Autograd-Math
mean
Mean of a Variable
, alongside the specified axis.
- axis
axis to compute the mean. 0-based indexed.
- keepDims
A boolean, whether to keep the dimensions or not.
If keepDims
is False
, the rank of the Variable
is reduced
by 1. If keepDims
is True
,
the reduced dimensions are retained with length 1.
Scala example
mean(x: Variable[T], axis: Int = 0, keepDims: Boolean = false)
Python example
mean(x, axis=0, keepDims=False):
abs
Element-wise absolute value.
- x
A Variable
.
Scala example
abs(x: Variable[T])
Python example
abs(x):
sum
Sum of the values in a Variable
, alongside the specified axis.
- axis
axis to compute the mean. 0-based indexed.
- keepDims
A boolean, whether to keep the dimensions or not.
If keepDims
is False
, the rank of the Variable
is reduced
by 1. If keepDims
is True
,
the reduced dimensions are retained with length 1.
Scala example
sum(x: Variable[T], axis: Int = 0, keepDims: Boolean = false)
Python example
sum(x, axis=0, keepDims=False):
clip
Element-wise value clipping.
- x
A Variable
.
- min
Double
- max
Double
Scala example
clip(x: Variable[T], min: Double, max: Double)
Python example
clip(x, min, max)
square
Element-wise square.
- x
A Variable
.
Scala example
square(x: Variable[T])
Python example
square(x):
sqrt
Element-wise square root.
- x
A Variable
.
Scala example
sqrt(x: Variable[T])
Python example
sqrt(x):
maximum
Element-wise maximum of two Variables
.
- x
A Variable
.
- y
A Variable
or Double.
Scala example
maximum(x: Variable[T], y: Variable[T])
Python example
maximum(x, y):
log
Element-wise log.
- x
A Variable
.
Scala example
log(x: Variable[T])
Python example
log(x):
exp
Element-wise exponential.
- x
A Variable
.
Scala example
exp(x: Variable[T])
Python example
exp(x):
pow
Element-wise exponentiation.
- x
A Variable
.
- a
Double.
Scala example
pow(x: Variable[T])
Python example
pow(x):
softsign
Softsign of a Variable
.
Scala example
softsign(x: Variable[T])
Python example
softsign(x):
softplus
Softplus of a Variable
.
Scala example
softplus(x: Variable[T])
Python example
softplus(x):
stack
Stacks a list of rank R
tensors into a rank R+1
tensor.
You should start from 1 as dim 0 is for batch.
- inputs: List of variables (tensors)
- axis: xis along which to perform stacking.
Scala example
def stack[T: ClassTag](inputs: List[Variable[T]], axis: Int = 1)
Python example
def stack(inputs, axis=1)
expand_dims
Adds a 1-sized dimension at index "axis".
Scala example
def expandDims[T: ClassTag](x: Variable[T], axis: Int)
Python example
expand_dims(x, axis)
contiguous
Turn the output and grad to be contiguous for the input Variable
Scala example
def contiguous[T: ClassTag](input: Variable[T])
Python example
def contiguous(x)
mm
Module to perform matrix multiplication on two mini-batch inputs, producing a mini-batch.
- x
A variable.
- y
A variable.
- axes
Axes along which to perform multiplication.
Scala example
def mm[T: ClassTag](x: Variable[T], y: Variable[T], axes: List[Int])
Python example
def mm(x, y, axes)
batch_dot
Operator that computes a dot product between samples in two tensors.
- x
Shape should only be [batch, xx]
- y
Shape should only be [batch, xx]
- axes
Integer or tuple of integers, axis or axes along which to take the dot product.
- normalize
Whether to L2-normalize samples along the
dot product axis before taking the dot product.
If set to True, then the output of the dot product
is the cosine proximity between the two samples.
Scala example
def batchDot[T: ClassTag](x: Variable[T], y: Variable[T], axes: List[Int], normalize: Boolean = false)
Python example
def batch_dot(x, y, axes=1, normalize=False)
l2_normalize
Normalizes a tensor wrt the L2 norm alongside the specified axis.
- x
A variable.
- axis
Axis along which to perform normalization.
Scala example
def l2Normalize[T: ClassTag](x: Variable[T], axis: Int)
Python example
def l2_normalize(x, axis)