Struct Value 

pub struct Value {
    pub id: ValueId,
    pub name: Option<String>,
    pub dtype: DataType,
    pub shape: Shape,
    pub layout: TensorLayout,
    pub device: Option<DeviceId>,
    pub producer: Option<NodeId>,
    pub consumers: Consumers,
    pub is_graph_input: bool,
    pub is_graph_output: bool,
}

A value flowing through the graph — the output of one node and the input of zero or more others (SSA form: at most one producer).

Carries a first-class TensorLayout and optional DeviceId placement, beyond ONNX’s plain type+shape.

Fields

id: ValueId

name: Option<String>

Optional name (graph inputs/outputs and initializers are named; interior SSA values may be anonymous).

dtype: DataType

shape: Shape

layout: TensorLayout

device: Option<DeviceId>

Device placement, filled in by the placement pass.

producer: Option<NodeId>

The node that produces this value, or None for graph inputs and initializers.

consumers: Consumers

Input slots that consume this value, keyed by (node, input_index).

is_graph_input: bool

Whether this value is present in Graph::inputs.

is_graph_output: bool

Whether this value is present in Graph::outputs.

Implementations

impl Value

pub fn new(id: ValueId, dtype: DataType, shape: Shape) -> Self

A new anonymous value with a contiguous default layout and no edges.

pub fn is_source(&self) -> bool

Whether this value has no producing node (graph input / initializer).

pub fn rank(&self) -> usize

The static rank (number of dimensions).

Trait Implementations

impl Clone for Value

fn clone(&self) -> Value

Returns a duplicate of the value.

fn clone_from(&mut self, source: &Self)

Performs copy-assignment from source.

impl Debug for Value

fn fmt(&self, f: &mut Formatter<'_>) -> Result

Formats the value using the given formatter.

impl PartialEq for Value

fn eq(&self, other: &Value) -> bool

Equality operator ==.

fn ne(&self, other: &Rhs) -> bool

Inequality operator !=.

impl StructuralPartialEq for Value