[functorch] refactor: get_exhaustive_batched_inputs (#85965)
`get_exhaustive_batched_inputs_batch_norm_is_training` and `get_exhaustive_batched_inputs` are same except for a couple of lines.
We move the above functionality into `generate_vmap_inputs` (which is now only function to create batched inputs)
Pull Request resolved: https://github.com/pytorch/pytorch/pull/85965
Approved by: https://github.com/zou3519
