Embedding_bag operator on GPU

Hello,

Nvidia MLPerf suggests to use [TensorRT](https://github.com/NVIDIA/TensorRT) framework for a performant inference deployment. For DLRM (DL based Recommendation Systems) inference on GPU, I have the following questions:

* Does TensorRT modify the backend (CUDA/C++ source code) of Embedding bag operator or it uses the exact same vanilla PyTorch CUDA kernels?

* What are the benefits of using vanilla PyTorch over TensorRT for DLRM inference?

Please let me know your comments. Thanks

The issue names no repository files, tests, or entry points. Start by locating the EmbeddingBag GPU implementations used by PyTorch and TensorRT, then compare their DLRM inference behavior and performance. Done means documenting whether the kernels differ and the practical trade-offs between vanilla PyTorch and TensorRT.

Written by the indexing model from the issue text.