pytorch/test/quantization
Tugsbayasgalan Manlaibaatar 75ac6fdcdd Propogate dynamo shape_env to make_fx (#96437)
Currently, when we use assume_static_by_default flag, dynamo won't produce any symbols for input tensors. But when we pass the dynamo generated graph onto make_fx via torchdynamo.export(aten_graph=True), there is no way to pass this flag. We enable this by directly passing the fake tensors dynamo used to make_fx and call make_fx with "real" mode with fake tensors from dynamo.

Note that this is modified version of (https://github.com/pytorch/pytorch/pull/96143)

Differential Revision: [D44561753](https://our.internmc.facebook.com/intern/diff/D44561753)
Pull Request resolved: https://github.com/pytorch/pytorch/pull/96437
Approved by: https://github.com/jansel, https://github.com/ezyang
2023-04-04 20:37:30 +00:00
..
ao_migration ao migration: remove package test as this behavior is tested by other things (#94422) 2023-02-13 16:33:40 +00:00
bc [BE] [3/3] Rewrite super() calls in test (#94592) 2023-02-12 22:20:53 +00:00
core pt2e short term quant: respect qmin/qmax for linear weight (#96232) 2023-03-10 04:46:20 +00:00
eager [BE] [3/3] Rewrite super() calls in test (#94592) 2023-02-12 22:20:53 +00:00
fx Propogate dynamo shape_env to make_fx (#96437) 2023-04-04 20:37:30 +00:00
jit [BE] Move flatbuffer related python C bindings to script_init (#97476) 2023-03-28 17:56:32 +00:00
serialized [ao] fix incorrect integer cast on histogram observer bounds (#90355) 2022-12-12 20:30:44 +00:00
__init__.py