[CORE] Adding support for insertion of soft-tuned prompts (#4645)

Co-authored-by: Swapnil Parekh <swapnilp@ibm.com>
Co-authored-by: Joe G <joseph.granados@h2o.ai>
Co-authored-by: Antoni Baum <antoni.baum@protonmail.com>

tests/worker/test_model_runner.py

@@ -23,6 +23,7 @@ def _create_model_runner(model: str, *args, **kwargs) -> ModelRunner:
         cache_config=engine_config.cache_config,
         load_config=engine_config.load_config,
         lora_config=engine_config.lora_config,
+        prompt_adapter_config=engine_config.prompt_adapter_config,
         is_driver_worker=True,
     )
     return model_runner