Merge branch 'main' into fix_avg_reward_query

assume/common/base.py

@@ -743,40 +743,58 @@ class LearningConfig(TypedDict):
 
     :param observation_dimension: The observation dimension.
     :type observation_dimension: int
+
     :param action_dimension: The action dimension.
     :type action_dimension: int
+
     :param continue_learning: Whether to continue learning.
     :type continue_learning: bool
+
     :param load_model_path: The path to the model to load.
     :type load_model_path: str
+
     :param max_bid_price: The maximum bid price.
     :type max_bid_price: float
+
     :param learning_mode: Whether to use learning mode.
     :type learning_mode: bool
+
     :param algorithm: The algorithm to use.
     :type algorithm: str
+
     :param learning_rate: The learning rate.
     :type learning_rate: float
+
     :param training_episodes: The number of training episodes.
     :type training_episodes: int
+
     :param episodes_collecting_initial_experience: The number of episodes collecting initial experience.
     :type episodes_collecting_initial_experience: int
+
     :param train_freq: The training frequency.
     :type train_freq: int
+
     :param gradient_steps: The number of gradient steps.
     :type gradient_steps: int
+
     :param batch_size: The batch size.
     :type batch_size: int
-    :param gamme: The discount factor.
+
+    :param gamma: The discount factor.
     :type gamma: float
+
     :param device: The device to use.
     :type device: str
+
     :param noise_sigma : The standard deviation of the noise.
     :type noise_sigma: float
+
     :param noise_scale: Controls the initial strength of the noise.
     :type noise_scale: int
+
     :param noise_dt: Determines how quickly the noise weakens over time.
     :type noise_dt: int
+
     :param trained_actors_path: The path to the learned model to load.
     :type trained_actors_path: str
     """