Revert "Make torch_geometric models compatible with export (#123403)"…

… (#128377)

Summary:
This reverts commit d78991a7381adb3df5e9b63c365db4506643edce.

This PR reverts pytorch/pytorch#123403 to fix the performance regression as discussed in pytorch/pytorch#127513 (comment).

X-link: pytorch/pytorch#128377
Approved by: https://github.com/jgong5, https://github.com/angelayi, https://github.com/desertfire

Reviewed By: clee2000

Differential Revision: D58501783

fbshipit-source-id: 1e55cc2c0b315ed6869195ee6730e72eb6be9da9

userbenchmark/dynamo/dynamobench/common.py

@@ -1184,12 +1184,14 @@ def load(cls, model, example_inputs, device):
  else:
  _register_dataclass_output_as_pytree(example_outputs)
 
- gm = torch.export._trace._export(
+ # TODO(angelayi): change this to predispatch
+ # https://github.com/pytorch/pytorch/issues/127513 needs to be fixed before changing
+ # to predispatch to avoid performance regressions
+ gm = torch.export._trace._export_to_torch_ir(
  model,
  example_args,
  example_kwargs,
- pre_dispatch=True,
- ).module()
+ )
  with torch.no_grad():
  so_path = torch._inductor.aot_compile(
  gm, example_args, example_kwargs