Module ops

Operations

Modules

• indexing
  Indexing Arrays

Enums

• PadMode
  The padding mode.
• PadWidth
  Number of padding values to add to the edges of each axis.
• TensorDotDims
  The number of dimensions to sum over

Traits

• ClipBound
  A custom trait for the bound of the clip operation.

Functions

• abs
  Element-wise absolute value.
• abs_device
  Element-wise absolute value.
• acos
  Element-wise inverse cosine.
• acos_device
  Element-wise inverse cosine.
• acosh
  Element-wise inverse hyperbolic cosine.
• acosh_device
  Element-wise inverse hyperbolic cosine.
• add
  See Array::add.
• add_device
  See Array::add.
• addmm
  Matrix multiplication with addition and optional scaling.
• addmm_device
  Matrix multiplication with addition and optional scaling.
• all
  See Array::all
• all_close
  See Array::all_close
• all_close_device
  See Array::all_close
• all_device
  See Array::all
• any
  See Array::any
• any_device
  See Array::any
• arange
  See Array::arange
• arange_device
  See Array::arange
• argpartition
  Returns the indices that partition the array. Returns an error if the arguments are invalid.
• argpartition_all
  Returns the indices that partition the flattened array. Returns an error if the arguments are invalid.
• argpartition_all_device
  Returns the indices that partition the flattened array. Returns an error if the arguments are invalid.
• argpartition_device
  Returns the indices that partition the array. Returns an error if the arguments are invalid.
• argsort
  Returns the indices that sort the array. Returns an error if the arguments are invalid.
• argsort_all
  Returns the indices that sort the flattened array. Returns an error if the arguments are invalid.
• argsort_all_device
  Returns the indices that sort the flattened array. Returns an error if the arguments are invalid.
• argsort_device
  Returns the indices that sort the array. Returns an error if the arguments are invalid.
• array_eq
  See Array::array_eq
• array_eq_device
  See Array::array_eq
• as_strided
  Create a view into the array with the given shape and strides.
• as_strided_device
  Create a view into the array with the given shape and strides.
• asin
  Element-wise inverse sine.
• asin_device
  Element-wise inverse sine.
• asinh
  Element-wise inverse hyperbolic sine.
• asinh_device
  Element-wise inverse hyperbolic sine.
• at_least_1d
  Convert array to have at least one dimension.
• at_least_1d_device
  Convert array to have at least one dimension.
• at_least_2d
  Convert array to have at least two dimensions.
• at_least_2d_device
  Convert array to have at least two dimensions.
• at_least_3d
  Convert array to have at least three dimensions.
• at_least_3d_device
  Convert array to have at least three dimensions.
• atan
  Element-wise inverse tangent.
• atan_device
  Element-wise inverse tangent.
• atanh
  Element-wise inverse hyperbolic tangent.
• atanh_device
  Element-wise inverse hyperbolic tangent.
• block_masked_mm
  Matrix multiplication with block masking.
• block_masked_mm_device
  Matrix multiplication with block masking.
• broadcast_arrays
  Broadcast a vector of arrays against one another. Returns an error if the shapes are broadcastable.
• broadcast_arrays_device
  Broadcast a vector of arrays against one another. Returns an error if the shapes are broadcastable.
• broadcast_to
  Broadcast an array to the given shape. Returns an error if the shapes are not broadcastable.
• broadcast_to_device
  Broadcast an array to the given shape. Returns an error if the shapes are not broadcastable.
• ceil
  Element-wise ceiling.
• ceil_device
  Element-wise ceiling.
• clip
  Clip the values of the array between the given minimum and maximum.
• clip_device
  Clip the values of the array between the given minimum and maximum.
• concatenate
  Concatenate the arrays along the given axis. Returns an error if the shapes are invalid.
• concatenate_device
  Concatenate the arrays along the given axis. Returns an error if the shapes are invalid.
• conv1d
  1D convolution over an input with several channels returning an error if the inputs are invalid.
• conv1d_device
  1D convolution over an input with several channels returning an error if the inputs are invalid.
• conv2d
  2D convolution over an input with several channels returning an error if the inputs are invalid.
• conv2d_device
  2D convolution over an input with several channels returning an error if the inputs are invalid.
• conv3d
  3D convolution over an input with several channels.
• conv3d_device
  3D convolution over an input with several channels.
• conv_general
  General convolution over an input with several channels returning an error if the inputs are invalid.
• conv_general_device
  General convolution over an input with several channels returning an error if the inputs are invalid.
• conv_transpose1d
  1D transposed convolution over an input with several channels.
• conv_transpose1d_device
  1D transposed convolution over an input with several channels.
• conv_transpose2d
  2D transposed convolution over an input with several channels.
• conv_transpose2d_device
  2D transposed convolution over an input with several channels.
• conv_transpose3d
  3D transposed convolution over an input with several channels.
• conv_transpose3d_device
  3D transposed convolution over an input with several channels.
• cos
  Element-wise cosine.
• cos_device
  Element-wise cosine.
• cosh
  Element-wise hyperbolic cosine.
• cosh_device
  Element-wise hyperbolic cosine.
• cummax
  See Array::cummax
• cummax_device
  See Array::cummax
• cummin
  See Array::cummin
• cummin_device
  See Array::cummin
• cumprod
  See Array::cumprod
• cumprod_device
  See Array::cumprod
• cumsum
  See Array::cumsum
• cumsum_device
  See Array::cumsum
• degrees
  Convert angles from radians to degrees.
• degrees_device
  Convert angles from radians to degrees.
• dequantize
  Dequantize the matrix w using the provided scales and biases and the group_size and bits configuration.
• dequantize_device
  Dequantize the matrix w using the provided scales and biases and the group_size and bits configuration.
• diag
  See Array::diag
• diag_device
  See Array::diag
• diagonal
  See Array::diagonal
• diagonal_device
  See Array::diagonal
• divide
  See Array::divide.
• divide_device
  See Array::divide.
• divmod
  Element-wise quotient and remainder.
• divmod_device
  Element-wise quotient and remainder.
• einsum
  Perform the Einstein summation convention on the operands.
• einsum_device
  Perform the Einstein summation convention on the operands.
• eq
  See Array::eq
• eq_device
  See Array::eq
• erf
  Element-wise error function.
• erf_device
  Element-wise error function.
• erfinv
  Element-wise inverse error function.
• erfinv_device
  Element-wise inverse error function.
• exp
  See Array::exp.
• exp_device
  See Array::exp.
• expand_dims
  Add a size one dimension at the given axis, returns an error if the axes are invalid.
• expand_dims_device
  Add a size one dimension at the given axis, returns an error if the axes are invalid.
• expm1
  Element-wise exponential minus 1.
• expm1_device
  Element-wise exponential minus 1.
• eye
  See Array::eye
• eye_device
  See Array::eye
• flatten
  Flatten an array. Returns an error if the axes are invalid.
• flatten_device
  Flatten an array. Returns an error if the axes are invalid.
• floor
  See Array::floor.
• floor_device
  See Array::floor.
• floor_divide
  See Array::floor_divide.
• floor_divide_device
  See Array::floor_divide.
• full
  See Array::full
• full_device
  See Array::full
• ge
  See Array::ge
• ge_device
  See Array::ge
• gt
  See Array::gt
• gt_device
  See Array::gt
• identity
  See Array::identity
• identity_device
  See Array::identity
• inner
  Ordinary inner product of vectors for 1-D arrays, in higher dimensions a sum product over the last axes.
• inner_device
  Ordinary inner product of vectors for 1-D arrays, in higher dimensions a sum product over the last axes.
• is_close
  See Array::is_close
• is_close_device
  See Array::is_close
• is_inf
  Return a boolean array indicating which elements are +/- inifnity.
• is_inf_device
  Return a boolean array indicating which elements are +/- inifnity.
• is_nan
  Return a boolean array indicating which elements are NaN.
• is_nan_device
  Return a boolean array indicating which elements are NaN.
• is_neg_inf
  Return a boolean array indicating which elements are negative infinity.
• is_neg_inf_device
  Return a boolean array indicating which elements are negative infinity.
• is_pos_inf
  Return a boolean array indicating which elements are positive infinity.
• is_pos_inf_device
  Return a boolean array indicating which elements are positive infinity.
• le
  See Array::le
• le_device
  See Array::le
• linspace
  See Array::linspace
• linspace_device
  See Array::linspace
• log
  See Array::log.
• log2
  See Array::log2.
• log1p
  See Array::log1p.
• log1p_device
  See Array::log1p.
• log2_device
  See Array::log2.
• log10
  See Array::log10.
• log10_device
  See Array::log10.
• log_add_exp
  Element-wise log-add-exp.
• log_add_exp_device
  Element-wise log-add-exp.
• log_device
  See Array::log.
• log_sum_exp
  See Array::log_sum_exp
• log_sum_exp_device
  See Array::log_sum_exp
• logical_and
  See Array::logical_and
• logical_and_device
  See Array::logical_and
• logical_not
  See Array::logical_not
• logical_not_device
  See Array::logical_not
• logical_or
  See Array::logical_or
• logical_or_device
  See Array::logical_or
• lt
  See Array::lt
• lt_device
  See Array::lt
• matmul
  See Array::matmul.
• matmul_device
  See Array::matmul.
• max
  See Array::max
• max_device
  See Array::max
• maximum
  Element-wise maximum.
• maximum_device
  Element-wise maximum.
• mean
  See Array::mean
• mean_device
  See Array::mean
• min
  See Array::min
• min_device
  See Array::min
• minimum
  Element-wise minimum.
• minimum_device
  Element-wise minimum.
• move_axis
  Move an axis to a new position. Returns an error if the axes are invalid.
• move_axis_device
  Move an axis to a new position. Returns an error if the axes are invalid.
• multiply
  See Array::multiply.
• multiply_device
  See Array::multiply.
• ne
  See Array::ne
• ne_device
  See Array::ne
• negative
  See Array::negative.
• negative_device
  See Array::negative.
• ones
  See Array::ones
• ones_device
  See Array::ones
• ones_dtype_device
  Similar to Array::ones but with a specified dtype.
• ones_like
  An array of ones like the input.
• ones_like_device
  An array of ones like the input.
• outer
  Compute the outer product of two 1-D arrays, if the array’s passed are not 1-D a flatten op will be run beforehand.
• outer_device
  Compute the outer product of two 1-D arrays, if the array’s passed are not 1-D a flatten op will be run beforehand.
• pad
  Pad an array with a constant value. Returns an error if the width is invalid.
• pad_device
  Pad an array with a constant value. Returns an error if the width is invalid.
• partition
  Returns a partitioned copy of the array such that the smaller kth elements are first. Returns an error if the arguments are invalid.
• partition_all
  Returns a partitioned copy of the flattened array such that the smaller kth elements are first. Returns an error if the arguments are invalid.
• partition_all_device
  Returns a partitioned copy of the flattened array such that the smaller kth elements are first. Returns an error if the arguments are invalid.
• partition_device
  Returns a partitioned copy of the array such that the smaller kth elements are first. Returns an error if the arguments are invalid.
• power
  See Array::power.
• power_device
  See Array::power.
• prod
  See Array::prod
• prod_device
  See Array::prod
• quantize
  Quantize the matrix w using bits bits per element.
• quantize_device
  Quantize the matrix w using bits bits per element.
• quantized_matmul
  Perform the matrix multiplication with the quantized matrix w. The quantization uses one floating point scale and bias per group_size of elements. Each element in w takes bits bits and is packed in an unsigned 32 bit integer.
• quantized_matmul_device
  Perform the matrix multiplication with the quantized matrix w. The quantization uses one floating point scale and bias per group_size of elements. Each element in w takes bits bits and is packed in an unsigned 32 bit integer.
• radians
  Convert angles from degrees to radians.
• radians_device
  Convert angles from degrees to radians.
• reciprocal
  See Array::reciprocal.
• reciprocal_device
  See Array::reciprocal.
• remainder
  See Array::remainder.
• remainder_device
  See Array::remainder.
• repeat
  See Array::repeat
• repeat_all
  See Array::repeat_all
• repeat_all_device
  See Array::repeat_all
• repeat_device
  See Array::repeat
• reshape
  Reshape an array while preserving the size. Returns an error if the new shape is invalid.
• reshape_device
  Reshape an array while preserving the size. Returns an error if the new shape is invalid.
• round
  See Array::round.
• round_device
  See Array::round.
• rsqrt
  See Array::rsqrt.
• rsqrt_device
  See Array::rsqrt.
• sigmoid
  Element-wise logistic sigmoid.
• sigmoid_device
  Element-wise logistic sigmoid.
• sign
  Element-wise sign.
• sign_device
  Element-wise sign.
• sin
  See Array::sin.
• sin_device
  See Array::sin.
• sinh
  Element-wise hyperbolic sine.
• sinh_device
  Element-wise hyperbolic sine.
• softmax
  Perform the softmax along the given axis.
• softmax_all
  Perform the softmax along all axes.
• softmax_all_device
  Perform the softmax along all axes.
• softmax_device
  Perform the softmax along the given axis.
• sort
  Returns a sorted copy of the array. Returns an error if the arguments are invalid.
• sort_all
  Returns a sorted copy of the flattened array. Returns an error if the arguments are invalid.
• sort_all_device
  Returns a sorted copy of the flattened array. Returns an error if the arguments are invalid.
• sort_device
  Returns a sorted copy of the array. Returns an error if the arguments are invalid.
• split
  Split an array along a given axis. Returns an error if the indices are invalid.
• split_device
  Split an array along a given axis. Returns an error if the indices are invalid.
• split_equal
  Split an array into equal parts along a given axis. Returns an error if the array cannot be split into equal parts.
• split_equal_device
  Split an array into equal parts along a given axis. Returns an error if the array cannot be split into equal parts.
• sqrt
  See Array::sqrt.
• sqrt_device
  See Array::sqrt.
• square
  See Array::square.
• square_device
  See Array::square.
• squeeze
  Remove length one axes from an array. Returns an error if the axes are invalid.
• squeeze_device
  Remove length one axes from an array. Returns an error if the axes are invalid.
• stack
  Stacks the arrays along a new axis. Returns an error if the arguments are invalid.
• stack_all
  Stacks the arrays along a new axis. Returns an error if the arrays have different shapes.
• stack_all_device
  Stacks the arrays along a new axis. Returns an error if the arrays have different shapes.
• stack_device
  Stacks the arrays along a new axis. Returns an error if the arguments are invalid.
• std
  Compute the standard deviation(s) over the given axes.
• std_device
  Compute the standard deviation(s) over the given axes.
• subtract
  See Array::subtract.
• subtract_device
  See Array::subtract.
• sum
  See Array::sum
• sum_device
  See Array::sum
• swap_axes
  Swap two axes of an array. Returns an error if the axes are invalid.
• swap_axes_device
  Swap two axes of an array. Returns an error if the axes are invalid.
• tan
  See [Array::tan].
• tan_device
  See [Array::tan].
• tanh
  Element-wise hyperbolic tangent.
• tanh_device
  Element-wise hyperbolic tangent.
• tensordot
  Compute the tensor dot product along the specified axes.
• tensordot_device
  Compute the tensor dot product along the specified axes.
• tile
  Construct an array by repeating a the number of times given by reps.
• tile_device
  Construct an array by repeating a the number of times given by reps.
• transpose
  Transpose the dimensions of the array. Returns an error if the axes are invalid.
• transpose_all
  Transpose with all axes reversed
• transpose_all_device
  Transpose with all axes reversed
• transpose_device
  Transpose the dimensions of the array. Returns an error if the axes are invalid.
• tri
  See Array::tri
• tri_device
  See Array::tri
• variance
  See Array::variance
• variance_device
  See Array::variance
• where
  Select from a or b according to condition returning an error if the arrays are not broadcastable.
• where_device
  Select from a or b according to condition returning an error if the arrays are not broadcastable.
• which
  Alias for [r#where]
• which_device
  Alias for [r#where]
• zeros
  See Array::zeros
• zeros_device
  See Array::zeros
• zeros_dtype
  Similar to Array::zeros but with a specified dtype.
• zeros_dtype_device
  Similar to Array::zeros but with a specified dtype.
• zeros_like
  An array of zeros like the input.
• zeros_like_device
  An array of zeros like the input.