Tugsbayasgalan Manlaibaatar f3c3f3a3c3 Fix assigning tensor with requires_grad as constant in export (#137997) ... When we insert cojstants into unlifted graph, we need to detach them if they require grad BUT when we detach we need to preserve the original aliasing information. Differential Revision: [D64406859](https://our.internmc.facebook.com/intern/diff/D64406859/) Pull Request resolved: https://github.com/pytorch/pytorch/pull/137997 Approved by: https://github.com/avikchaudhuri		2024-10-17 06:41:10 +00:00
..
experimental	Deprecate _preserve_ops and consolidate with decomp_table (#135080)	2024-09-15 17:01:58 +00:00
passes	[export] Make move_to_device_pass function public (#134263)	2024-08-23 23:18:30 +00:00
__init__.py	Create export_for_inference API and expose core_aten as public facing API (#135912)	2024-09-15 17:05:07 +00:00
_remove_auto_functionalized_pass.py	Redesign custom op functionlaization for better re-inplace (#134409)	2024-09-04 17:08:58 +00:00
_remove_effect_tokens_pass.py	[export][fx] More robust DCE pass (#132764)	2024-08-06 22:27:22 +00:00
_safeguard.py	Flip default value for mypy disallow_untyped_defs [6/11] (#127843)	2024-06-08 18:49:29 +00:00
_swap.py	[export] Move swap to a different file (#137134)	2024-10-06 04:28:18 +00:00
_trace.py	tensor constant with wrapped method (#138091)	2024-10-17 00:00:04 +00:00
_tree_utils.py	[export] do not use tree_flatten_spec (#118608)	2024-01-30 19:14:04 +00:00
_unlift.py	Fix assigning tensor with requires_grad as constant in export (#137997)	2024-10-17 06:41:10 +00:00
custom_obj.py	[Export] Introduce class_fqn into CustomObjArgument (#118158)	2024-01-25 18:44:25 +00:00
dynamic_shapes.py	[export] clean up dynamic markers from tensors (#137230)	2024-10-04 06:50:45 +00:00
exported_program.py	Try unify training and inference (#136888)	2024-10-11 20:09:58 +00:00
graph_signature.py	[export] refactor ExportGraphSignature construction (#134059)	2024-08-23 23:29:28 +00:00
unflatten.py	[alt] fix unroll in successive unflatten (#137646)	2024-10-12 15:53:52 +00:00