transformers

e8cfae1d - Extend continuous batching memory prediction test to XPU (#47159)

Extend continuous batching memory prediction test to XPU (#47159) * Extend continuous batching memory prediction test to XPU Signed-off-by: Wang, Yi A <yi.a.wang@intel.com> * This PR updates the test to use the existing device-agnostic testing helpers `backend_empty_cache` and `backend_memory_allocated`, allowing the same memory prediction check to run on CUDA and XPU without hardcoded CUDA memory APIs. Signed-off-by: Wang, Yi A <yi.a.wang@intel.com> --------- Signed-off-by: Wang, Yi A <yi.a.wang@intel.com>