[BugFix] Use correct default cuda device

ghstack-source-id: 9afb5b03ddf75afec357e9e54caadfc92ebf4ded Pull Request resolved: #1161
pytorch · Jan 7, 2025 · ecb692e · ecb692e
1 parent 70d4ed1
commit ecb692e
Showing 1 changed file with 1 addition and 3 deletions.
diff --git a/tensordict/utils.py b/tensordict/utils.py
@@ -1373,8 +1373,6 @@ def _parse_to(*args, **kwargs):
             *args, **kwargs
         )
     else:
-        device = None
-        dtype = None
         non_blocking = kwargs.get("non_blocking", False)
         convert_to_format = kwargs.get("convert_to_format")
         if len(args) > 0:
@@ -1390,7 +1388,7 @@ def _parse_to(*args, **kwargs):
             device = torch.device(device)
 
     if device and device.type == "cuda" and device.index is None:
-        device = torch.device("cuda:0")
+        device = torch.device(f"cuda:{torch.cuda.current_device()}")
 
     if other is not None:
         if device is not None and device != other.device: