[FSDP2] Gate PT2 code for torch deploy (pytorch#142456)

See diff for internal details

Differential Revision: [D67003832](https://our.internmc.facebook.com/intern/diff/D67003832)
Pull Request resolved: pytorch#142456
Approved by: https://github.com/yf225, https://github.com/weifengpy, https://github.com/fegin

Author: Andrew Gu

torch/distributed/fsdp/_fully_shard/_fsdp_param.py

@@ -136,7 +136,8 @@ def copy__functionalize(tensor, data):
         torch.ops.fsdp.copy_.default(tensor_inner, data_inner)
 
 
-torch.fx.node.has_side_effect(torch.ops.fsdp.copy_.default)
+if not torch._running_with_deploy():
+    torch.fx.node.has_side_effect(torch.ops.fsdp.copy_.default)
 
 
 class ShardedState(Enum):