Tensor<T> class

A Tensor in this library is used as the fundamental unit to hold a structured collection of float values on the CPU. These values are stored for performance enhancement inside of Float32List. Globally a counter called idCounter ensures uniqueness of Tensors up to the integer limit of instances.

Each Tensor is both holding gradient value and its actual data value. Initialisation is to be done via its constructor.

IMPORTANT: The datatype T is not predefined as a standard value. To ensure compatibility with the framework the use of Scalar,Vector,Matrix and Tensor3D is recommended.

Tensors that end up orphaned are deleted using the Finalizer class but for maximum performance manual management is recommended. Boolean _freed is used as an additional safety feature allowing to see if the memory allocated has already been freed.

Constructors

Tensor(dynamic initialValue, {Node? creator})

Creates a Tensor object storing and managing the value initialValue given. Supported types for Tensor initialisation are: Scalar,Vector,Matrix and Tensor3D. Short versions of those types are Sc,Vec,Mat and T3D. Tensors can be created the following ways:

Properties

creator ↔ Node?

Node creator stores the node that is responsible for the creation of this Tensor. This allows for backward gradient propagation as well as tracing the origin path of this Tensor.

getter/setter pair

data ↔ Float32List

`Rolled out List of the Tensors value.

getter/setter pair

dataPtr ↔ Pointer<Float>

C pointer to the memory location of data.

getter/setter pair

grad ↔ Float32List

`Rolled out List of the Tensors gradient value.

getter/setter pair

gradPtr ↔ Pointer<Float>

C pointer to the memory location of grad.

getter/setter pair

gradValue → T

Converts grad via shape and returns the value as T.

no setter

hasData → bool

no setter

hashCode → int

The hash code for this object.

no setter, inherited

id ↔ String

Global identifier to find and query tensors. id is set to "t__idCounter by default. Id can be set anytime on CPU-bound Tensors.

getter/setter pair

isFreed → bool

no setter

runtimeType → Type

A representation of the runtime type of the object.

no setter, inherited

shape ↔ List<int>

Stores as which shape the Tensors value should be interpreted as. 0-3 dimensions are supported currently (2026-09).

getter/setter pair

value → T

Converts data via shape and returns the value as T.

no setter

Methods

backward({double startingGrad = 1.0}) → void

Starting point of gradient calculation. The network of Nodes and Tensors traverses in topological order over all Tensors affecting this Tensor adding a given gradient, calculated via their creator. The value of startingGrad affects the first gradient given into the backwards process and is set to 1.0 by default.

free() → void

Frees the allocated C memory and sets _freed to true.

noSuchMethod(Invocation invocation) → dynamic

Invoked when a nonexistent method or property is accessed.

inherited

printGraph() → void

Allows the exact printing of the current compute graph for this Tensor. Topological Depth is indicated via indents. Colors indicate tensor types. Green is indication of an input Tensor, blue an intermediate Tensor and Yellow the resulting output of the Network. Tensor name and shape are listed as t_tensorid [1,4,5,...] Example for a simple neuronal network:

toString() → String

A string representation of this object.

inherited

zeroGrad() → void

Traverses the computational Graph which created this Tensor, setting every value of every gradient to 0.0. This method in machine learning is to be used in between training iterations. Forgetting this method may result in oscillating or exploding gradient values.

Operators

operator ==(Object other) → bool

The equality operator.

inherited