lit_sb3.core.noise

source module lit_sb3.core.noise

Classes

• ActionNoise — The action noise base class.

• NormalActionNoise — A Gaussian action noise.

• OrnsteinUhlenbeckActionNoise — An Ornstein Uhlenbeck action noise, this is designed to approximate Brownian motion with friction.

• VectorizedActionNoise — A Vectorized action noise for parallel environments.

source class ActionNoise()

Bases : ABC

The action noise base class.

Methods

• reset — Call end of episode reset for the noise.

source method ActionNoise.reset() → None

Call end of episode reset for the noise.

source class NormalActionNoise(mean: np.ndarray, sigma: np.ndarray, dtype: DTypeLike = np.float32)

Bases : ActionNoise

A Gaussian action noise.

Parameters

• mean : np.ndarray — Mean value of the noise

• sigma : np.ndarray — Scale of the noise (std here)

• dtype : DTypeLike — Type of the output noise

source class OrnsteinUhlenbeckActionNoise(mean: np.ndarray, sigma: np.ndarray, theta: float = 0.15, dt: float = 0.01, initial_noise: np.ndarray | None = None, dtype: DTypeLike = np.float32)

Bases : ActionNoise

An Ornstein Uhlenbeck action noise, this is designed to approximate Brownian motion with friction.

Based on http://math.stackexchange.com/questions/1287634/implementing-ornstein-uhlenbeck-in-matlab

Parameters

• mean : np.ndarray — Mean of the noise

• sigma : np.ndarray — Scale of the noise

• theta : float — Rate of mean reversion

• dt : float — Timestep for the noise

• initial_noise : np.ndarray | None — Initial value for the noise output, (if None: 0)

• dtype : DTypeLike — Type of the output noise

Methods

• reset — Reset the Ornstein Uhlenbeck noise, to the initial position.

source method OrnsteinUhlenbeckActionNoise.reset() → None

Reset the Ornstein Uhlenbeck noise, to the initial position.

source class VectorizedActionNoise(base_noise: ActionNoise, n_envs: int)

Bases : ActionNoise

A Vectorized action noise for parallel environments.

Parameters

• base_noise : ActionNoise — Noise generator to use

• n_envs : int — Number of parallel environments

Methods

• reset — Reset all the noise processes, or those listed in indices.

source method VectorizedActionNoise.reset(indices: Iterable[int] | None = None) → None

Reset all the noise processes, or those listed in indices.

Parameters

• indices : Iterable[int] | None — The indices to reset. Default: None. If the parameter is None, then all processes are reset to their initial position.

source attribute VectorizedActionNoise.base_noise

source attribute VectorizedActionNoise.noises