scalation/scalation/scalation.modeling/scalation.modeling.autograd/Function

Function

scalation.modeling.autograd.Function

The Function base trait for all differentiable operations in the autograd system. A Function encapsulates both the forward computation (producing outputs) and the backward computation (propagating gradients). It also provides utility methods for handling unbroadcasting of shapes during the backward pass, ensuring correct gradient flow. Every custom operation should extend this trait and implement forward and backward.

Attributes

Graph
Supertypes: class Object

trait Matchable

class Any
Known subtypes: class Abs

class Add

class AddConstant

class BatchMatMul

class Ceil

class Clip

class Concat

class Div

class DivConstant

class Dot

class ELU

class Exp

class Floor

class GRUCellFused

class GeLU

class Identity

class LeakyReLU

class Log

class LogBase

class MAELoss

class MSELoss

class MatMul

class Max

class MaxScalar

class MaxValue

class Mean

class MeanAlongAxis

class Min

class MinScalar

class MinValue

class Mul

class MulConstant

class Neg

class Permute

class Pow

class RNNCellFused

class RNNFused

class ReLU

class Reciprocal

class Reshape

class Round

class SSELoss

class Sigmoid

class Sign

class Slice

class Softmax

class Sqrt

class Std

class StdAlongAxis

class Sub

class SubConstant

class Sum

class Tanh

class Transpose

class Variance

class VarianceAlongAxis
Show all

Members list

Value members

Abstract methods

Performs the backward pass given the upstream gradient.

Value parameters

gradOutput: the gradient tensor from the next layer.

Attributes

Performs the forward pass to compute the output variable.

Attributes

Returns: a Variabl containing the output data and gradient function.

Concrete methods

Map of attributes for visualization/debugging (default: empty).

Attributes

Backpropagates gradients for functions with two inputs.

Value parameters

computeGrad1: function to compute the gradient for v1.
computeGrad2: function to compute the gradient for v2.
gradOutput: the upstream gradient tensor.
v1: the first input variable.
v2: the second input variable.

Attributes

Returns the input variables to this Function. This works automatically for all case-class ops by iterating over their constructor fields and collecting those of type Variabl.

Attributes

See also: Case classes and Product: https://scala-lang.org/api/3.x/scala/Product.html

Human-readable name of this op (defaults to simple class name).

Attributes

Unbroadcasts a variable's tensor data to a specified old shape.

Value parameters

oldShape: the target shape.
v: the variable to unbroadcast.

Attributes

Returns: a new Variabl with data unbroadcasted.

Unbroadcasts a tensor to a given shape by summing across reduced dimensions.

Value parameters

data: the tensor data.
oldShape: the original shape.

Attributes

Returns: a TensorD with shape adjusted to oldShape.
Throws: Exception
if unbroadcasting is not feasible.

Concrete fields

Unique numeric ID for this Function node (for graph viz/debugging).

Attributes

In this article

Generated with