Struct ProdigyOptimizer 

pub struct ProdigyOptimizer { /* private fields */ }

Prodigy optimizer

Auto-tuning adaptive learning rate optimizer that estimates the distance scale D from initialization to automatically set appropriate step sizes.

Key features:

• No manual LR tuning needed
• Adapts to problem scale automatically
• Combines Adam-style updates with D-Adaptation
• Maintains first and second moment estimates

Implementations

impl ProdigyOptimizer

pub fn new(config: ProdigyConfig) -> Self

Create new Prodigy optimizer with given config

pub fn get_d(&self) -> f64

Get current D estimate

pub fn get_step(&self) -> usize

Get current step count

Trait Implementations

impl Optimizer for ProdigyOptimizer

fn zero_grad(&mut self)

Zero all gradients.

fn get_lr(&self) -> f64

Get current learning rate.

fn set_lr(&mut self, lr: f64)

Set learning rate.

fn step( &mut self, parameters: &mut HashMap<String, Array2<f64>>, gradients: &HashMap<String, Array2<f64>>, ) -> TrainResult<()>

Update parameters with computed gradients.

fn state_dict(&self) -> HashMap<String, Vec<f64>>

Get optimizer state for checkpointing.

fn load_state_dict(&mut self, state: HashMap<String, Vec<f64>>)

Load optimizer state from checkpoint.