder.Tensor`](/documentation/Accelerate/BNNSGraph/Builder/Tensor/fma(y:z:)) Adds an element-wise fused multiply-add operation to the current graph. [`func gather(indices: some BNNSGraph.Builder.OperationParameter, axis: Int, batchDimensionCount: Int) -> BNNSGraph.Builder.Tensor`](/documentation/Accelerate/BNNSGraph/Builder/Tensor/gather(indices:axis:batchDimensionCount:)) Adds a gather operation to the current graph. [`func gatherAlongAxis(indices: some BNNSGraph.Builder.OperationParameter, axis: Int) -> BNNSGraph.Builder.Tensor`](/documentation/Accelerate/BNNSGraph/Builder/Tensor/gatherAlongAxis(indices:axis:)) Adds a gather-along-axis operation to the current graph. [`func gatherND(indices: some BNNSGraph.Builder.OperationParameter, batchDimensionCount: Int) -> BNNSGraph.Builder.Tensor`](/documentation/Accelerate/BNNSGraph/Builder/Tensor/gatherND(indices:batchDimensionCount:)) Adds a gather-nd operation to the current graph. [`func gelu() -> BNNSGraph.Builder.Tensor`](/documentation/Accelerate/BNNSGraph/Builder/Tensor/gelu()) Adds a Gaussian Error Linear Unit (GELU) activation operation to the current graph. [`func geluSigmoidApproximation() -> BNNSGraph.Builder.Tensor`](/documentation/Accelerate/BNNSGraph/Builder/Tensor/geluSigmoidApproximation()) Adds a Gaussian Error Linear Unit (GELU) sigmoid approximation activation operation to the current graph. [`func geluTanhApproximation() -> BNNSGraph.Builder.Tensor`](/documentation/Accelerate/BNNSGraph/Builder/Tensor/geluTanhApproximation()) Adds a Gaussian Error Linear Unit (GELU) tanh approximation activation operation to the current graph. [`func gru(initialHiddenStates: BNNSGraph.Builder.Tensor, inputHiddenWeight: BNNSGraph.Builder.Tensor, hiddenHiddenWeight: BNNSGraph.Builder.Tensor, bias: BNNSGraph.Builder.Tensor, inputBias: BNNSGraph.Builder.Tensor, direction: BNNSGraph.Builder.Direction, activation: BNNSGraph.Builder.Activation, recurrentActivation: BNNSGraph.Builder.Activation, applyResetGateAfterMatMul: Bool, outputSequence: Bool) -> (output: BNNSGraph.Builder.Tensor, hiddenStates: BNNSGraph.Builder.Tensor)`](/documentation/Accelerate/BNNSGraph/Builder/Tensor/gru(initialHiddenStates:inputHiddenWeight:hiddenHiddenWeight:bias:inputBias:direction:activation:recurrentActivation:applyResetGateAfterMatMul:outputSequence:)) Adds a GRU operation to the current graph. [`func hardSigmoid(alpha: Float, beta: Float) -> BNNSGraph.Builder.Tensor`](/documentation/Accelerate/BNNSGraph/Builder/Tensor/hardSigmoid(alpha:beta:)) Adds a hard sigmoid activation operation to the current graph. [`func hardSwish() -> BNNSGraph.Builder.Tensor`](/documentation/Accelerate/BNNSGraph/Builder/Tensor/hardSwish()) Adds a hard swish activation operation to the current graph. [`func instanceNorm(epsilon: Float) -> BNNSGraph.Builder.Tensor`](/documentation/Accelerate/BNNSGraph/Builder/Tensor/instanceNorm(epsilon:)) Adds a instance normalization operation to the current graph. [`func instanceNorm(weight: some BNNSGraph.Builder.OperationParameter, bias: some BNNSGraph.Builder.OperationParameter, epsilon: Float) -> BNNSGraph.Builder.Tensor`](/documentation/Accelerate/BNNSGraph/Builder/Tensor/instanceNorm(weight:bias:epsilon:)) Adds a instance normalization operation to the current graph. [`func l1Norm(axes: [Int], keepDimensions: Bool) -> BNNSGraph.Builder.Tensor`](/documentation/Accelerate/BNNSGraph/Builder/Tensor/l1Norm(axes:keepDimensions:)) Adds a sum-of-absolutes reduction operation along the given axis operation to the current graph. [`func l2Norm(axes: [Int], keepDimensions: Bool) -> BNNSGraph.Builder.Tensor`](/documentation/Accelerate/BNNSGraph/Builder/Tensor/l2Norm(axes:keepDimensions:)) Adds a Euclidean-norm reduction operation along the given axis operation to the current graph. [`func l2Norm(epsilon: Float) -> BNNSGraph.Builder.Tensor`](/documentation/Accelerate/BNNSGraph/Builder/Tensor/l2Norm(epsilon:)) Adds an L2 spatial normalization operation to the current graph. [`func layerNorm(axes: [Int], epsilon: Float) -> BNNSGraph.Builder.Tensor`](/documentation/Accelerate/BNNSGraph/Builder/Tensor/layerNorm(axes:epsilon:)) Adds a layer normalization operation to the current graph. [`func layerNorm(weight: some BNNSGraph.Builder.OperationParameter, bias: some BNNSGraph.Builder.OperationParameter, axes: [Int], epsilon: Float) -> BNNSGraph.Builder.Tensor`](/documentation/Accelerate/BNNSGraph/Builder/Tensor/layerNorm(weight:bias:axes:epsilon:)) Adds a layer normalization operation to the current graph. [`func leakyReLU(alpha: Float) -> BNNSGraph.Builder.Tensor`](/documentation/Accelerate/BNNSGraph/Builder/Tensor/leakyReLU(alpha:)) Adds a Leaky Rectified Linear Unit (ReLU) activation operation to the current graph. [`func linear(weight: some BNNSGraph.Builder.OperationParameter) -> BNNSGraph.Builder.Tensor`](/documentation/Accelerate/BNNSGraph/Builder/Tensor/linear(weight:)) Adds a linear transformation operation to the current graph. [`func linear(weight: some BNNSGraph.Builder.OperationParameter, bias: some BNNSGraph.Builder.OperationParameter) -> BNNSGraph.Builder.Tensor`](/documentation/Accelerate/BNNSGraph/Builder/Tensor/linear(weight:bias:)) Adds a linear transformation operation to the current graph. [`func log(epsilon: Float) -> BNNSGraph.Builder.Tensor`](/documentation/Accelerate/BNNSGraph/Builder/Tensor/log(epsilon:)) Adds an element-wise log operation to the current graph. [`func logSoftmax(axis: Int) -> BNNSGraph.Builder.Tensor`](/documentation/Accelerate/BNNSGraph/Builder/Tensor/logSoftmax(axis:)) Adds a log-softmax along the given axis operation to the current graph. [`func logSum(axes: [Int], keepDimensions: Bool) -> BNNSGraph.Builder.Tensor`](/documentation/Accelerate/BNNSGraph/Builder/Tensor/logSum(axes:keepDimensions:)) Adds a log-sum reduction operation along the given axis operation to the current graph. [`func logSumExp(axes: [Int], keepDimensions: Bool) -> BNNSGraph.Builder.Tensor`](/documentation/Accelerate/BNNSGraph/Builder/Tensor/logSumExp(axes:keepDimensions:)) Adds a log-sum-exp reduction operation along the given axis operation to the current graph. [`func lstm(initialHiddenStates: BNNSGraph.Builder.Tensor, initialCellStates: BNNSGraph.Builder.Tensor, inputHiddenWeight: BNNSGraph.Builder.Tensor, hiddenHiddenWeight: BNNSGraph.Builder.Tensor, bias: BNNSGraph.Builder.Tensor, direction: BNNSGraph.Builder.Direction, activation: BNNSGraph.Builder.Activation, recurrentActivation: BNNSGraph.Builder.Activation, cellActivation: BNNSGraph.Builder.Activation, outputSequence: Bool) -> (output: BNNSGraph.Builder.Tensor, hiddenStates: BNNSGraph.Builder.Tensor, memoryStates: BNNSGraph.Builder.Tensor)`](/documentation/Accelerate/BNNSGraph/Builder/Tensor/lstm(initialHiddenStates:initialCellStates:inputHiddenWeight:hiddenHiddenWeight:bias:direction:activation:recurrentActivation:cellActivation:outputSequence:)) Adds an LSTM operation to the current graph. [`func matmul(transpose: Bool, other: some BNNSGraph.Builder.OperationParameter, transposeOther: Bool) -> BNNSGraph.Builder.Tensor`](/documentation/Accelerate/BNNSGraph/Builder/Tensor/matmul(transpose:other:transposeOther:)) Adds a matrix-matrix multiplication operation to the current graph. [`func matmul(transpose: Bool, other: some BNNSGraph.Builder.OperationParameter, transposeOther: Bool, bias: some BNNSGraph.Builder.OperationParameter) -> BNNSGraph.Builder.Tensor`](/documentation/Accelerate/BNNSGraph/Builder/Tensor/matmul(transpose:other:transposeOther:bias:)) Adds a matrix-matrix multiplication operation to the current graph. [`func max(y: some BNNSGraph.Builder.OperationParameter) -> BNNSGraph.Builder.Tensor`](/documentation/Accelerate/BNNSGraph/Builder/Tensor/max(y:)) Adds an element-wise maximum operation to the current graph. [`func maximum(axes: [Int], keepDimensions: Bool) -> BNNSGraph.Builder.Tensor`](/documentation/Accelerate/BNNSGraph/Builder/Tensor/maximum(axes:keepDimensions:)) Adds a maximum reduction operation along the given axis operation to the current graph. [`func mean(axes: [Int], keepDimensions: Bool) -> BNNSGraph.Builder.Tensor`](/documentation/Accelerate/BNNSGraph/Builder/Tensor/mean(axes:keepDimensions:)) Adds a mean reduction operation along the given axis operation to the current graph. [`func min(y: some BNNSGraph.Builder.OperationParameter) -> BNNSGraph.Builder.Tensor`](/documentation/Accelerate/BNNSGraph/Builder/Tensor/min(y:)) Adds an element-wise minimum operation to the current graph. [`func minimum(axes: [Int], keepDimensions: Bool) -> BNNSGraph.Builder.Tensor`](/documentation/Accelerate/BNNSGraph/Builder/Tensor/minimum(axes:keepDimensions:)) Adds a minimum reduction operation along the given axis operation to the current graph. [`func pad(BNNSGraph.Builder.Padding, padding: [Int]) -> BNNSGraph.Builder.Tensor`](/documentation/Accelerate/BNNSGraph/Builder/Tensor/pad(_:padding:)) Adds a padding operation to the graph [`func pooling(BNNSGraph.Builder.PoolingFunction, kernelSize: [Int], strides: [Int], padding: BNNSGraph.Builder.PoolingPadding, ceilingMode: BNNSGraph.Builder.CeilingMode) -> BNNSGraph.Builder.Tensor`](/documentation/Accelerate/BNNSGraph/Builder/Tensor/pooling(_:kernelSize:strides:padding:ceilingMode:)) Adds a pooling operation to the current graph. [`func pow(y: some BNNSGraph.Builder.OperationParameter) -> BNNSGraph.Builder.Tensor`](/documentation/Accelerate/BNNSGraph/Builder/Tensor/pow(y:)) Adds an element-wise power operation to the current graph. [`func prelu(alpha: [Float]) -> BNNSGraph.Builder.Tensor`](/documentation/Accelerate/BNNSGraph/Builder/Tensor/prelu(alpha:)) Adds a Parametric ReLU (PReLU) activation operation to the current graph. [`func product(axes: [Int], keepDimensions: Bool) -> BNNSGraph.Builder.Tensor`](/documentation/Accelerate/BNNSGraph/Builder/Tensor/product(axes:keepDimensions:)) Adds a product reduction operation along the given axis operation to the current graph. [`func reciprocal(epsilon: Float) -> BNNSGraph.Builder.Tensor`](/documentation/Accelerate/BNNSGraph/Builder/Tensor/reciprocal(epsilon:)) Adds an element-wise reciprocal operation to the current graph. [`func relu() -> BNNSGraph.Builder.Tensor`](/documentation/Accelerate/BNNSGraph/Builder/Tensor/relu()) Adds a Rectified Linear Unit (ReLU) activation operation to the current graph. [`func relu6() -> BNNSGraph.Builder.Tensor`](/documentation/Accelerate/BNNSGraph/Builder/Tensor/relu6()) Adds a Rectified Linear Unit 6 (ReLU6) activation operation to the current graph. [`func reshape(to: some BNNSGraph.Builder.OperationParameter) -> BNNSGraph.Builder.Tensor`](/documentation/Accelerate/BNNSGraph/Builder/Tensor/reshape(to:)-5oniw) Adds a dynamic reshape operation to the current graph. [`func reshape(to: [Int]) -> BNNSGraph.Builder.Tensor`](/documentation/Accelerate/BNNSGraph/Builder/Tensor/reshape(to:)-9stle) Adds a reshape operation to the current graph. [`func rmsNorm(scale: some BNNSGraph.Builder.OperationParameter, epsilon: Float) -> BNNSGraph.Builder.Tensor`](/documentation/Accelerate/BNNSGraph/Builder/Tensor/rmsNorm(scale:epsilon:)) Adds an RMS spatial normalization operation to the current graph. [`func rnn(initialHiddenStates: BNNSGraph.Builder.Tensor, inputHiddenWeight: BNNSGraph.Builder.Tensor, hiddenHiddenWeight: BNNSGraph.Builder.Tensor, bias: BNNSGraph.Builder.Tensor, direction: BNNSGraph.Builder.Direction, activation: BNNSGraph.Builder.Activation, outputSequence: Bool) -> (output: BNNSGraph.Builder.Tensor, hiddenStates: BNNSGraph.Builder.Tensor)`](/documentation/Accelerate/BNNSGraph/Builder/Tensor/rnn(initialHiddenStates:inputHiddenWeight:hiddenHiddenWeight:bias:direction:activation:outputSequence:)) Adds an RNN operation to the current graph. [`func round() -> BNNSGraph.Builder.Tensor`](/documentation/Accelerate/BNNSGraph/Builder/Tensor/round()) Adds an element-wise round operation to the current graph. [`func rsqrt(epsilon: Float) -> BNNSGraph.Builder.Tensor`](/documentation/Accelerate/BNNSGraph/Builder/Tensor/rsqrt(epsilon:)) Adds an element-wise rsqrt operation to the current graph. [`func scaledTanh(alpha: Float, beta: Float) -> BNNSGraph.Builder.Tensor`](/documentation/Accelerate/BNNSGraph/Builder/Tensor/scaledTanh(alpha:beta:)) Adds a scaled tanh activation operation to the current graph. [`func scatter(updates: BNNSGraph.Builder.Tensor, indices: some BNNSGraph.Builder.OperationParameter, mode: BNNSGraph.Builder.ScatterMode, axis: Int) -> BNNSGraph.Builder.Tensor`](/documentation/Accelerate/BNNSGraph/Builder/Tensor/scatter(updates:indices:mode:axis:)) Adds a scatter operation to the current graph. [`func scatterAlongAxis(updates: BNNSGraph.Builder.Tensor, indices: some BNNSGraph.Builder.OperationParameter, mode: BNNSGraph.Builder.ScatterMode, axis: Int) -> BNNSGraph.Builder.Tensor`](/documentation/Accelerate/BNNSGraph/Builder/Tensor/scatterAlongAxis(updates:indices:mode:axis:)) Adds a scatter-along-axis operation to the current graph [`func scatterND(updates: BNNSGraph.Builder.Tensor, indices: some BNNSGraph.Builder.OperationParameter, mode: BNNSGraph.Builder.ScatterMode) -> BNNSGraph.Builder.Tensor`](/documentation/Accelerate/BNNSGraph/Builder/Tensor/scatterND(updates:indices:mode:)) Adds a scatter-nd operation to the current graph. [`func select(BNNSGraph.Builder.Tensor, BNNSGraph.Builder.Tensor) -> BNNSGraph.Builder.Tensor`](/documentation/Accelerate/BNNSGraph/Builder/Tensor/select(_:_:)-1fiqe) Adds a tensor-tensor select operation to the current graph. [`func select(BNNSGraph.Builder.Tensor, any BNNSGraph.Builder.OperationParameter) -> BNNSGraph.Builder.Tensor`](/documentation/Accelerate/BNNSGraph/Builder/Tensor/select(_:_:)-3ha1f) Adds a tensor-scalar select operation to the current graph. [`func select(any BNNSGraph.Builder.OperationParameter, BNNSGraph.Builder.Tensor) -> BNNSGraph.Builder.Tensor`](/documentation/Accelerate/BNNSGraph/Builder/Tensor/select(_:_:)-77itm) Adds a scalar-tensor select operation to the current graph. [`func sigmoid() -> BNNSGraph.Builder.Tensor`](/documentation/Accelerate/BNNSGraph/Builder/Tensor/sigmoid()) Adds a sigmoid activation operation to the current graph. [`func silu() -> BNNSGraph.Builder.Tensor`](/documentation/Accelerate/BNNSGraph/Builder/Tensor/silu()) Adds a Sigmoid Linear Unit (SiLU) activation operation to the current graph. [`func sin() -> BNNSGraph.Builder.Tensor`](/documentation/Accelerate/BNNSGraph/Builder/Tensor/sin()) Adds an element-wise sin operation to the current graph. [`func sinh() -> BNNSGraph.Builder.Tensor`](/documentation/Accelerate/BNNSGraph/Builder/Tensor/sinh()) Adds an element-wise sinh operation to the current graph. [`func softmax(axis: Int) -> BNNSGraph.Builder.Tensor`](/documentation/Accelerate/BNNSGraph/Builder/Tensor/softmax(axis:)) Adds a softmax along the given axis operation to the current graph. [`func softplus(alpha: Float) -> BNNSGraph.Builder.Tensor`](/documentation/Accelerate/BNNSGraph/Builder/Tensor/softplus(alpha:)) Adds a softplus activation operation to the current graph. [`func softsign() -> BNNSGraph.Builder.Tensor`](/documentation/Accelerate/BNNSGraph/Builder/Tensor/softsign()) Adds a softsign activation operation to the current graph. [`func sqrt() -> BNNSGraph.Builder.Tensor`](/documentation/Accelerate/BNNSGraph/Builder/Tensor/sqrt()) Adds an element-wise sqrt operation to the current graph. [`func squeeze(axis: Int) -> BNNSGraph.Builder.Tensor`](/documentation/Accelerate/BNNSGraph/Builder/Tensor/squeeze(axis:)) Adds a squeeze operation in the graph. [`func sum(axes: [Int], keepDimensions: Bool) -> BNNSGraph.Builder.Tensor`](/documentation/Accelerate/BNNSGraph/Builder/Tensor/sum(axes:keepDimensions:)) Adds a sum reduction operation along the given axis operation to the current graph. [`func sumOfSquares(axes: [Int], keepDimensions: Bool) -> BNNSGraph.Builder.Tensor`](/documentation/Accelerate/BNNSGraph/Builder/Tensor/sumOfSquares(axes:keepDimensions:)) Adds a sum-of-squares reduction operation along the given axis operation to the current graph. [`func tan() -> BNNSGraph.Builder.Tensor`](/documentation/Accelerate/BNNSGraph/Builder/Tensor/tan()) Adds an element-wise tan operation to the current graph. [`func tanh() -> BNNSGraph.Builder.Tensor`](/documentation/Accelerate/BNNSGraph/Builder/Tensor/tanh()) Adds an element-wise hyperbolic tangent operation to the current graph. [`func tensorShape() -> BNNSGraph.Builder.Tensor`](/documentation/Accelerate/BNNSGraph/Builder/Tensor/tensorShape()) Adds a shape operation to the current graph. [`func threshold(to: Float) -> BNNSGraph.Builder.Tensor`](/documentation/Accelerate/BNNSGraph/Builder/Tensor/threshold(to:)) Adds an element-wise round operation to the current graph. [`func thresholdedReLU(alpha: Float) -> BNNSGraph.Builder.Tensor`](/documentation/Accelerate/BNNSGraph/Builder/Tensor/thresholdedReLU(alpha:)) Adds a Thresholded Leaky Rectified Linear Unit (ReLU) activation operation to the current graph. [`func topK(Int, axis: Int, findLargest: Bool) -> (values: BNNSGraph.Builder.Tensor, indices: BNNSGraph.Builder.Tensor)`](/documentation/Accelerate/BNNSGraph/Builder/Tensor/topK(_:axis:findLargest:)) Adds a top-k operation to the current graph. [`func transpose(axes: [Int]) -> BNNSGraph.Builder.Tensor`](/documentation/Accelerate/BNNSGraph/Builder/Tensor/transpose(axes:)) Adds a transpose operation to the current graph. [`func unsqueeze(axis: Int) -> BNNSGraph.Builder.Tensor`](/documentation/Accelerate/BNNSGraph/Builder/Tensor/unsqueeze(axis:)) Adds an unsqueeze operation in the graph. ### Subscripts [`subscript(any BNNSGraph.Builder.SliceIndex...) -> BNNSGraph.Builder.Tensor`](/documentation/Accelerate/BNNSGraph/Builder/Tensor/subscript(_:)) Adds a slice operation to the current graph. ### Default Implementations [OperationParameter Implementations](/documentation/Accelerate/BNNSGraph/Builder/Tensor/OperationParameter-Implementations) ## Relationships ### Conforms To [`Escapable`](/documentation/Swift/Escapable) [`TensorDescriptor`](/documentation/Accelerate/BNNSGraph/TensorDescriptor) [`Copyable`](/documentation/Swift/Copyable) [`CustomStringConvertible`](/documentation/Swift/CustomStringConvertible) [`OperationParameter`](/documentation/Accelerate/BNNSGraph/Builder/OperationParameter) --- Copyright © 2026 Apple Inc. All rights reserved. | [Terms of Use](https://www.apple.com/legal/internet-services/terms/site.html) | [Privacy Policy](https://www.apple.com/privacy/privacy-policy)