Struct LearnerReport 

pub struct LearnerReport {
    pub updates_completed: usize,
    pub env_steps_consumed: usize,
    pub episodes_completed: usize,
    pub broadcasts_sent: usize,
    pub last_policy_version: u64,
    pub episode_rewards: Vec<f32>,
    pub final_stats: Option<TrainingStats>,
}

Summary returned by learner_loop alongside the trained crate::train::ppo::trainer::PPOTrainerBurn.

Fields

updates_completed: usize

PPO updates completed.

env_steps_consumed: usize

Env steps consumed into updates (num_steps * num_actors each).

episodes_completed: usize

Episodes completed within the consumed steps.

broadcasts_sent: usize

Policy broadcasts sent (one per broadcast_every updates).

last_policy_version: u64

Version number of the newest broadcast policy.

episode_rewards: Vec<f32>

Total reward of every completed episode, in consumption order.

final_stats: Option<TrainingStats>

Stats from the final PPO update, if any ran.

Implementations

impl LearnerReport

pub fn mean_recent_episode_reward(&self, n: usize) -> f32

Mean total reward over the most recent n completed episodes (or all of them when fewer have completed). Returns 0.0 before the first episode completes.

Trait Implementations

impl Clone for LearnerReport

fn clone(&self) -> LearnerReport

Returns a duplicate of the value.

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

Performs copy-assignment from source.

impl Debug for LearnerReport

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

Formats the value using the given formatter.

impl Default for LearnerReport

fn default() -> LearnerReport

Returns the “default value” for a type.