[ROCm] enabling miopen_batch_norm lowering in inductor (#105740)
Enabling miopen_batch_norm lowering for inductor only.
This is to avoid errors observed in some models and perf difference is very close from initial benchmarks.
```
LoweringException: RuntimeError: Expected contiguous tensor, but got non-contiguous tensor for argument #1 'input' (while checking arguments for miopen_batch_norm)
target: aten.miopen_batch_norm.default
```
Pull Request resolved: https://github.com/pytorch/pytorch/pull/105740
Approved by: https://github.com/jithunnair-amd, https://github.com/malfet
