Change compile for pipeline module torch.compile

We have encountered and issue with torch.compile and the pipeline
module. modifying a member of the module duing the run will cause
torch compile to restart the analysis and treat the module as dynamic.
this happens because the fwd function will modify the micro_offset
attribute of the pipeline module.
in order to bypass this issue without significantly changing the way
the pipeline module works we propose to compile only the layers in the
pipeline module instead of the pipeline module itslef.
this will bypass the issue, and should still give most of the benefit of
torch compiling the pipeline module while avoiding the issue.

deepspeed/runtime/pipe/module.py

@@ -662,3 +662,8 @@ def get_additional_losses(self):
          Return a dictionary of {"loss name": loss_value} or None if no additional losses.
         """
         return None
+
+    def compile(self, *args, **kwargs):
+        for idx, layer in enumerate(self.forward_funcs):
+            new_layer = torch.compile(layer, *args, **kwargs)
+            self.forward_funcs[idx] = new_layer