Struct SafeVariable 

pub struct SafeVariable<F: Float + Send + Sync> {
    pub id: VariableID,
    pub env: Arc<SafeVariableEnvironment<F>>,
    pub requires_grad: bool,
}

PyTorch-compatible Variable wrapper for ToRSh integration

This provides a PyTorch-style Variable interface that wraps the SciRS2 autograd system. Unlike the RefCell-based Variable, this is thread-safe and can be used globally.

Fields

id: VariableID

Variable ID in the environment

env: Arc<SafeVariableEnvironment<F>>

Reference to the thread-safe environment

requires_grad: bool

Whether this variable requires gradients

Implementations

impl<F: Float + Send + Sync> SafeVariable<F>

pub fn new( data: NdArray<F>, env: Arc<SafeVariableEnvironment<F>>, requires_grad: bool, ) -> Result<Self, Box<dyn Error + Send + Sync>>

Create a new variable with gradient requirement

pub fn backward(&self) -> Result<(), Box<dyn Error + Send + Sync>>

PyTorch-compatible backward() method

pub fn data(&self) -> Result<NdArray<F>, Box<dyn Error + Send + Sync>>

Get the current data (read-only)

pub fn requires_grad(&self) -> bool

Check if gradients are required

pub fn set_requires_grad(&mut self, requires_grad: bool)

Set gradient requirement

Trait Implementations

impl<F: Float + Send + Sync> AutogradTensor<F> for SafeVariable<F>

fn backward(&self) -> Result<(), Box<dyn Error + Send + Sync>>

fn grad(&self) -> Option<&NdArray<F>>

fn requires_grad(&self) -> bool

fn set_requires_grad(&mut self, requires_grad: bool)

impl<F: Clone + Float + Send + Sync> Clone for SafeVariable<F>

fn clone(&self) -> SafeVariable<F>

Returns a duplicate of the value.

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

Performs copy-assignment from source.

impl<F: Float + Send + Sync> Send for SafeVariable<F>

Implement Send + Sync for thread safety

impl<F: Float + Send + Sync> Sync for SafeVariable<F>