* Fix missing key for cpu_embedding * Remove 512 as it stuck for now * Small fix
* enable previous models * test mistral model * for test * run models separately * test all models * for test * revert the llm_performance_test.yaml