Skip to main content

SparseOptimizer

torsh_sparse::optimizers

Trait SparseOptimizer 

Source 
pub trait SparseOptimizer {
    // Required methods
    fn step(
        &mut self,
        parameters: &mut [&mut CsrTensor],
        gradients: &[&CsrTensor],
    ) -> TorshResult<()>;
    fn lr(&self) -> f32;
    fn set_lr(&mut self, lr: f32);

    // Provided method
    fn zero_grad(&mut self) { ... }
}
Expand description

Trait for sparse optimizers

Required Methods§

Source

fn step( &mut self, parameters: &mut [&mut CsrTensor], gradients: &[&CsrTensor], ) -> TorshResult<()>

Update sparse parameters with sparse gradients

Source

fn lr(&self) -> f32

Get current learning rate

Source

fn set_lr(&mut self, lr: f32)

Set learning rate

Provided Methods§

Source

fn zero_grad(&mut self)

Zero gradients (if applicable)

Dyn Compatibility§

This trait is dyn compatible.

In older versions of Rust, dyn compatibility was called "object safety".

Implementors§