destack.sh
menu

destack:tensor

import destack:tensor

Structs

  1. AxisSetAxis set used by reductions and broadcasts.
  2. ConvolutionAxesAxis mapping for tensor convolution.
  3. ConvolutionGroupsGrouping parameters for tensor convolution.
  4. ConvolutionWindowWindow parameters for tensor convolution.
  5. CoordinateTensor coordinate with Rank indices.
  6. DotAxesAxis mapping for tensor dot.
  7. FactorizationOptionsOptions for matrix factorizations.
  8. GatherAxesAxis mapping for tensor gather.
  9. NamedAxisNamed axis metadata used by frontends and debugging tools.
  10. NormalizationOptionsOptions for normalization operations.
  11. PaddingPadding for a tensor with Rank axes.
  12. PermutationAxis permutation with Rank entries.
  13. PoolOptionsWindow options for pooling.
  14. PoolWindowWindow parameters for pooling.
  15. ScatterAxesAxis mapping for tensor scatter.
  16. ShapeTensor shape with Rank axes.
  17. SliceWindowSlice window for a tensor with Rank axes.
  18. SoftmaxOptionsOptions for softmax-like operations.
  19. TensorConvertOptionsOutput options for tensor conversion.
  20. TensorConvolutionOptionsOutput options for tensor convolution.
  21. TensorCreateOptionsAllocation options for a new tensor.
  22. TensorExecutionOptionsExecution options for tensor work.
  23. TensorOutputOptionsOutput options for tensor operations.
  24. TensorReduceOptionsOutput options for tensor reductions.
  25. TensorScatterOptionsOutput options for tensor scatter.

Enums

  1. ConvertModeConversion mode for tensor element conversion.
  2. FactorizationModeMatrix factorization mode.
  3. MemoryOrderMemory order for dense tensor storage.
  4. ReductionReduction used when partial tensor shards are combined.
  5. ScatterModeUpdate mode for tensor scatter.

Newtypes

  1. AxisAxis index inside a tensor rank.
  2. DenseDense contiguous tensor format.
  3. ExtentRuntime size along one tensor axis.
  4. PartialStore partial results across one mesh axis.
  5. ReplicateReplicate one tensor axis across the mesh.
  6. ShardSplit one tensor axis across one mesh axis.
  7. ShardingTensor storage partitioned or replicated across a compute mesh.
  8. StrideTensor element stride.
  9. StridedExplicit strided tensor view format.
  10. TensorOwning tensor value.
  11. TensorViewBorrowed view over tensor storage.
  12. UnshardedTensor storage on one device.

Newtype interfaces

  1. PlacementTensor placement descriptor.
  2. ShardingAxisPer-axis placement descriptor used by a sharded tensor.
  3. TensorFormatFormat implemented by owning tensor storage.
  4. TensorViewFormatFormat implemented by borrowed tensor views.

Type aliases

  1. ColumnMajorFormatColumn-major dense tensor format.
  2. RowMajorFormatRow-major dense tensor format.

Functions

  1. absoluteReturn elementwise absolute value.
  2. allReturn whether all selected boolean values are true.
  3. anyReturn whether any selected boolean value is true.
  4. argmaxReturn indices of maximum values along axis.
  5. argminReturn indices of minimum values along axis.
  6. avgPoolReturn average-pooled values.
  7. bernoulliReturn values sampled from Bernoulli trials.
  8. bitAndReturn bitwise AND along the axes in axisSet.
  9. bitOrReturn bitwise OR along the axes in axisSet.
  10. bitXorReturn bitwise XOR along the axes in axisSet.
  11. broadcastBroadcast tensor to shape.
  12. castCast tensor elements to U by bit reinterpretation.
  13. choleskyReturn the Cholesky factorization.
  14. concatConcatenate tensors along axis.
  15. convertConvert tensor elements to U.
  16. convolutionConvolve input with kernel.
  17. cosReturn elementwise cosine.
  18. detReturn the determinant of a matrix.
  19. diagonalReturn the diagonal of a matrix.
  20. dotMultiply and reduce two tensors over configured axes.
  21. equalReturn elementwise equality.
  22. expReturn elementwise exponential.
  23. expandDimsInsert a length-one axis.
  24. fillFill view with value.
  25. gatherGather slices from operand.
  26. geluReturn Gaussian error linear activation.
  27. greaterReturn elementwise greater-than comparison.
  28. greaterEqualReturn elementwise greater-than-or-equal comparison.
  29. innerReturn the inner product of two vectors.
  30. inverseReturn the inverse of a matrix.
  31. layerNormReturn layer-normalized values.
  32. lessReturn elementwise less-than comparison.
  33. lessEqualReturn elementwise less-than-or-equal comparison.
  34. loadLoad one tensor element from view.
  35. logReturn elementwise natural logarithm.
  36. logSoftmaxReturn log-softmax over one axis.
  37. matmulMatrix multiply rank-two tensors.
  38. maxReturn maximum values of tensor along the axes in axisSet.
  39. maxPoolReturn max-pooled values.
  40. meanReturn the mean of tensor along the axes in axisSet.
  41. minReturn minimum values of tensor along the axes in axisSet.
  42. negativeReturn elementwise negation.
  43. normReturn vector or matrix norm along axes.
  44. normalReturn values sampled from the standard normal distribution.
  45. notEqualReturn elementwise inequality.
  46. outerReturn the outer product of two vectors.
  47. padPad tensor.
  48. productMultiply elements of tensor along the axes in axisSet.
  49. qrReturn the QR factorization.
  50. reformatCopy tensor data into a new format.
  51. reluReturn rectified linear activation.
  52. reshapeReshape tensor to shape.
  53. scatterScatter updates into operand.
  54. selectSelect tensor elements from thenValue or elseValue.
  55. sigmoidReturn logistic sigmoid activation.
  56. sinReturn elementwise sine.
  57. sliceSlice tensor by window.
  58. softmaxReturn softmax over one axis.
  59. solveSolve a linear system.
  60. splatBuild a tensor by repeating value.
  61. sqrtReturn elementwise square root.
  62. squeezeRemove length-one axes.
  63. stackStack tensors along a new axis.
  64. storeStore one tensor element into view.
  65. sumSum tensor along the axes in axisSet.
  66. svdReturn the singular value decomposition.
  67. tanReturn elementwise tangent.
  68. traceReturn the trace of a matrix.
  69. transposeTranspose tensor by permutation.
  70. uniformReturn values sampled uniformly from [0, 1).
  71. uniformRangeReturn values sampled uniformly from [low, high).