Naveen Suda afdd4247a2 [torchao][pt2e] Make prepare and convert faster by caching (#162550) ...

Summary: D79674759 tried to fix the expensive prepare and convert steps, as `assert_and_get_unique_device` was called multiple times. This change fixes that issue by using `functools.cache` decorator.

Test Plan:
Verified on llm export to QNN.
LLM Quantization prepare time of ~20min reduced to ~3min.

Rollback Plan:

Differential Revision: D82073679

Pull Request resolved: https://github.com/pytorch/pytorch/pull/162550
Approved by: https://github.com/andrewor14

2025-09-11 07:59:22 +00:00

..

nn

[nn]: updated type alias for padddingmode in module/conv.py (#158843)

2025-07-25 23:05:02 +00:00

ns

[BE][13/16] fix typos in torch/ (torch/ao/) (#156603)

2025-06-29 04:34:04 +00:00

pruning

[BE]: ruff PLC0207 - use maxsplit kwarg (#160107)

2025-08-08 03:14:59 +00:00

quantization

[torchao][pt2e] Make prepare and convert faster by caching (#162550)

2025-09-11 07:59:22 +00:00

__init__.py

[BE][Easy] improve submodule discovery for torch.ao type annotations (#144680)

2025-01-13 17:16:19 +00:00