Back out "[quant][graphmode][fx] Separate handling Copy operator to a helper function" (#55388)
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/55388
temporarily revert D27314678 (https://github.com/pytorch/pytorch/commit/c57541ce060e2043d0b30b3b382e2da7c1355695), it appears to cause a perf regression that makes quantization of some models take too long to complete tests.
Reviewed By: houseroad
Differential Revision: D27583809
fbshipit-source-id: e9c088ccbfd3bfb3a1d4c7eafee3eca29ee7717b
