Board is always copied when returned as an observation to avoid in-pl…

…ace modification of the observation when state of the environment changes
MultiAgentLearning · Feb 14, 2020 · 4ba44f6 · 4ba44f6
1 parent 69da607
commit 4ba44f6
Showing 1 changed file with 1 addition and 2 deletions.
diff --git a/pommerman/forward_model.py b/pommerman/forward_model.py
@@ -546,9 +546,8 @@ def in_view_range(position, v_row, v_col):
         observations = []
         for agent in agents:
             agent_obs = {'alive': alive_agents}
-            board = curr_board
+            board = curr_board.copy()
             if is_partially_observable:
-                board = board.copy()
                 for row in range(board_size):
                     for col in range(board_size):
                         if not in_view_range(agent.position, row, col):