polish(pu): polish comments

Author: puyuan1996

ding/model/template/q_learning.py

@@ -58,6 +58,8 @@ def __init__(
             - dropout (:obj:`Optional[float]`): The dropout rate of the dropout layer. \
                 if ``None`` then default disable dropout layer.
             - init_bias (:obj:`Optional[float]`): The initial value of the last layer bias in the head network. \
+            - noise (:obj:`bool`): Whether use ``NoiseLinearLayer`` as ``layer_fn`` in Q networks' MLP. \
+                Default ``False``.
         """
         super(DQN, self).__init__()
         # Squeeze data from tuple, list or dict to single object. For example, from (4, ) to 4

ding/policy/common_utils.py

@@ -9,7 +9,8 @@
 
 def set_noise_mode(module: nn.Module, noise_enabled: bool):
     """
-    Recursively set the 'force_noise' flag on all NoiseLinearLayer modules within the given module.
+    Overview:
+        Recursively set the 'force_noise' flag on all NoiseLinearLayer modules within the given module.
     """
     for m in module.modules():
         if isinstance(m, NoiseLinearLayer):

ding/policy/dqn.py

@@ -386,7 +386,7 @@ def _forward_collect(self, data: Dict[int, Any], eps: float) -> Dict[int, Any]:
         data = default_collate(list(data.values()))
         if self._cuda:
             data = to_device(data, self._device)
-        # Use the new config parameter to decide noise mode.
+        # Use the add_noise parameter to decide noise mode.
         # Default to True if the parameter is not provided.
         if self._cfg.collect.get("add_noise", True):
             set_noise_mode(self._collect_model, True)

dizoo/atari/config/serial/demon_attack/demon_attack_dqn_config.py

@@ -1,6 +1,6 @@
 from easydict import EasyDict
 
-pong_dqn_config = dict(
+demon_attack_dqn_config = dict(
     exp_name='DemonAttack_dqn_collect-not-noise_seed0',
     env=dict(
         collector_env_num=8,
@@ -41,20 +41,20 @@
         ),
     ),
 )
-pong_dqn_config = EasyDict(pong_dqn_config)
-main_config = pong_dqn_config
-pong_dqn_create_config = dict(
+demon_attack_dqn_config = EasyDict(demon_attack_dqn_config)
+main_config = demon_attack_dqn_config
+demon_attack_dqn_create_config = dict(
     env=dict(
         type='atari',
         import_names=['dizoo.atari.envs.atari_env'],
     ),
     env_manager=dict(type='subprocess'),
     policy=dict(type='dqn'),
 )
-pong_dqn_create_config = EasyDict(pong_dqn_create_config)
-create_config = pong_dqn_create_config
+demon_attack_dqn_create_config = EasyDict(demon_attack_dqn_create_config)
+create_config = demon_attack_dqn_create_config
 
 if __name__ == '__main__':
-    # or you can enter `ding -m serial -c pong_dqn_config.py -s 0`
+    # or you can enter `ding -m serial -c demon_attack_dqn_config.py -s 0`
     from ding.entry import serial_pipeline
     serial_pipeline((main_config, create_config), seed=0, max_env_step=int(10e6))

dizoo/atari/config/serial/pong/pong_dqn_config.py

@@ -27,7 +27,7 @@
             learning_rate=0.0001,
             target_update_freq=500,
         ),
-        collect=dict(n_sample=96,),
+        collect=dict(n_sample=96, ),
         eval=dict(evaluator=dict(eval_freq=4000, )),
         other=dict(
             eps=dict(