fix: normalize_reward modifies the ends in-placely

typoverflow · Jul 18, 2023 · f69cbf1 · f69cbf1
1 parent 9297785
commit f69cbf1
Show file tree

Hide file tree

Showing 2 changed files with 2 additions and 2 deletions.
diff --git a/offlinerllib/__init__.py b/offlinerllib/__init__.py
@@ -1,2 +1,2 @@
 
-__version__ = "0.0.10"
+__version__ = "0.0.11"
diff --git a/offlinerllib/utils/d4rl.py b/offlinerllib/utils/d4rl.py
@@ -10,7 +10,7 @@ def antmaze_normalize_reward(dataset):
     return dataset, {}
 
 def mujoco_normalize_reward(dataset):
-    split_points = dataset["ends"]
+    split_points = dataset["ends"].copy()
     split_points[-1] = False   # the last traj may be incomplete, so we discard them
     reward = dataset["rewards"]
     returns, lengths = [], []