[ez][inductor][fx pass] strengthen numerical check for batch fusion (#106744)
Summary:
As title.
For batch fusion, we use torch op to fuse and the result should be exactly same as the original ones.
pull request: https://github.com/pytorch/pytorch/pull/106731#issuecomment-1668662078
Test Plan:
```
buck test mode/dev-nosan //caffe2/test/inductor:group_batch_fusion
File changed: fbcode//caffe2/test/inductor/test_group_batch_fusion.py
File changed: fbsource//xplat/caffe2/test/inductor/test_group_batch_fusion.py
Buck UI: https://www.internalfb.com/buck2/cf14a2dd-faee-417a-8d26-0b9326c944e4
Test UI: https://www.internalfb.com/intern/testinfra/testrun/6755399617159540
Network: Up: 0B Down: 0B
Jobs completed: 12. Time elapsed: 2:55.5s.
Tests finished: Pass 4. Fail 0. Fatal 0. Skip 0. Build failure 0
```
Reviewed By: dshi7
Differential Revision: D48132255
@diff-train-skip-merge
Pull Request resolved: https://github.com/pytorch/pytorch/pull/106744
Approved by: https://github.com/kit1980