Struct CandleDeterministicPolicy 

pub struct CandleDeterministicPolicy { /* private fields */ }

Implementations

impl CandleDeterministicPolicy

pub fn new( obs_dim: usize, act_dim: usize, hidden: usize, max_action: f32, lr: f64, device: Device, ) -> Result<Self, NNError>

impl CandleDeterministicPolicy

pub fn td3_actor_step( &mut self, obs: &TensorData, critic: &CandleTwinQ, ) -> Result<TrainMetrics, NNError>

TD3 actor gradient step with autograd flowing through the critic.

Takes concrete CandleTwinQ to preserve gradient flow from Q1(s, a) back to actor parameters via the deterministic actions.

Trait Implementations

impl DeterministicPolicy for CandleDeterministicPolicy

fn act(&self, obs: &TensorData) -> Result<TensorData, NNError>

Compute deterministic action. Returns [batch_size, act_dim].

fn target_act(&self, obs: &TensorData) -> Result<TensorData, NNError>

Compute target policy action (from target network).

fn soft_update_target(&mut self, tau: f32)

Polyak soft update of target network.

fn learning_rate(&self) -> f32

fn set_learning_rate(&mut self, lr: f32)

fn save(&self, path: &Path) -> Result<(), NNError>

fn load(&mut self, path: &Path) -> Result<(), NNError>

impl Send for CandleDeterministicPolicy

impl Sync for CandleDeterministicPolicy