.export_helpers

experimental_experiment.export_helpers.torch_export(mod: Module, args: Tuple[Any, ...], kwargs: Mapping[str, Any] | None = None, *, dynamic_shapes: Dict[str, Any] | Tuple[Any, ...] | List[Any] | None = None, strict: bool = False, preserve_module_call_signature: Tuple[str, ...] = (), backed_size_oblivious: bool | str = False, prefer_deferred_runtime_asserts_over_guards: bool = False, verbose: int = 0, **other_kwargs)

Wrapper around torch.export.export(). backed_size_oblivious can be boolean then it calls torch.fx.experimental._config.patch(backed_size_oblivious=True) or not. It can be 'auto' to let select automatically the best mode. It can be 'half' to disable some non oblivious exceptions.