ops

Element-wise negation.

Element-wise sign function.

Element-wise natural exponential.

Element-wise natural logarithm.

Element-wise base-2 logarithm.

Element-wise square root.

Element-wise square.

Element-wise multiplicative inverse.

Element-wise floor (round toward $-\infty$).

Element-wise ceiling (round toward $+\infty$).

Element-wise sine (input in radians).

Element-wise cosine (input in radians).

Element-wise tangent (input in radians).

Element-wise inverse sine in radians.

Element-wise inverse cosine in radians.

Element-wise inverse tangent in radians.

Element-wise hyperbolic sine.

Element-wise hyperbolic cosine.

Element-wise hyperbolic tangent.

Element-wise rectified linear unit.

Element-wise logistic sigmoid.

Return a tensor with C-contiguous (row-major) memory layout.

Element-wise addition with broadcasting and dtype promotion.

Element-wise subtraction with broadcasting and dtype promotion.

Element-wise multiplication (Hadamard product) with broadcasting.

Element-wise true division with broadcasting and dtype promotion.

Element-wise maximum with broadcasting.

Element-wise minimum with broadcasting.

Matrix product with full broadcasting semantics.

Return the cumulative sum of elements along `dim`.

Return the cumulative product of elements along `dim`.

Return the sum of the diagonal entries of a 2-D tensor.

Reshape a tensor to `shape`, returning a view when possible.

Insert a new size-1 dimension at position `dim`.

Collapse a contiguous range of dimensions into a single dim.

Return a view with dimensions rearranged.

Swap the last two dimensions of `input`.

Broadcast a tensor to a target shape as a *view*.

Return a view with singleton dimensions broadcast to new sizes.

Construct a tensor by repeating `input` `reps` times per axis.

Reverse the order of elements along the specified `dims`.

Flip a 2-D-or-higher tensor left-to-right (along axis 1).

Flip a 1-D-or-higher tensor upside-down (along axis 0).

Cyclically shift the elements of `input` along `dims`.

Return the lower-triangular part of a 2-D-or-higher tensor.

Return the upper-triangular part of a 2-D-or-higher tensor.

Pad a tensor with the specified border widths.

Gather values along `dim` according to `indices`.

Sort `input` along `dim` in ascending order and return values + indices.

Return the indices that would sort `input` along `dim`.

Return the indices of all non-zero (or True) elements.

Return the sorted unique elements of `input`.

Return the `k` largest values along `dim` and their indices.

Extract a (off-)diagonal from each matrix in a batched input.

Element-wise equality predicate.

Element-wise inequality predicate.

Element-wise strict greater-than predicate.

Element-wise greater-than-or-equal predicate.

Element-wise strict less-than predicate.

Element-wise less-than-or-equal predicate.

Elementwise selection between `x` and `y` based on `cond`.

Return `input` with positions where `mask` is True set to `value`.

Concatenate tensors along an existing dimension.

Stack tensors along a new dimension.

Stack tensors horizontally (column-wise).

Stack tensors vertically (row-wise).

Split a tensor into `chunks` roughly equal pieces along `dim`.

Remove `dim` and return a tuple of views along it.

Element-wise reciprocal square root.

Element-wise clamping to a closed interval.

Element-wise IEEE 754 infinity check.

Element-wise NaN check.

Element-wise IEEE 754 finiteness check.

Replace NaN, positive and negative infinity entries with finite values.

Return a tensor that shares data but is detached from the autograd graph.

Return a deep copy of `input` with its own storage.

Element-wise clamping to a closed interval.

Accumulate `src` into `input` along `dim` at positions `index`.

Element-wise base-10 logarithm.

Element-wise natural logarithm of `1 + x`.

Element-wise base-2 exponential.

Element-wise truncation toward zero.

Element-wise fractional part.

Element-wise quadrant-aware inverse tangent.

Element-wise C-style floating-point remainder.

Element-wise Python-style modulo.

Element-wise Euclidean norm of two operands, computed in a numerically stable way.

Element-wise log-sum-exp of two operands.

Element-wise IEEE 754 next representable floating-point value.

Compute the numerically stable log-sum-exp along `dim`.

Strict 2-D matrix multiplication.

Batched matrix multiplication.

Kronecker product of two tensors.

Element-wise logical AND.

Element-wise logical OR.

Element-wise logical exclusive-OR.

Element-wise logical NOT.

Return elements of `input` at the given flat `indices`.

Select slices of `input` along `dim` given `index`.

Return a contiguous slice of `input` along `dim`.

Write `src` into `base_impl` along `dim` at positions `index`.

Return the `k`-th smallest value along `dim` and its index.

Move one or more dimensions to new positions.

Inverse of `flatten` — split a single dim into a tuple of dims.

Compute a histogram of `input` with `bins` equally-spaced bins.

Return the Cartesian product of the given 1-D tensors.

Find insertion points to maintain order in a sorted sequence.

Return the bucket index of each element of `values` in `boundaries`.

Element-wise equality predicate.

Element-wise inequality predicate.

Element-wise strict less-than predicate.

Element-wise less-than-or-equal predicate.

Element-wise strict greater-than predicate.

Element-wise greater-than-or-equal predicate.

Element-wise inverse sine in radians.

Element-wise inverse cosine in radians.

Element-wise inverse tangent in radians.

Element-wise bitwise NOT.

Element-wise bitwise AND.

Element-wise bitwise OR.

Element-wise bitwise XOR.

Element-wise bitwise left shift.

Element-wise bitwise right shift.

Return a 1-D tensor of elements where `mask` is True.

Element-wise approximate equality with absolute and relative tolerance.

Return a view of `input` with a different shape.

Alias for `cat`.

Real part of a complex tensor.

Imaginary part of a complex tensor.

Element-wise complex conjugate.

Compute the arithmetic mean along `dim`.

Compute the product of elements along `dim`.

Compute the sample variance along `dim`.

Compute the sample standard deviation along `dim`.

Return the indices of the maximum values along `dim`.

Return the indices of the minimum values along `dim`.

Remove dimensions of size 1.

Tile a tensor along an existing dimension.

Repeat each element of `x` `repeats` times along `dim`.

Split a tensor into pieces of explicit sizes along `dim`.

Generalized tensor contraction over arbitrary axes.

Construct coordinate grids from 1-D coordinate vectors.

Compute the `n`-th order finite difference along an axis.

Linearly interpolate between `input` and an endpoint with a weight.