memorax.algorithms.GradientPPOConfig

memorax.algorithms.GradientPPOConfig#

class memorax.algorithms.GradientPPOConfig[source]#

Bases: object

GradientPPOConfig(num_envs: int, num_steps: int, gamma: float, gae_lambda: float, num_minibatches: int, update_epochs: int, normalize_advantage: bool, clip_coefficient: float, clip_value_loss: bool, entropy_coefficient: float, regularization_coefficient: float, truncation_length: int, burn_in_length: int = 0)

num_envs: int#
num_steps: int#
gamma: float#
gae_lambda: float#
num_minibatches: int#
update_epochs: int#
normalize_advantage: bool#
clip_coefficient: float#
clip_value_loss: bool#
entropy_coefficient: float#
regularization_coefficient: float#
truncation_length: int#
burn_in_length: int = 0#
property batch_size#
__init__(num_envs, num_steps, gamma, gae_lambda, num_minibatches, update_epochs, normalize_advantage, clip_coefficient, clip_value_loss, entropy_coefficient, regularization_coefficient, truncation_length, burn_in_length=0)#
Parameters:

  • num_envs (int)

  • num_steps (int)

  • gamma (float)

  • gae_lambda (float)

  • num_minibatches (int)

  • update_epochs (int)

  • normalize_advantage (bool)

  • clip_coefficient (float)

  • clip_value_loss (bool)

  • entropy_coefficient (float)

  • regularization_coefficient (float)

  • truncation_length (int)

  • burn_in_length (int)

Return type:

None

replace(**updates)#

Returns a new object replacing the specified fields with new values.