Struct Parameter 

pub struct Parameter { /* private fields */ }

A learnable parameter of a neural network module.

Parameters wrap Variables and provide additional functionality:

• Automatic requires_grad=true by default
• Registration with parent modules
• Easy access to data and gradients

Implementations

impl Parameter

pub fn new(data: Tensor<f32>, requires_grad: bool) -> Self

Creates a new parameter from a tensor.

Arguments

• data - The tensor data
• requires_grad - Whether to track gradients (default true)

pub fn named( name: impl Into<String>, data: Tensor<f32>, requires_grad: bool, ) -> Self

Creates a new parameter with a name.

pub fn from_variable(var: Variable) -> Self

Creates a parameter from an existing Variable.

pub fn name(&self) -> &str

Returns the parameter name.

pub fn set_name(&mut self, name: impl Into<String>)

Sets the parameter name.

pub fn variable(&self) -> Variable

Returns a clone of the underlying Variable.

pub fn data(&self) -> Tensor<f32>

Returns a clone of the tensor data.

pub fn shape(&self) -> Vec<usize>

Returns the shape of the parameter.

pub fn numel(&self) -> usize

Returns the number of elements.

pub fn requires_grad(&self) -> bool

Returns whether this parameter requires gradients.

pub fn grad(&self) -> Option<Tensor<f32>>

Returns the gradient if available.

pub fn zero_grad(&self)

Zeros the gradient.

pub fn update_data(&self, new_data: Tensor<f32>)

Updates the parameter data in-place.

Used by optimizers to update weights.

pub fn apply_update<F>(&self, f: F)

where F: FnOnce(&Tensor<f32>) -> Tensor<f32>,

Applies a function to the parameter data.

Trait Implementations

impl Clone for Parameter

fn clone(&self) -> Parameter

Returns a duplicate of the value.

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

Performs copy-assignment from source.

impl Debug for Parameter

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

Formats the value using the given formatter.