Fix TuningStep validation error by adding V3 HyperparameterTuner duck typing support (#5344)

Author: aviruthen

sagemaker-core/src/sagemaker/core/workflow/pipeline_context.py

@@ -382,11 +382,11 @@ def retrieve_caller_name(job_instance):
     if isinstance(job_instance, Transformer):
         return "transform"
 
-    # Duck typing for HyperparameterTuner: has 'fit' method and 'best_estimator' method
-    # This avoids importing from sagemaker.train which would violate architecture
-    if hasattr(job_instance, 'fit') and hasattr(job_instance, 'best_estimator'):
+    # Duck typing for HyperparameterTuner: has 'tune' method and 'model_trainer' attribute
+    # This covers both V2 (fit/best_estimator) and V3 (tune/model_trainer) implementations
+    if (hasattr(job_instance, 'fit') and hasattr(job_instance, 'best_estimator')) or \
+       (hasattr(job_instance, 'tune') and hasattr(job_instance, 'model_trainer')):
         return "tune"
-    
     # if isinstance(job_instance, AutoML):
     #     return "auto_ml"