fix: minor jumpstart dev ex improvements

src/sagemaker/chainer/model.py

@@ -261,7 +261,11 @@ def register(
         )
 
     def prepare_container_def(
-        self, instance_type=None, accelerator_type=None, serverless_inference_config=None
+        self,
+        instance_type=None,
+        accelerator_type=None,
+        serverless_inference_config=None,
+        accept_eula=None,
     ):
         """Return a container definition with framework configuration set in model environment.
 
@@ -274,6 +278,11 @@ def prepare_container_def(
             serverless_inference_config (sagemaker.serverless.ServerlessInferenceConfig):
                 Specifies configuration related to serverless endpoint. Instance type is
                 not provided in serverless inference. So this is used to find image URIs.
+            accept_eula (bool): For models that require a Model Access Config, specify True or
+                False to indicate whether model terms of use have been accepted.
+                The `accept_eula` value must be explicitly defined as `True` in order to
+                accept the end-user license agreement (EULA) that some
+                models require. (Default: None).
 
         Returns:
             dict[str, str]: A container definition object usable with the
@@ -303,7 +312,12 @@ def prepare_container_def(
             deploy_env[MODEL_SERVER_WORKERS_PARAM_NAME.upper()] = to_string(
                 self.model_server_workers
             )
-        return sagemaker.container_def(deploy_image, self.model_data, deploy_env)
+        return sagemaker.container_def(
+            deploy_image,
+            self.model_data,
+            deploy_env,
+            accept_eula=accept_eula,
+        )
 
     def serving_image_uri(
         self, region_name, instance_type, accelerator_type=None, serverless_inference_config=None

src/sagemaker/djl_inference/model.py

@@ -717,6 +717,7 @@ def prepare_container_def(
         instance_type=None,
         accelerator_type=None,
         serverless_inference_config=None,
+        accept_eula=None,
     ):  # pylint: disable=unused-argument
         """A container definition with framework configuration set in model environment variables.
 

src/sagemaker/huggingface/model.py

@@ -457,6 +457,7 @@ def prepare_container_def(
         accelerator_type=None,
         serverless_inference_config=None,
         inference_tool=None,
+        accept_eula=None,
     ):
         """A container definition with framework configuration set in model environment variables.
 
@@ -471,6 +472,11 @@ def prepare_container_def(
                 not provided in serverless inference. So this is used to find image URIs.
             inference_tool (str): the tool that will be used to aid in the inference.
                 Valid values: "neuron, neuronx, None" (default: None).
+            accept_eula (bool): For models that require a Model Access Config, specify True or
+                False to indicate whether model terms of use have been accepted.
+                The `accept_eula` value must be explicitly defined as `True` in order to
+                accept the end-user license agreement (EULA) that some
+                models require. (Default: None).
 
         Returns:
             dict[str, str]: A container definition object usable with the
@@ -502,7 +508,10 @@ def prepare_container_def(
                 self.model_server_workers
             )
         return sagemaker.container_def(
-            deploy_image, self.repacked_model_data or self.model_data, deploy_env
+            deploy_image,
+            self.repacked_model_data or self.model_data,
+            deploy_env,
+            accept_eula=accept_eula,
         )
 
     def serving_image_uri(

src/sagemaker/jumpstart/exceptions.py

@@ -118,8 +118,9 @@ def __init__(
                 self.message = (
                     f"Version '{version}' of JumpStart model '{model_id}' "  # type: ignore
                     "has at least 1 vulnerable dependency in the training script. "
-                    "Please try targeting a higher version of the model or using a "
-                    "different model. List of vulnerabilities: "
+                    "Please try targeting a higher version of the model, upgrading "
+                    "SageMaker Python SDK version or using a different model. "
+                    "List of vulnerabilities: "
                     f"{', '.join(vulnerabilities)}"  # type: ignore
                 )
             else:
@@ -154,8 +155,8 @@ def __init__(
                 raise RuntimeError("Must specify `model_id` and `version` arguments.")
             self.message = (
                 f"Version '{version}' of JumpStart model '{model_id}' is deprecated. "
-                "Please try targeting a higher version of the model or using a "
-                "different model."
+                "Please try targeting a higher version of the model, upgrading "
+                "SageMaker Python SDK version or using a different model."
             )
 
         super().__init__(self.message)

src/sagemaker/jumpstart/payload_utils.py

@@ -115,80 +115,6 @@ def _construct_payload(
     return payload_to_use
 
 
-def _extract_generated_text_from_response(
-    response: dict,
-    model_id: str,
-    model_version: str,
-    region: Optional[str] = None,
-    tolerate_vulnerable_model: bool = False,
-    tolerate_deprecated_model: bool = False,
-    sagemaker_session: Session = DEFAULT_JUMPSTART_SAGEMAKER_SESSION,
-    accept_type: Optional[str] = None,
-) -> str:
-    """Returns generated text extracted from full response payload.
-
-    Args:
-        response (dict): Dictionary-valued response from which to extract
-            generated text.
-        model_id (str): JumpStart model ID of the JumpStart model from which to extract
-            generated text.
-        model_version (str): Version of the JumpStart model for which to extract generated
-            text.
-        region (Optional[str]): Region for which to extract generated
-            text. (Default: None).
-        tolerate_vulnerable_model (bool): True if vulnerable versions of model
-            specifications should be tolerated (exception not raised). If False, raises an
-            exception if the script used by this version of the model has dependencies with known
-            security vulnerabilities. (Default: False).
-        tolerate_deprecated_model (bool): True if deprecated versions of model
-            specifications should be tolerated (exception not raised). If False, raises
-            an exception if the version of the model is deprecated. (Default: False).
-        sagemaker_session (sagemaker.session.Session): A SageMaker Session
-            object, used for SageMaker interactions. If not
-            specified, one is created using the default AWS configuration
-            chain. (Default: sagemaker.jumpstart.constants.DEFAULT_JUMPSTART_SAGEMAKER_SESSION).
-        accept_type (Optional[str]): The accept type to optionally specify for the response.
-            (Default: None).
-
-    Returns:
-        str: extracted generated text from the endpoint response payload.
-
-    Raises:
-        ValueError: If the model is invalid, the model does not support generated text extraction,
-        or if the response is malformed.
-    """
-
-    if not isinstance(response, dict):
-        raise ValueError(f"Response must be dictionary. Instead, got: {type(response)}")
-
-    payloads: Optional[Dict[str, JumpStartSerializablePayload]] = _retrieve_example_payloads(
-        model_id=model_id,
-        model_version=model_version,
-        region=region,
-        tolerate_vulnerable_model=tolerate_vulnerable_model,
-        tolerate_deprecated_model=tolerate_deprecated_model,
-        sagemaker_session=sagemaker_session,
-    )
-    if payloads is None or len(payloads) == 0:
-        raise ValueError(f"Model ID '{model_id}' does not support generated text extraction.")
-
-    for payload in payloads.values():
-        if accept_type is None or payload.accept == accept_type:
-            generated_text_response_key: Optional[str] = payload.generated_text_response_key
-            if generated_text_response_key is None:
-                raise ValueError(
-                    f"Model ID '{model_id}' does not support generated text extraction."
-                )
-
-            generated_text_response_key_split = generated_text_response_key.split(".")
-            try:
-                return _extract_field_from_json(response, generated_text_response_key_split)
-            except KeyError:
-                raise ValueError(f"Response is malformed: {response}")
-
-    raise ValueError(f"Model ID '{model_id}' does not support generated text extraction.")
-
-
 class PayloadSerializer:
     """Utility class for serializing payloads associated with JumpStart models.
 

src/sagemaker/jumpstart/types.py

@@ -339,7 +339,6 @@ class JumpStartSerializablePayload(JumpStartDataHolderType):
         "content_type",
         "accept",
         "body",
-        "generated_text_response_key",
         "prompt_key",
     ]
 
@@ -371,7 +370,6 @@ def from_json(self, json_obj: Optional[Dict[str, Any]]) -> None:
         self.content_type = json_obj["content_type"]
         self.body = json_obj["body"]
         accept = json_obj.get("accept")
-        self.generated_text_response_key = json_obj.get("generated_text_response_key")
         self.prompt_key = json_obj.get("prompt_key")
         if accept:
             self.accept = accept

src/sagemaker/jumpstart/utils.py

@@ -109,7 +109,8 @@ def get_jumpstart_gated_content_bucket(
     accessors.JumpStartModelsAccessor.set_jumpstart_gated_content_bucket(gated_bucket_to_return)
 
     if gated_bucket_to_return != old_gated_content_bucket:
-        accessors.JumpStartModelsAccessor.reset_cache()
+        if old_gated_content_bucket is not None:
+            accessors.JumpStartModelsAccessor.reset_cache()
         for info_log in info_logs:
             constants.JUMPSTART_LOGGER.info(info_log)
 
@@ -153,7 +154,8 @@ def get_jumpstart_content_bucket(
     accessors.JumpStartModelsAccessor.set_jumpstart_content_bucket(bucket_to_return)
 
     if bucket_to_return != old_content_bucket:
-        accessors.JumpStartModelsAccessor.reset_cache()
+        if old_content_bucket is not None:
+            accessors.JumpStartModelsAccessor.reset_cache()
         for info_log in info_logs:
             constants.JUMPSTART_LOGGER.info(info_log)
     return bucket_to_return

src/sagemaker/model.py

@@ -547,6 +547,7 @@ def create(
         accelerator_type: Optional[str] = None,
         serverless_inference_config: Optional[ServerlessInferenceConfig] = None,
         tags: Optional[List[Dict[str, Union[str, PipelineVariable]]]] = None,
+        accept_eula: Optional[bool] = None,
     ):
         """Create a SageMaker Model Entity
 
@@ -570,6 +571,11 @@ def create(
                 For more information about tags, see
                 `boto3 documentation <https://boto3.amazonaws.com/v1/documentation/\
 api/latest/reference/services/sagemaker.html#SageMaker.Client.add_tags>`_
+            accept_eula (bool): For models that require a Model Access Config, specify True or
+                False to indicate whether model terms of use have been accepted.
+                The `accept_eula` value must be explicitly defined as `True` in order to
+                accept the end-user license agreement (EULA) that some
+                models require. (Default: None).
 
         Returns:
             None or pipeline step arguments in case the Model instance is built with
@@ -581,6 +587,7 @@ def create(
             accelerator_type=accelerator_type,
             tags=tags,
             serverless_inference_config=serverless_inference_config,
+            accept_eula=accept_eula,
         )
 
     def _init_sagemaker_session_if_does_not_exist(self, instance_type=None):
@@ -601,6 +608,7 @@ def prepare_container_def(
         instance_type=None,
         accelerator_type=None,
         serverless_inference_config=None,
+        accept_eula=None,
     ):  # pylint: disable=unused-argument
         """Return a dict created by ``sagemaker.container_def()``.
 
@@ -618,6 +626,11 @@ def prepare_container_def(
             serverless_inference_config (sagemaker.serverless.ServerlessInferenceConfig):
                 Specifies configuration related to serverless endpoint. Instance type is
                 not provided in serverless inference. So this is used to find image URIs.
+            accept_eula (bool): For models that require a Model Access Config, specify True or
+                False to indicate whether model terms of use have been accepted.
+                The `accept_eula` value must be explicitly defined as `True` in order to
+                accept the end-user license agreement (EULA) that some
+                models require. (Default: None).
 
         Returns:
             dict: A container definition object usable with the CreateModel API.
@@ -635,7 +648,7 @@ def prepare_container_def(
             self.repacked_model_data or self.model_data,
             deploy_env,
             image_config=self.image_config,
-            accept_eula=getattr(self, "accept_eula", None),
+            accept_eula=accept_eula or getattr(self, "accept_eula", None),
         )
 
     def is_repack(self) -> bool:
@@ -777,6 +790,7 @@ def _create_sagemaker_model(
         accelerator_type=None,
         tags=None,
         serverless_inference_config=None,
+        accept_eula=None,
     ):
         """Create a SageMaker Model Entity
 
@@ -796,6 +810,11 @@ def _create_sagemaker_model(
             serverless_inference_config (sagemaker.serverless.ServerlessInferenceConfig):
                 Specifies configuration related to serverless endpoint. Instance type is
                 not provided in serverless inference. So this is used to find image URIs.
+            accept_eula (bool): For models that require a Model Access Config, specify True or
+                False to indicate whether model terms of use have been accepted.
+                The `accept_eula` value must be explicitly defined as `True` in order to
+                accept the end-user license agreement (EULA) that some
+                models require. (Default: None).
         """
 
         if self.model_package_arn is not None or self.algorithm_arn is not None:
@@ -827,6 +846,7 @@ def _create_sagemaker_model(
                 instance_type,
                 accelerator_type=accelerator_type,
                 serverless_inference_config=serverless_inference_config,
+                accept_eula=accept_eula,
             )
 
             if not isinstance(self.sagemaker_session, PipelineSession):
@@ -1434,7 +1454,12 @@ def deploy(
                 "serverless_inference_config needs to be a ServerlessInferenceConfig object"
             )
 
-        if instance_type and instance_type.startswith("ml.inf") and not self._is_compiled_model:
+        if (
+            getattr(self, "model_id", None) in {"", None}
+            and instance_type
+            and instance_type.startswith("ml.inf")
+            and not self._is_compiled_model
+        ):
             LOGGER.warning(
                 "Your model is not compiled. Please compile your model before using Inferentia."
             )

src/sagemaker/multidatamodel.py

@@ -121,7 +121,11 @@ def __init__(
             )
 
     def prepare_container_def(
-        self, instance_type=None, accelerator_type=None, serverless_inference_config=None
+        self,
+        instance_type=None,
+        accelerator_type=None,
+        serverless_inference_config=None,
+        accept_eula=None,
     ):
         """Return a container definition set.
 
@@ -149,6 +153,7 @@ def prepare_container_def(
             env=environment,
             model_data_url=self.model_data_prefix,
             container_mode=self.container_mode,
+            accept_eula=accept_eula,
         )
 
     def deploy(

src/sagemaker/mxnet/model.py

@@ -263,7 +263,11 @@ def register(
         )
 
     def prepare_container_def(
-        self, instance_type=None, accelerator_type=None, serverless_inference_config=None
+        self,
+        instance_type=None,
+        accelerator_type=None,
+        serverless_inference_config=None,
+        accept_eula=None,
     ):
         """Return a container definition with framework configuration.
 
@@ -278,6 +282,11 @@ def prepare_container_def(
             serverless_inference_config (sagemaker.serverless.ServerlessInferenceConfig):
                 Specifies configuration related to serverless endpoint. Instance type is
                 not provided in serverless inference. So this is used to find image URIs.
+            accept_eula (bool): For models that require a Model Access Config, specify True or
+                False to indicate whether model terms of use have been accepted.
+                The `accept_eula` value must be explicitly defined as `True` in order to
+                accept the end-user license agreement (EULA) that some
+                models require. (Default: None).
 
         Returns:
             dict[str, str]: A container definition object usable with the
@@ -308,7 +317,10 @@ def prepare_container_def(
                 self.model_server_workers
             )
         return sagemaker.container_def(
-            deploy_image, self.repacked_model_data or self.model_data, deploy_env
+            deploy_image,
+            self.repacked_model_data or self.model_data,
+            deploy_env,
+            accept_eula=accept_eula,
         )
 
     def serving_image_uri(