diff --git a/GPT_SoVITS/s2_train.py b/GPT_SoVITS/s2_train.py
index cb978181..dc9e7643 100644
--- a/GPT_SoVITS/s2_train.py
+++ b/GPT_SoVITS/s2_train.py
@@ -17,7 +17,7 @@ import torch.multiprocessing as mp
 from torch.cuda.amp import GradScaler
 musa_ddp = False
 if musa_utils.is_available():
-    os.environ["MUSA_LAUNCH_BLOCKING"] = 1
+    os.environ["MUSA_LAUNCH_BLOCKING"] = "1"
     autocast = torch.musa.amp.autocast
     musa_ddp = musa_utils.should_ddp()
 elif torch.cuda.is_available():
diff --git a/musa_accelerator.py b/musa_accelerator.py
index 4512bc25..249e152c 100644
--- a/musa_accelerator.py
+++ b/musa_accelerator.py
@@ -18,6 +18,8 @@ from lightning_fabric.utilities.types import _DEVICE
 from pytorch_lightning.accelerators.accelerator import Accelerator
 from pytorch_lightning.utilities.exceptions import MisconfigurationException
 
+from musa_utils import _musa_available
+
 import logging
 _log = logging.getLogger(__name__)
 
diff --git a/musa_utils.py b/musa_utils.py
index c5200023..ebc14ac4 100644
--- a/musa_utils.py
+++ b/musa_utils.py
@@ -68,8 +68,7 @@ def get_device_dtype(device_idx: int = 0) -> tuple[torch.dtype, float, float]:
     return dtype, version_val, mem_gb
 
 def should_ddp(device_idx: int = 0) -> bool:
-    device_name = torch.musa.get_device_properties(device_idx).name
-    numbers_in_name = [int(n) for n in re.findall(r'\d+', device_name)]
+    _, version_val, _ = get_device_dtype(device_idx)
     if version_val <= 2.1:
         return False
     else: