Improve evaluation api

mlflow/models/__init__.py

@@ -24,14 +24,13 @@
 from .model import Model
 from .flavor_backend import FlavorBackend
 from ..utils.environment import infer_pip_requirements
-from .evaluation import evaluate, EvaluationDataset
+from .evaluation import evaluate
 
 __all__ = [
     "Model",
     "FlavorBackend",
     "infer_pip_requirements",
     "evaluate",
-    "EvaluationDataset",
 ]
 
 

mlflow/models/evaluation/__init__.py

@@ -1,19 +1,17 @@
 from mlflow.models.evaluation.base import (
     ModelEvaluator,
-    EvaluationDataset,
     EvaluationResult,
-    EvaluationMetrics,
     EvaluationArtifact,
     evaluate,
     list_evaluators,
+    get_last_failed_evaluator,
 )
 
 __all__ = [
     "ModelEvaluator",
-    "EvaluationDataset",
     "EvaluationResult",
-    "EvaluationMetrics",
     "EvaluationArtifact",
     "evaluate",
     "list_evaluators",
+    "get_last_failed_evaluator",
 ]