Add policy documentation links to policy_kwargs parameter

docs/modules/ppo_mask.rst

@@ -245,6 +245,7 @@ Parameters
   :members:
   :inherited-members:
 
+.. _ppo_mask_policies:
 
 MaskablePPO Policies
 --------------------

docs/modules/ppo_recurrent.rst

@@ -125,6 +125,7 @@ Parameters
   :members:
   :inherited-members:
 
+.. _ppo_recurrent_policies:
 
 RecurrentPPO Policies
 ---------------------

sb3_contrib/__init__.py

@@ -15,10 +15,10 @@
 
 __all__ = [
     "ARS",
-    "CrossQ",
-    "MaskablePPO",
-    "RecurrentPPO",
     "QRDQN",
     "TQC",
     "TRPO",
+    "CrossQ",
+    "MaskablePPO",
+    "RecurrentPPO",
 ]

sb3_contrib/ars/ars.py

@@ -40,7 +40,7 @@ class ARS(BaseAlgorithm):
     :param zero_policy: Boolean determining if the passed policy should have it's weights zeroed before training.
     :param alive_bonus_offset: Constant added to the reward at each step, used to cancel out alive bonuses.
     :param n_eval_episodes: Number of episodes to evaluate each candidate.
-    :param policy_kwargs: Keyword arguments to pass to the policy on creation
+    :param policy_kwargs: Keyword arguments to pass to the policy on creation. See :ref:`ars_policies`
     :param stats_window_size: Window size for the rollout logging, specifying the number of episodes to average
         the reported success rate, mean episode length, and mean reward over
     :param tensorboard_log: String with the directory to put tensorboard logs:

sb3_contrib/common/torch_layers.py

@@ -1,6 +1,6 @@
 import torch
 
-__all__ = ["BatchRenorm1d", "BatchRenorm"]
+__all__ = ["BatchRenorm", "BatchRenorm1d"]
 
 
 class BatchRenorm(torch.nn.Module):

sb3_contrib/crossq/crossq.py

@@ -56,7 +56,7 @@ class CrossQ(OffPolicyAlgorithm):
     :param stats_window_size: Window size for the rollout logging, specifying the number of episodes to average
         the reported success rate, mean episode length, and mean reward over
     :param tensorboard_log: the log location for tensorboard (if None, no logging)
-    :param policy_kwargs: additional arguments to be passed to the policy on creation
+    :param policy_kwargs: additional arguments to be passed to the policy on creation. See :ref:`crossq_policies`
     :param verbose: Verbosity level: 0 for no output, 1 for info messages (such as device or wrappers used), 2 for
         debug messages
     :param seed: Seed for the pseudo random generators

sb3_contrib/ppo_mask/__init__.py

@@ -1,4 +1,4 @@
 from sb3_contrib.ppo_mask.policies import CnnPolicy, MlpPolicy, MultiInputPolicy
 from sb3_contrib.ppo_mask.ppo_mask import MaskablePPO
 
-__all__ = ["CnnPolicy", "MlpPolicy", "MultiInputPolicy", "MaskablePPO"]
+__all__ = ["CnnPolicy", "MaskablePPO", "MlpPolicy", "MultiInputPolicy"]

sb3_contrib/ppo_mask/ppo_mask.py

@@ -57,7 +57,7 @@ class MaskablePPO(OnPolicyAlgorithm):
     :param stats_window_size: Window size for the rollout logging, specifying the number of episodes to average
         the reported success rate, mean episode length, and mean reward over
     :param tensorboard_log: the log location for tensorboard (if None, no logging)
-    :param policy_kwargs: additional arguments to be passed to the policy on creation
+    :param policy_kwargs: additional arguments to be passed to the policy on creation. See :ref:`ppo_mask_policies`
     :param verbose: the verbosity level: 0 no output, 1 info, 2 debug
     :param seed: Seed for the pseudo random generators
     :param device: Device (cpu, cuda, ...) on which the code should be run.

sb3_contrib/ppo_recurrent/ppo_recurrent.py

@@ -57,7 +57,7 @@ class RecurrentPPO(OnPolicyAlgorithm):
     :param stats_window_size: Window size for the rollout logging, specifying the number of episodes to average
         the reported success rate, mean episode length, and mean reward over
     :param tensorboard_log: the log location for tensorboard (if None, no logging)
-    :param policy_kwargs: additional arguments to be passed to the policy on creation
+    :param policy_kwargs: additional arguments to be passed to the policy on creation. See :ref:`ppo_recurrent_policies`
     :param verbose: the verbosity level: 0 no output, 1 info, 2 debug
     :param seed: Seed for the pseudo random generators
     :param device: Device (cpu, cuda, ...) on which the code should be run.

sb3_contrib/qrdqn/__init__.py

@@ -1,4 +1,4 @@
 from sb3_contrib.qrdqn.policies import CnnPolicy, MlpPolicy, MultiInputPolicy
 from sb3_contrib.qrdqn.qrdqn import QRDQN
 
-__all__ = ["CnnPolicy", "MlpPolicy", "MultiInputPolicy", "QRDQN"]
+__all__ = ["QRDQN", "CnnPolicy", "MlpPolicy", "MultiInputPolicy"]

sb3_contrib/qrdqn/qrdqn.py

@@ -53,7 +53,7 @@ class QRDQN(OffPolicyAlgorithm):
     :param stats_window_size: Window size for the rollout logging, specifying the number of episodes to average
         the reported success rate, mean episode length, and mean reward over
     :param tensorboard_log: the log location for tensorboard (if None, no logging)
-    :param policy_kwargs: additional arguments to be passed to the policy on creation
+    :param policy_kwargs: additional arguments to be passed to the policy on creation. See :ref:`qrdqn_policies`
     :param verbose: the verbosity level: 0 no output, 1 info, 2 debug
     :param seed: Seed for the pseudo random generators
     :param device: Device (cpu, cuda, ...) on which the code should be run.

sb3_contrib/tqc/__init__.py

@@ -1,4 +1,4 @@
 from sb3_contrib.tqc.policies import CnnPolicy, MlpPolicy, MultiInputPolicy
 from sb3_contrib.tqc.tqc import TQC
 
-__all__ = ["CnnPolicy", "MlpPolicy", "MultiInputPolicy", "TQC"]
+__all__ = ["TQC", "CnnPolicy", "MlpPolicy", "MultiInputPolicy"]

sb3_contrib/tqc/tqc.py

@@ -60,7 +60,7 @@ class TQC(OffPolicyAlgorithm):
     :param stats_window_size: Window size for the rollout logging, specifying the number of episodes to average
         the reported success rate, mean episode length, and mean reward over
     :param tensorboard_log: the log location for tensorboard (if None, no logging)
-    :param policy_kwargs: additional arguments to be passed to the policy on creation
+    :param policy_kwargs: additional arguments to be passed to the policy on creation. See :ref:`tqc_policies`
     :param verbose: the verbosity level: 0 no output, 1 info, 2 debug
     :param seed: Seed for the pseudo random generators
     :param device: Device (cpu, cuda, ...) on which the code should be run.

sb3_contrib/trpo/__init__.py

@@ -1,4 +1,4 @@
 from sb3_contrib.trpo.policies import CnnPolicy, MlpPolicy, MultiInputPolicy
 from sb3_contrib.trpo.trpo import TRPO
 
-__all__ = ["CnnPolicy", "MlpPolicy", "MultiInputPolicy", "TRPO"]
+__all__ = ["TRPO", "CnnPolicy", "MlpPolicy", "MultiInputPolicy"]

sb3_contrib/trpo/trpo.py

@@ -64,7 +64,7 @@ class TRPO(OnPolicyAlgorithm):
     :param stats_window_size: Window size for the rollout logging, specifying the number of episodes to average
         the reported success rate, mean episode length, and mean reward over
     :param tensorboard_log: the log location for tensorboard (if None, no logging)
-    :param policy_kwargs: additional arguments to be passed to the policy on creation
+    :param policy_kwargs: additional arguments to be passed to the policy on creation. See :ref:`trpo_policies`
     :param verbose: the verbosity level: 0 no output, 1 info, 2 debug
     :param seed: Seed for the pseudo random generators
     :param device: Device (cpu, cuda, ...) on which the code should be run.