snowflake-ml-python 1.6.0__py3-none-any.whl → 1.6.2__py3-none-any.whl

snowflake/ml/model/_packager/model_handlers/snowmlmodel.py

@@ -1,20 +1,27 @@
 import os
 import warnings
-from typing import TYPE_CHECKING, Callable, Dict, Optional, Type, cast, final
+from typing import TYPE_CHECKING, Any, Callable, Dict, Optional, Type, cast, final
 
 import cloudpickle
 import numpy as np
 import pandas as pd
+from packaging import version
 from typing_extensions import TypeGuard, Unpack
 
 from snowflake.ml._internal import type_utils
+from snowflake.ml._internal.exceptions import exceptions
 from snowflake.ml.model import custom_model, model_signature, type_hints as model_types
 from snowflake.ml.model._packager.model_env import model_env
-from snowflake.ml.model._packager.model_handlers import _base
+from snowflake.ml.model._packager.model_handlers import (
+    _base,
+    _utils as handlers_utils,
+    model_objective_utils,
+)
 from snowflake.ml.model._packager.model_handlers_migrator import base_migrator
 from snowflake.ml.model._packager.model_meta import (
     model_blob_meta,
     model_meta as model_meta_api,
+    model_meta_schema,
 )
 from snowflake.ml.model._signatures import numpy_handler, utils as model_signature_utils
 
@@ -62,6 +69,53 @@ class SnowMLModelHandler(_base.BaseModelHandler["BaseEstimator"]):
 
         return cast("BaseEstimator", model)
 
+    @classmethod
+    def _get_local_version_package(cls, pkg_name: str) -> Optional[version.Version]:
+        import importlib_metadata
+        from packaging import version
+
+        local_version = None
+
+        try:
+            local_dist = importlib_metadata.distribution(pkg_name)  # type: ignore[no-untyped-call]
+            local_version = version.parse(local_dist.version)
+        except importlib_metadata.PackageNotFoundError:
+            pass
+
+        return local_version
+
+    @classmethod
+    def _can_support_xgb(cls, enable_explainability: Optional[bool]) -> bool:
+
+        local_xgb_version = cls._get_local_version_package("xgboost")
+
+        if local_xgb_version and local_xgb_version >= version.parse("2.1.0"):
+            if enable_explainability:
+                warnings.warn(
+                    f"This version of xgboost {local_xgb_version} does not work with shap 0.42.1."
+                    + "If you want model explanations, lower the xgboost version to <2.1.0.",
+                    category=UserWarning,
+                    stacklevel=1,
+                )
+            return False
+        return True
+
+    @classmethod
+    def _get_supported_object_for_explainability(
+        cls, estimator: "BaseEstimator", enable_explainability: Optional[bool]
+    ) -> Any:
+        methods = ["to_xgboost", "to_lightgbm"]
+        for method_name in methods:
+            if hasattr(estimator, method_name):
+                try:
+                    result = getattr(estimator, method_name)()
+                    if method_name == "to_xgboost" and not cls._can_support_xgb(enable_explainability):
+                        return None
+                    return result
+                except exceptions.SnowflakeMLException:
+                    pass  # Do nothing and continue to the next method
+        return None
+
     @classmethod
     def save_model(
         cls,
@@ -73,6 +127,9 @@ class SnowMLModelHandler(_base.BaseModelHandler["BaseEstimator"]):
         is_sub_model: Optional[bool] = False,
         **kwargs: Unpack[model_types.SNOWModelSaveOptions],
     ) -> None:
+
+        enable_explainability = kwargs.get("enable_explainability", None)
+
         from snowflake.ml.modeling.framework.base import BaseEstimator
 
         assert isinstance(model, BaseEstimator)
@@ -101,15 +158,35 @@ class SnowMLModelHandler(_base.BaseModelHandler["BaseEstimator"]):
                         raise ValueError(f"Target method {method_name} does not exist in the model.")
                 model_meta.signatures = temp_model_signature_dict
 
+        if enable_explainability or enable_explainability is None:
+            python_base_obj = cls._get_supported_object_for_explainability(model, enable_explainability)
+            if python_base_obj is None:
+                if enable_explainability:  # if user set enable_explainability to True, throw error else silently skip
+                    raise ValueError("Explain only support for xgboost or lightgbm Snowpark ML models.")
+                # set None to False so we don't include shap in the environment
+                enable_explainability = False
+            else:
+                model_objective_and_output_type = model_objective_utils.get_model_objective_and_output_type(
+                    python_base_obj
+                )
+                model_meta.model_objective = model_objective_and_output_type.objective
+                model_meta = handlers_utils.add_explain_method_signature(
+                    model_meta=model_meta,
+                    explain_method="explain",
+                    target_method="predict",
+                    output_return_type=model_objective_and_output_type.output_type,
+                )
+                enable_explainability = True
+
         model_blob_path = os.path.join(model_blobs_dir_path, name)
         os.makedirs(model_blob_path, exist_ok=True)
-        with open(os.path.join(model_blob_path, cls.MODELE_BLOB_FILE_OR_DIR), "wb") as f:
+        with open(os.path.join(model_blob_path, cls.MODEL_BLOB_FILE_OR_DIR), "wb") as f:
             cloudpickle.dump(model, f)
         base_meta = model_blob_meta.ModelBlobMeta(
             name=name,
             model_type=cls.HANDLER_TYPE,
             handler_version=cls.HANDLER_VERSION,
-            path=cls.MODELE_BLOB_FILE_OR_DIR,
+            path=cls.MODEL_BLOB_FILE_OR_DIR,
         )
         model_meta.models[name] = base_meta
         model_meta.min_snowpark_ml_version = cls._MIN_SNOWPARK_ML_VERSION
@@ -118,7 +195,29 @@ class SnowMLModelHandler(_base.BaseModelHandler["BaseEstimator"]):
         model_dependencies = model._get_dependencies()
         for dep in model_dependencies:
             pkg_name = dep.split("==")[0]
-            _include_if_absent_pkgs.append(model_env.ModelDependency(requirement=pkg_name, pip_name=pkg_name))
+            if pkg_name != "xgboost":
+                _include_if_absent_pkgs.append(model_env.ModelDependency(requirement=pkg_name, pip_name=pkg_name))
+                continue
+
+            local_xgb_version = cls._get_local_version_package("xgboost")
+            if local_xgb_version and local_xgb_version >= version.parse("2.0.0") and enable_explainability:
+                model_meta.env.include_if_absent(
+                    [
+                        model_env.ModelDependency(requirement="xgboost==2.0.*", pip_name="xgboost"),
+                    ],
+                    check_local_version=False,
+                )
+            else:
+                model_meta.env.include_if_absent(
+                    [
+                        model_env.ModelDependency(requirement="xgboost", pip_name="xgboost"),
+                    ],
+                    check_local_version=True,
+                )
+
+        if enable_explainability:
+            model_meta.env.include_if_absent([model_env.ModelDependency(requirement="shap", pip_name="shap")])
+            model_meta.explain_algorithm = model_meta_schema.ModelExplainAlgorithm.SHAP
         model_meta.env.include_if_absent(_include_if_absent_pkgs, check_local_version=True)
 
     @classmethod
@@ -146,6 +245,7 @@ class SnowMLModelHandler(_base.BaseModelHandler["BaseEstimator"]):
         cls,
         raw_model: "BaseEstimator",
         model_meta: model_meta_api.ModelMetadata,
+        background_data: Optional[pd.DataFrame] = None,
         **kwargs: Unpack[model_types.SNOWModelLoadOptions],
     ) -> custom_model.CustomModel:
         from snowflake.ml.model import custom_model
@@ -172,6 +272,24 @@ class SnowMLModelHandler(_base.BaseModelHandler["BaseEstimator"]):
 
                     return model_signature_utils.rename_pandas_df(df, signature.outputs)
 
+                @custom_model.inference_api
+                def explain_fn(self: custom_model.CustomModel, X: pd.DataFrame) -> pd.DataFrame:
+                    import shap
+
+                    methods = ["to_xgboost", "to_lightgbm"]
+                    for method_name in methods:
+                        try:
+                            base_model = getattr(raw_model, method_name)()
+                            explainer = shap.TreeExplainer(base_model)
+                            df = pd.DataFrame(explainer(X).values)
+                            return model_signature_utils.rename_pandas_df(df, signature.outputs)
+                        except exceptions.SnowflakeMLException:
+                            pass  # Do nothing and continue to the next method
+                    raise ValueError("The model must be an xgboost or lightgbm estimator.")
+
+                if target_method == "explain":
+                    return explain_fn
+
                 return fn
 
             type_method_dict = {}

snowflake/ml/model/_packager/model_handlers/tensorflow.py

@@ -36,7 +36,7 @@ class TensorFlowHandler(_base.BaseModelHandler["tensorflow.Module"]):
     _MIN_SNOWPARK_ML_VERSION = "1.0.12"
     _HANDLER_MIGRATOR_PLANS: Dict[str, Type[base_migrator.BaseModelHandlerMigrator]] = {}
 
-    MODELE_BLOB_FILE_OR_DIR = "model"
+    MODEL_BLOB_FILE_OR_DIR = "model"
     DEFAULT_TARGET_METHODS = ["__call__"]
 
     @classmethod
@@ -68,6 +68,10 @@ class TensorFlowHandler(_base.BaseModelHandler["tensorflow.Module"]):
         is_sub_model: Optional[bool] = False,
         **kwargs: Unpack[model_types.TensorflowSaveOptions],
     ) -> None:
+        enable_explainability = kwargs.get("enable_explainability", False)
+        if enable_explainability:
+            raise NotImplementedError("Explainability is not supported for Tensorflow model.")
+
         import tensorflow
 
         assert isinstance(model, tensorflow.Module)
@@ -114,15 +118,15 @@ class TensorFlowHandler(_base.BaseModelHandler["tensorflow.Module"]):
         model_blob_path = os.path.join(model_blobs_dir_path, name)
         os.makedirs(model_blob_path, exist_ok=True)
         if isinstance(model, tensorflow.keras.Model):
-            tensorflow.keras.models.save_model(model, os.path.join(model_blob_path, cls.MODELE_BLOB_FILE_OR_DIR))
+            tensorflow.keras.models.save_model(model, os.path.join(model_blob_path, cls.MODEL_BLOB_FILE_OR_DIR))
         else:
-            tensorflow.saved_model.save(model, os.path.join(model_blob_path, cls.MODELE_BLOB_FILE_OR_DIR))
+            tensorflow.saved_model.save(model, os.path.join(model_blob_path, cls.MODEL_BLOB_FILE_OR_DIR))
 
         base_meta = model_blob_meta.ModelBlobMeta(
             name=name,
             model_type=cls.HANDLER_TYPE,
             handler_version=cls.HANDLER_VERSION,
-            path=cls.MODELE_BLOB_FILE_OR_DIR,
+            path=cls.MODEL_BLOB_FILE_OR_DIR,
         )
         model_meta.models[name] = base_meta
         model_meta.min_snowpark_ml_version = cls._MIN_SNOWPARK_ML_VERSION
@@ -156,6 +160,7 @@ class TensorFlowHandler(_base.BaseModelHandler["tensorflow.Module"]):
         cls,
         raw_model: "tensorflow.Module",
         model_meta: model_meta_api.ModelMetadata,
+        background_data: Optional[pd.DataFrame] = None,
         **kwargs: Unpack[model_types.TensorflowLoadOptions],
     ) -> custom_model.CustomModel:
         import tensorflow

snowflake/ml/model/_packager/model_handlers/torchscript.py

@@ -34,7 +34,7 @@ class TorchScriptHandler(_base.BaseModelHandler["torch.jit.ScriptModule"]):  # t
     _MIN_SNOWPARK_ML_VERSION = "1.0.12"
     _HANDLER_MIGRATOR_PLANS: Dict[str, Type[base_migrator.BaseModelHandlerMigrator]] = {}
 
-    MODELE_BLOB_FILE_OR_DIR = "model.pt"
+    MODEL_BLOB_FILE_OR_DIR = "model.pt"
     DEFAULT_TARGET_METHODS = ["forward"]
 
     @classmethod
@@ -66,6 +66,10 @@ class TorchScriptHandler(_base.BaseModelHandler["torch.jit.ScriptModule"]):  # t
         is_sub_model: Optional[bool] = False,
         **kwargs: Unpack[model_types.TorchScriptSaveOptions],
     ) -> None:
+        enable_explainability = kwargs.get("enable_explainability", False)
+        if enable_explainability:
+            raise NotImplementedError("Explainability is not supported for Torch Script model.")
+
         import torch
 
         assert isinstance(model, torch.jit.ScriptModule)  # type:ignore[attr-defined]
@@ -106,13 +110,13 @@ class TorchScriptHandler(_base.BaseModelHandler["torch.jit.ScriptModule"]):  # t
 
         model_blob_path = os.path.join(model_blobs_dir_path, name)
         os.makedirs(model_blob_path, exist_ok=True)
-        with open(os.path.join(model_blob_path, cls.MODELE_BLOB_FILE_OR_DIR), "wb") as f:
-            torch.jit.save(model, f)  # type:ignore[attr-defined]
+        with open(os.path.join(model_blob_path, cls.MODEL_BLOB_FILE_OR_DIR), "wb") as f:
+            torch.jit.save(model, f)  # type:ignore[no-untyped-call, attr-defined]
         base_meta = model_blob_meta.ModelBlobMeta(
             name=name,
             model_type=cls.HANDLER_TYPE,
             handler_version=cls.HANDLER_VERSION,
-            path=cls.MODELE_BLOB_FILE_OR_DIR,
+            path=cls.MODEL_BLOB_FILE_OR_DIR,
         )
         model_meta.models[name] = base_meta
         model_meta.min_snowpark_ml_version = cls._MIN_SNOWPARK_ML_VERSION
@@ -137,7 +141,7 @@ class TorchScriptHandler(_base.BaseModelHandler["torch.jit.ScriptModule"]):  # t
         model_blob_metadata = model_blobs_metadata[name]
         model_blob_filename = model_blob_metadata.path
         with open(os.path.join(model_blob_path, model_blob_filename), "rb") as f:
-            m = torch.jit.load(  # type:ignore[attr-defined]
+            m = torch.jit.load(  # type:ignore[no-untyped-call, attr-defined]
                 f, map_location="cuda" if kwargs.get("use_gpu", False) else "cpu"
             )
         assert isinstance(m, torch.jit.ScriptModule)  # type:ignore[attr-defined]
@@ -152,6 +156,7 @@ class TorchScriptHandler(_base.BaseModelHandler["torch.jit.ScriptModule"]):  # t
         cls,
         raw_model: "torch.jit.ScriptModule",  # type:ignore[name-defined]
         model_meta: model_meta_api.ModelMetadata,
+        background_data: Optional[pd.DataFrame] = None,
         **kwargs: Unpack[model_types.TorchScriptLoadOptions],
     ) -> custom_model.CustomModel:
         from snowflake.ml.model import custom_model

snowflake/ml/model/_packager/model_handlers/xgboost.py

@@ -1,6 +1,6 @@
 # mypy: disable-error-code="import"
-import json
 import os
+import warnings
 from typing import (
     TYPE_CHECKING,
     Any,
@@ -13,14 +13,20 @@ from typing import (
     final,
 )
 
+import importlib_metadata
 import numpy as np
 import pandas as pd
+from packaging import version
 from typing_extensions import TypeGuard, Unpack
 
 from snowflake.ml._internal import type_utils
 from snowflake.ml.model import custom_model, model_signature, type_hints as model_types
 from snowflake.ml.model._packager.model_env import model_env
-from snowflake.ml.model._packager.model_handlers import _base, _utils as handlers_utils
+from snowflake.ml.model._packager.model_handlers import (
+    _base,
+    _utils as handlers_utils,
+    model_objective_utils,
+)
 from snowflake.ml.model._packager.model_handlers_migrator import base_migrator
 from snowflake.ml.model._packager.model_meta import (
     model_blob_meta,
@@ -45,41 +51,8 @@ class XGBModelHandler(_base.BaseModelHandler[Union["xgboost.Booster", "xgboost.X
     _MIN_SNOWPARK_ML_VERSION = "1.0.12"
     _HANDLER_MIGRATOR_PLANS: Dict[str, Type[base_migrator.BaseModelHandlerMigrator]] = {}
 
-    MODELE_BLOB_FILE_OR_DIR = "model.ubj"
+    MODEL_BLOB_FILE_OR_DIR = "model.ubj"
     DEFAULT_TARGET_METHODS = ["predict", "predict_proba"]
-    _BINARY_CLASSIFICATION_OBJECTIVE_PREFIX = ["binary:"]
-    _MULTI_CLASSIFICATION_OBJECTIVE_PREFIX = ["multi:"]
-    _RANKING_OBJECTIVE_PREFIX = ["rank:"]
-    _REGRESSION_OBJECTIVE_PREFIX = ["reg:"]
-
-    @classmethod
-    def get_model_objective(cls, model: Union["xgboost.Booster", "xgboost.XGBModel"]) -> _base.ModelObjective:
-        import xgboost
-
-        if isinstance(model, xgboost.XGBClassifier) or isinstance(model, xgboost.XGBRFClassifier):
-            num_classes = handlers_utils.get_num_classes_if_exists(model)
-            if num_classes == 2:
-                return _base.ModelObjective.BINARY_CLASSIFICATION
-            return _base.ModelObjective.MULTI_CLASSIFICATION
-        if isinstance(model, xgboost.XGBRegressor) or isinstance(model, xgboost.XGBRFRegressor):
-            return _base.ModelObjective.REGRESSION
-        if isinstance(model, xgboost.XGBRanker):
-            return _base.ModelObjective.RANKING
-        model_params = json.loads(model.save_config())
-        model_objective = model_params["learner"]["objective"]
-        for classification_objective in cls._BINARY_CLASSIFICATION_OBJECTIVE_PREFIX:
-            if classification_objective in model_objective:
-                return _base.ModelObjective.BINARY_CLASSIFICATION
-        for classification_objective in cls._MULTI_CLASSIFICATION_OBJECTIVE_PREFIX:
-            if classification_objective in model_objective:
-                return _base.ModelObjective.MULTI_CLASSIFICATION
-        for ranking_objective in cls._RANKING_OBJECTIVE_PREFIX:
-            if ranking_objective in model_objective:
-                return _base.ModelObjective.RANKING
-        for regression_objective in cls._REGRESSION_OBJECTIVE_PREFIX:
-            if regression_objective in model_objective:
-                return _base.ModelObjective.REGRESSION
-        return _base.ModelObjective.UNKNOWN
 
     @classmethod
     def can_handle(
@@ -114,10 +87,29 @@ class XGBModelHandler(_base.BaseModelHandler[Union["xgboost.Booster", "xgboost.X
         is_sub_model: Optional[bool] = False,
         **kwargs: Unpack[model_types.XGBModelSaveOptions],
     ) -> None:
+        enable_explainability = kwargs.get("enable_explainability", True)
+
         import xgboost
 
         assert isinstance(model, xgboost.Booster) or isinstance(model, xgboost.XGBModel)
 
+        local_xgb_version = None
+
+        try:
+            local_dist = importlib_metadata.distribution("xgboost")  # type: ignore[no-untyped-call]
+            local_xgb_version = version.parse(local_dist.version)
+        except importlib_metadata.PackageNotFoundError:
+            pass
+
+        if local_xgb_version and local_xgb_version >= version.parse("2.1.0") and enable_explainability:
+            warnings.warn(
+                f"This version of xgboost {local_xgb_version} does not work with shap 0.42.1."
+                + "If you want model explanations, lower the xgboost version to <2.1.0.",
+                category=UserWarning,
+                stacklevel=1,
+            )
+            enable_explainability = False
+
         if not is_sub_model:
             target_methods = handlers_utils.get_target_methods(
                 model=model,
@@ -146,25 +138,29 @@ class XGBModelHandler(_base.BaseModelHandler[Union["xgboost.Booster", "xgboost.X
                 sample_input_data=sample_input_data,
                 get_prediction_fn=get_prediction,
             )
-            if kwargs.get("enable_explainability", False):
-                output_type = model_signature.DataType.DOUBLE
-                if cls.get_model_objective(model) == _base.ModelObjective.MULTI_CLASSIFICATION:
-                    output_type = model_signature.DataType.STRING
+            model_objective_and_output = model_objective_utils.get_model_objective_and_output_type(model)
+            model_meta.model_objective = handlers_utils.validate_model_objective(
+                model_meta.model_objective, model_objective_and_output.objective
+            )
+            if enable_explainability:
                 model_meta = handlers_utils.add_explain_method_signature(
                     model_meta=model_meta,
                     explain_method="explain",
                     target_method="predict",
-                    output_return_type=output_type,
+                    output_return_type=model_objective_and_output.output_type,
                 )
+                model_meta.function_properties = {
+                    "explain": {model_meta_schema.FunctionProperties.PARTITIONED.value: False}
+                }
 
         model_blob_path = os.path.join(model_blobs_dir_path, name)
         os.makedirs(model_blob_path, exist_ok=True)
-        model.save_model(os.path.join(model_blob_path, cls.MODELE_BLOB_FILE_OR_DIR))
+        model.save_model(os.path.join(model_blob_path, cls.MODEL_BLOB_FILE_OR_DIR))
         base_meta = model_blob_meta.ModelBlobMeta(
             name=name,
             model_type=cls.HANDLER_TYPE,
             handler_version=cls.HANDLER_VERSION,
-            path=cls.MODELE_BLOB_FILE_OR_DIR,
+            path=cls.MODEL_BLOB_FILE_OR_DIR,
             options=model_meta_schema.XgboostModelBlobOptions({"xgb_estimator_type": model.__class__.__name__}),
         )
         model_meta.models[name] = base_meta
@@ -173,15 +169,27 @@ class XGBModelHandler(_base.BaseModelHandler[Union["xgboost.Booster", "xgboost.X
         model_meta.env.include_if_absent(
             [
                 model_env.ModelDependency(requirement="scikit-learn", pip_name="scikit-learn"),
-                model_env.ModelDependency(requirement="xgboost", pip_name="xgboost"),
             ],
             check_local_version=True,
         )
-        if kwargs.get("enable_explainability", False):
+        if local_xgb_version and local_xgb_version >= version.parse("2.0.0") and enable_explainability:
             model_meta.env.include_if_absent(
-                [model_env.ModelDependency(requirement="shap", pip_name="shap")],
+                [
+                    model_env.ModelDependency(requirement="xgboost==2.0.*", pip_name="xgboost"),
+                ],
+                check_local_version=False,
+            )
+        else:
+            model_meta.env.include_if_absent(
+                [
+                    model_env.ModelDependency(requirement="xgboost", pip_name="xgboost"),
+                ],
                 check_local_version=True,
             )
+
+        if enable_explainability:
+            model_meta.env.include_if_absent([model_env.ModelDependency(requirement="shap", pip_name="shap")])
+            model_meta.explain_algorithm = model_meta_schema.ModelExplainAlgorithm.SHAP
         model_meta.env.cuda_version = kwargs.get("cuda_version", model_env.DEFAULT_CUDA_VERSION)
 
     @classmethod
@@ -224,6 +232,7 @@ class XGBModelHandler(_base.BaseModelHandler[Union["xgboost.Booster", "xgboost.X
         cls,
         raw_model: Union["xgboost.Booster", "xgboost.XGBModel"],
         model_meta: model_meta_api.ModelMetadata,
+        background_data: Optional[pd.DataFrame] = None,
         **kwargs: Unpack[model_types.XGBModelLoadOptions],
     ) -> custom_model.CustomModel:
         import xgboost
@@ -260,7 +269,7 @@ class XGBModelHandler(_base.BaseModelHandler[Union["xgboost.Booster", "xgboost.X
                     import shap
 
                     explainer = shap.TreeExplainer(raw_model)
-                    df = pd.DataFrame(explainer(X).values)
+                    df = handlers_utils.convert_explanations_to_2D_df(raw_model, explainer(X).values)
                     return model_signature_utils.rename_pandas_df(df, signature.outputs)
 
                 if target_method == "explain":

snowflake/ml/model/_packager/model_meta/model_meta.py

@@ -55,6 +55,7 @@ def create_model_metadata(
     conda_dependencies: Optional[List[str]] = None,
     pip_requirements: Optional[List[str]] = None,
     python_version: Optional[str] = None,
+    model_objective: model_types.ModelObjective = model_types.ModelObjective.UNKNOWN,
     **kwargs: Any,
 ) -> Generator["ModelMetadata", None, None]:
     """Create a generator for model metadata object. Use generator to ensure correct register and unregister for
@@ -74,6 +75,9 @@ def create_model_metadata(
         pip_requirements: List of pip Python packages requirements for running the model. Defaults to None.
         python_version: A string of python version where model is run. Used for user override. If specified as None,
             current version would be captured. Defaults to None.
+        model_objective: The objective of the Model Version. It is an enum class ModelObjective with values REGRESSION,
+            BINARY_CLASSIFICATION, MULTI_CLASSIFICATION, RANKING, or UNKNOWN. By default it is set to
+            ModelObjective.UNKNOWN and may be overridden by inferring from the Model Object.
         **kwargs: Dict of attributes and values of the metadata. Used when loading from file.
 
     Raises:
@@ -131,6 +135,7 @@ def create_model_metadata(
         model_type=model_type,
         signatures=signatures,
         function_properties=function_properties,
+        model_objective=model_objective,
     )
 
     code_dir_path = os.path.join(model_dir_path, MODEL_CODE_DIR)
@@ -237,6 +242,7 @@ class ModelMetadata:
         function_properties: A dict mapping function names to dict mapping function property key to value.
         metadata: User provided key-value metadata of the model. Defaults to None.
         creation_timestamp: Unix timestamp when the model metadata is created.
+        model_objective: Model objective like regression, classification etc.
     """
 
     def telemetry_metadata(self) -> ModelMetadataTelemetryDict:
@@ -260,6 +266,8 @@ class ModelMetadata:
         min_snowpark_ml_version: Optional[str] = None,
         models: Optional[Dict[str, model_blob_meta.ModelBlobMeta]] = None,
         original_metadata_version: Optional[str] = model_meta_schema.MODEL_METADATA_VERSION,
+        model_objective: model_types.ModelObjective = model_types.ModelObjective.UNKNOWN,
+        explain_algorithm: Optional[model_meta_schema.ModelExplainAlgorithm] = None,
     ) -> None:
         self.name = name
         self.signatures: Dict[str, model_signature.ModelSignature] = dict()
@@ -284,6 +292,9 @@ class ModelMetadata:
 
         self.original_metadata_version = original_metadata_version
 
+        self.model_objective: model_types.ModelObjective = model_objective
+        self.explain_algorithm: Optional[model_meta_schema.ModelExplainAlgorithm] = explain_algorithm
+
     @property
     def min_snowpark_ml_version(self) -> str:
         return self._min_snowpark_ml_version.base_version
@@ -321,9 +332,11 @@ class ModelMetadata:
         model_dict = model_meta_schema.ModelMetadataDict(
             {
                 "creation_timestamp": self.creation_timestamp,
-                "env": self.env.save_as_dict(pathlib.Path(model_dir_path)),
+                "env": self.env.save_as_dict(
+                    pathlib.Path(model_dir_path), default_channel_override=env_utils.SNOWFLAKE_CONDA_CHANNEL_URL
+                ),
                 "runtimes": {
-                    runtime_name: runtime.save(pathlib.Path(model_dir_path))
+                    runtime_name: runtime.save(pathlib.Path(model_dir_path), default_channel_override="conda-forge")
                     for runtime_name, runtime in self.runtimes.items()
                 },
                 "metadata": self.metadata,
@@ -333,6 +346,13 @@ class ModelMetadata:
                 "signatures": {func_name: sig.to_dict() for func_name, sig in self.signatures.items()},
                 "version": model_meta_schema.MODEL_METADATA_VERSION,
                 "min_snowpark_ml_version": self.min_snowpark_ml_version,
+                "model_objective": self.model_objective.value,
+                "explainability": (
+                    model_meta_schema.ExplainabilityMetadataDict(algorithm=self.explain_algorithm.value)
+                    if self.explain_algorithm
+                    else None
+                ),
+                "function_properties": self.function_properties,
             }
         )
 
@@ -370,6 +390,9 @@ class ModelMetadata:
             signatures=loaded_meta["signatures"],
             version=original_loaded_meta_version,
             min_snowpark_ml_version=loaded_meta_min_snowpark_ml_version,
+            model_objective=loaded_meta.get("model_objective", model_types.ModelObjective.UNKNOWN.value),
+            explainability=loaded_meta.get("explainability", None),
+            function_properties=loaded_meta.get("function_properties", {}),
         )
 
     @classmethod
@@ -406,6 +429,11 @@ class ModelMetadata:
         else:
             runtimes = None
 
+        explanation_algorithm_dict = model_dict.get("explainability", None)
+        explanation_algorithm = None
+        if explanation_algorithm_dict:
+            explanation_algorithm = model_meta_schema.ModelExplainAlgorithm(explanation_algorithm_dict["algorithm"])
+
         return cls(
             name=model_dict["name"],
             model_type=model_dict["model_type"],
@@ -417,4 +445,9 @@ class ModelMetadata:
             min_snowpark_ml_version=model_dict["min_snowpark_ml_version"],
             models=models,
             original_metadata_version=model_dict["version"],
+            model_objective=model_types.ModelObjective(
+                model_dict.get("model_objective", model_types.ModelObjective.UNKNOWN.value)
+            ),
+            explain_algorithm=explanation_algorithm,
+            function_properties=model_dict.get("function_properties", {}),
         )

snowflake/ml/model/_packager/model_meta/model_meta_schema.py

@@ -71,6 +71,10 @@ ModelBlobOptions = Union[
 ]
 
 
+class ExplainabilityMetadataDict(TypedDict):
+    algorithm: Required[str]
+
+
 class ModelBlobMetadataDict(TypedDict):
     name: Required[str]
     model_type: Required[type_hints.SupportedModelHandlerType]
@@ -92,3 +96,10 @@ class ModelMetadataDict(TypedDict):
     signatures: Required[Dict[str, Dict[str, Any]]]
     version: Required[str]
     min_snowpark_ml_version: Required[str]
+    model_objective: Required[str]
+    explainability: NotRequired[Optional[ExplainabilityMetadataDict]]
+    function_properties: NotRequired[Dict[str, Dict[str, Any]]]
+
+
+class ModelExplainAlgorithm(Enum):
+    SHAP = "shap"

snowflake/ml/model/_packager/model_packager.py

@@ -47,6 +47,7 @@ class ModelPackager:
         ext_modules: Optional[List[ModuleType]] = None,
         code_paths: Optional[List[str]] = None,
         options: Optional[model_types.ModelSaveOption] = None,
+        model_objective: model_types.ModelObjective = model_types.ModelObjective.UNKNOWN,
     ) -> model_meta.ModelMetadata:
         if (signatures is None) and (sample_input_data is None) and not model_handler.is_auto_signature_model(model):
             raise snowml_exceptions.SnowflakeMLException(
@@ -84,6 +85,7 @@ class ModelPackager:
             conda_dependencies=conda_dependencies,
             pip_requirements=pip_requirements,
             python_version=python_version,
+            model_objective=model_objective,
             **options,
         ) as meta:
             model_blobs_path = os.path.join(self.local_dir_path, ModelPackager.MODEL_BLOBS_DIR)
@@ -146,7 +148,8 @@ class ModelPackager:
         m = handler.load_model(self.meta.name, self.meta, model_blobs_path, **options)
 
         if as_custom_model:
-            m = handler.convert_as_custom_model(m, self.meta, **options)
+            background_data = handler.load_background_data(self.meta.name, model_blobs_path)
+            m = handler.convert_as_custom_model(m, self.meta, background_data, **options)
             assert isinstance(m, custom_model.CustomModel)
 
         self.model = m

snowflake/ml/model/_packager/model_runtime/model_runtime.py

@@ -67,7 +67,9 @@ class ModelRuntime:
     def runtime_rel_path(self) -> pathlib.PurePosixPath:
         return pathlib.PurePosixPath(ModelRuntime.RUNTIME_DIR_REL_PATH) / self.name
 
-    def save(self, packager_path: pathlib.Path) -> model_meta_schema.ModelRuntimeDict:
+    def save(
+        self, packager_path: pathlib.Path, default_channel_override: str = env_utils.SNOWFLAKE_CONDA_CHANNEL_URL
+    ) -> model_meta_schema.ModelRuntimeDict:
         runtime_base_path = packager_path / self.runtime_rel_path
         runtime_base_path.mkdir(parents=True, exist_ok=True)
 
@@ -80,7 +82,7 @@ class ModelRuntime:
         self.runtime_env.conda_env_rel_path = self.runtime_rel_path / self.runtime_env.conda_env_rel_path
         self.runtime_env.pip_requirements_rel_path = self.runtime_rel_path / self.runtime_env.pip_requirements_rel_path
 
-        env_dict = self.runtime_env.save_as_dict(packager_path)
+        env_dict = self.runtime_env.save_as_dict(packager_path, default_channel_override=default_channel_override)
 
         return model_meta_schema.ModelRuntimeDict(
             imports=list(map(str, self.imports)),

snowflake/ml/model/_signatures/pytorch_handler.py

@@ -30,7 +30,7 @@ class SeqOfPyTorchTensorHandler(base_handler.BaseDataHandler[Sequence["torch.Ten
 
     @staticmethod
     def count(data: Sequence["torch.Tensor"]) -> int:
-        return min(data_col.shape[0] for data_col in data)
+        return min(data_col.shape[0] for data_col in data)  # type: ignore[no-any-return]
 
     @staticmethod
     def truncate(data: Sequence["torch.Tensor"]) -> Sequence["torch.Tensor"]: