protomotions.agents.evaluators.config module#

Configuration classes for evaluators.

class protomotions.agents.evaluators.config.EvaluatorConfig( _target_='protomotions.agents.evaluators.base_evaluator.BaseEvaluator', evaluation_components=<factory>, max_eval_steps=600, eval_metrics_every=200, )[source]#

Bases: object

Configuration for base evaluator.

Attributes:: evaluation_components: Dictionary of MdpComponent evaluation metrics for success/failure tracking. max_eval_steps: Maximum steps per evaluation episode. eval_metrics_every: Evaluate metrics every N epochs. None = disabled.

evaluation_components: Dict[str, MdpComponent]#

max_eval_steps: int = 600#

eval_metrics_every: int | None = 200#

__init__( _target_='protomotions.agents.evaluators.base_evaluator.BaseEvaluator', evaluation_components=<factory>, max_eval_steps=600, eval_metrics_every=200, )#

class protomotions.agents.evaluators.config.MotionWeightsRulesConfig( motion_weights_update_success_discount=0.999, motion_weights_update_failure_discount=0.999, min_motion_weight='1/num_motions', )[source]#

Bases: object

Configuration for motion weights update rule.

Attributes:: motion_weights_update_success_discount: Discount factor for successful motion weights. motion_weights_update_failure_discount: Discount for failed motions. 0 = set weight straight to 1. min_motion_weight: Minimum weight for any motion. ‘1/num_motions’ or float value.

motion_weights_update_success_discount: float = 0.999#

motion_weights_update_failure_discount: float = 0.999#

min_motion_weight: float | str = '1/num_motions'#

__init__( motion_weights_update_success_discount=0.999, motion_weights_update_failure_discount=0.999, min_motion_weight='1/num_motions', )#

class protomotions.agents.evaluators.config.MimicEvaluatorConfig( _target_='protomotions.agents.evaluators.mimic_evaluator.MimicEvaluator', evaluation_components=<factory>, max_eval_steps=600, eval_metrics_every=200, save_predicted_motion_lib_every=3, motion_weights_rules=<factory>, eval_action_ema_alpha=None, )[source]#

Bases: EvaluatorConfig

Configuration for Mimic evaluator.

Attributes:: evaluation_components: Dictionary of MdpComponent evaluation metrics for success/failure tracking. max_eval_steps: Maximum steps per evaluation episode. eval_metrics_every: Evaluate metrics every N epochs. None = disabled. save_predicted_motion_lib_every: Save pred_motion_lib every M evals. None = disabled. motion_weights_rules: Rules for updating motion sampling weights. eval_action_ema_alpha: EMA smoothing factor for actions during evaluation only. Simulates deployment low-pass filtering. a_applied = alpha * a_policy + (1-alpha) * a_prev. None = disabled (raw actions). Typical values: 0.5-0.8.Smaller alpha = more smoothing.

save_predicted_motion_lib_every: int | None = 3#

motion_weights_rules: MotionWeightsRulesConfig#

eval_action_ema_alpha: float | None = None#

__init__( _target_='protomotions.agents.evaluators.mimic_evaluator.MimicEvaluator', evaluation_components=<factory>, max_eval_steps=600, eval_metrics_every=200, save_predicted_motion_lib_every=3, motion_weights_rules=<factory>, eval_action_ema_alpha=None, )#