From 65d3d62e38700997e2d7c7c98dc22d8b23d8a9f7 Mon Sep 17 00:00:00 2001
From: whning <whning@zju.edu.cn>
Date: Fri, 5 Jun 2026 02:44:37 +0800
Subject: [PATCH] Fix mutable default arguments in RL files (2 files: PPO
 actor/critic, simulator simple)

---
 qlib/rl/order_execution/policy.py           | 8 ++++++--
 qlib/rl/order_execution/simulator_simple.py | 9 +++++++--
 2 files changed, 13 insertions(+), 4 deletions(-)

diff --git a/qlib/rl/order_execution/policy.py b/qlib/rl/order_execution/policy.py
index a46b587aa11..539afbb2b5c 100644
--- a/qlib/rl/order_execution/policy.py
+++ b/qlib/rl/order_execution/policy.py
@@ -76,8 +76,10 @@ def forward(
         self,
         obs: torch.Tensor,
         state: torch.Tensor = None,
-        info: dict = {},
+        info: dict = None,
     ) -> Tuple[torch.Tensor, Optional[torch.Tensor]]:
+        if info is None:
+            info = {}
         feature = self.extractor(to_torch(obs, device=auto_device(self)))
         out = self.layer_out(feature)
         return out, state
@@ -93,8 +95,10 @@ def forward(
         self,
         obs: torch.Tensor,
         state: torch.Tensor = None,
-        info: dict = {},
+        info: dict = None,
     ) -> torch.Tensor:
+        if info is None:
+            info = {}
         feature = self.extractor(to_torch(obs, device=auto_device(self)))
         return self.value_out(feature).squeeze(dim=-1)
 
diff --git a/qlib/rl/order_execution/simulator_simple.py b/qlib/rl/order_execution/simulator_simple.py
index 48aa03a1707..b96889873f1 100644
--- a/qlib/rl/order_execution/simulator_simple.py
+++ b/qlib/rl/order_execution/simulator_simple.py
@@ -79,8 +79,8 @@ def __init__(
         self,
         order: Order,
         data_dir: Path,
-        feature_columns_today: List[str] = [],
-        feature_columns_yesterday: List[str] = [],
+        feature_columns_today: List[str] = None,
+        feature_columns_yesterday: List[str] = None,
         data_granularity: int = 1,
         ticks_per_step: int = 30,
         vol_threshold: Optional[float] = None,
@@ -89,6 +89,11 @@ def __init__(
 
         assert ticks_per_step % data_granularity == 0
 
+        if feature_columns_today is None:
+            feature_columns_today = []
+        if feature_columns_yesterday is None:
+            feature_columns_yesterday = []
+
         self.order = order
         self.data_dir = data_dir
         self.feature_columns_today = feature_columns_today