Fix replay buffer device at load time

sbx/common/off_policy_algorithm.py

@@ -1,3 +1,5 @@
+import io
+import pathlib
 from typing import Any, Dict, List, Optional, Tuple, Type, Union
 
 import jax
@@ -9,6 +11,7 @@
 from stable_baselines3.common.off_policy_algorithm import OffPolicyAlgorithm
 from stable_baselines3.common.policies import BasePolicy
 from stable_baselines3.common.type_aliases import GymEnv, Schedule
+from stable_baselines3.common.utils import get_device
 
 
 class OffPolicyAlgorithmJax(OffPolicyAlgorithm):
@@ -116,3 +119,13 @@ def _setup_model(self) -> None:
         )
         # Convert train freq parameter to TrainFreq object
         self._convert_train_freq()
+
+    def load_replay_buffer(
+        self,
+        path: Union[str, pathlib.Path, io.BufferedIOBase],
+        truncate_last_traj: bool = True,
+    ) -> None:
+        super().load_replay_buffer(path, truncate_last_traj)
+        # Override replay buffer device to be always cpu for conversion to numpy
+        assert self.replay_buffer is not None
+        self.replay_buffer.device = get_device("cpu")

sbx/version.txt

@@ -1 +1 @@
-0.9.0
+0.9.1

tests/test_buffers.py

@@ -0,0 +1,14 @@
+import pytest
+import torch as th
+
+from sbx import SAC
+
+
+def test_force_cpu_device(tmp_path):
+    if not th.cuda.is_available():
+        pytest.skip("No CUDA device")
+    model = SAC("MlpPolicy", "Pendulum-v1", buffer_size=200)
+    assert model.replay_buffer.device == th.device("cpu")
+    model.save_replay_buffer(tmp_path / "replay")
+    model.load_replay_buffer(tmp_path / "replay")
+    assert model.replay_buffer.device == th.device("cpu")