Merge pull request google-deepmind#988 from NightMachinery:master

PiperOrigin-RevId: 500691669 Change-Id: I8900f86d97b8f49262abe1077d7866f606f39452
KezhiAdore · Jan 9, 2023 · 521d01a · 521d01a
2 parents 434a1cb + 56af2b7
commit 521d01a
Show file tree

Hide file tree

Showing 3 changed files with 10 additions and 5 deletions.
diff --git a/open_spiel/python/algorithms/tabular_qlearner.py b/open_spiel/python/algorithms/tabular_qlearner.py
@@ -22,6 +22,9 @@
 
 
 def valuedict():
+  # The default factory is called without arguments to produce a new value when
+  # a key is not present, in __getitem__ only. This value is added to the dict,
+  # so modifying it will modify the dict.
   return collections.defaultdict(float)
 
 

diff --git a/open_spiel/python/examples/tic_tac_toe_dqn_vs_tabular.py b/open_spiel/python/examples/tic_tac_toe_dqn_vs_tabular.py
@@ -34,7 +34,7 @@
 
 flags.DEFINE_integer("num_episodes", int(5e4), "Number of train episodes.")
 flags.DEFINE_boolean(
-    "iteractive_play", True,
+    "interactive_play", True,
     "Whether to run an interactive play with the agent after training.")
 
 
@@ -135,7 +135,7 @@ def main(_):
     r_mean = eval_against_random_bots(env, agents, random_agents, 1000)
     logging.info("Mean episode rewards: %s", r_mean)
 
-    if not FLAGS.iteractive_play:
+    if not FLAGS.interactive_play:
       return
 
     # Play from the command line against the trained DQN agent.

diff --git a/open_spiel/python/examples/tic_tac_toe_qlearner.py b/open_spiel/python/examples/tic_tac_toe_qlearner.py
@@ -35,8 +35,10 @@
 
 flags.DEFINE_integer("num_episodes", int(5e4), "Number of train episodes.")
 flags.DEFINE_boolean(
-    "iteractive_play", True,
-    "Whether to run an interactive play with the agent after training.")
+    "interactive_play",
+    True,
+    "Whether to run an interactive play with the agent after training.",
+)
 
 
 def pretty_board(time_step):
@@ -120,7 +122,7 @@ def main(_):
     for agent in agents:
       agent.step(time_step)
 
-  if not FLAGS.iteractive_play:
+  if not FLAGS.interactive_play:
     return
 
   # 2. Play from the command line against the trained agent.