[quant] Add benchmakrs for embedding_bag coversion ops (#43291)
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/43291
Test Float2Fused and Fused2Float conversion operators for embedding_bag byte and 4-bit ops
Test Plan:
```
python -m pt.qembedding_pack_tes
```
Imported from OSS
Reviewed By: radkris-git
Differential Revision: D23231641
fbshipit-source-id: a2afe51bba52980d2e96dfd7dbc183327e9349fd
