pytorch

366f1b2c - [xnnpack][lite-int] Freeze/Inline module to remove reference to self (#88863)

[xnnpack][lite-int] Freeze/Inline module to remove reference to self (#88863) We need to inline graph before converting from torchscript to xnnpack flatubuffer. Remove graph dependence on self. This will later help us work with constant data. Differential Revision: [D41049858](https://our.internmc.facebook.com/intern/diff/D41049858/) Pull Request resolved: https://github.com/pytorch/pytorch/pull/88863 Approved by: https://github.com/digantdesai