PrepareQuant step (#24425)
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/24425
- _jit_pass_prepare_quant: clone the observer module in argument and insert that to the
module we want to quantize, insert observer calls for the Tensor values we want to observe
Differential Revision: D16933120
fbshipit-source-id: 7248de6132429ba943a09831a76486f7a3cd52a3