Merge remote-tracking branch 'upstream/master' into modernize-metadata

Author: ofek

CHANGELOG.md

@@ -1,5 +1,49 @@
 # Changelog
 
+## v2.226.0 (2024-07-12)
+
+### Features
+
+ * Curated hub improvements
+ * InferenceSpec support for MMS and testing
+
+### Bug Fixes and Other Changes
+
+ * ModelBuilder not passing HF_TOKEN to model.
+ * update image_uri_configs  07-10-2024 07:18:04 PST
+
+## v2.225.0 (2024-07-10)
+
+### Features
+
+ * model optimization
+
+### Bug Fixes and Other Changes
+
+ * fix integ test
+ * update uris for v1.1.1
+ * update image_uri_configs  07-04-2024 07:17:24 PST
+
+## v2.224.4 (2024-07-04)
+
+### Bug Fixes and Other Changes
+
+ * allow for inf spec and server override to be passed
+
+## v2.224.3 (2024-07-03)
+
+### Bug Fixes and Other Changes
+
+ * Upgrade local dependencies
+ * Improve docstrings for estimator tags
+
+## v2.224.2 (2024-06-27)
+
+### Bug Fixes and Other Changes
+
+ * Update DJLModel class for latest container releases
+ * list_models() for python3.8
+
 ## v2.224.1 (2024-06-21)
 
 ### Bug Fixes and Other Changes

VERSION

@@ -1 +1 @@
-2.224.2.dev0
+2.226.1.dev0

requirements/extras/local_requirements.txt

@@ -1,3 +1,3 @@
 urllib3>=1.26.8,<3.0.0
-docker>=5.0.2,<7.0.0
+docker>=5.0.2,<8.0.0
 PyYAML>=5.4.1,<7

requirements/extras/test_requirements.txt

@@ -39,3 +39,4 @@ accelerate>=0.24.1,<=0.27.0
 schema==0.7.5
 tensorflow>=2.1,<=2.16
 mlflow>=2.12.2,<2.13
+huggingface_hub>=0.23.4

src/sagemaker/accept_types.py

@@ -82,6 +82,7 @@ def retrieve_default(
     tolerate_deprecated_model: bool = False,
     sagemaker_session: Session = DEFAULT_JUMPSTART_SAGEMAKER_SESSION,
     model_type: JumpStartModelType = JumpStartModelType.OPEN_WEIGHTS,
+    config_name: Optional[str] = None,
 ) -> str:
     """Retrieves the default accept type for the model matching the given arguments.
 
@@ -105,6 +106,7 @@ def retrieve_default(
             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).
+        config_name (Optional[str]): Name of the JumpStart Model config to apply. (Default: None).
     Returns:
         str: The default accept type to use for the model.
 
@@ -125,4 +127,5 @@ def retrieve_default(
         tolerate_deprecated_model=tolerate_deprecated_model,
         sagemaker_session=sagemaker_session,
         model_type=model_type,
+        config_name=config_name,
     )

src/sagemaker/content_types.py

@@ -82,6 +82,7 @@ def retrieve_default(
     tolerate_deprecated_model: bool = False,
     sagemaker_session: Session = DEFAULT_JUMPSTART_SAGEMAKER_SESSION,
     model_type: JumpStartModelType = JumpStartModelType.OPEN_WEIGHTS,
+    config_name: Optional[str] = None,
 ) -> str:
     """Retrieves the default content type for the model matching the given arguments.
 
@@ -105,6 +106,7 @@ def retrieve_default(
             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).
+        config_name (Optional[str]): Name of the JumpStart Model config to apply. (Default: None).
     Returns:
         str: The default content type to use for the model.
 
@@ -125,6 +127,7 @@ def retrieve_default(
         tolerate_deprecated_model=tolerate_deprecated_model,
         sagemaker_session=sagemaker_session,
         model_type=model_type,
+        config_name=config_name,
     )
 
 

src/sagemaker/deserializers.py

@@ -102,6 +102,7 @@ def retrieve_default(
     tolerate_deprecated_model: bool = False,
     sagemaker_session: Session = DEFAULT_JUMPSTART_SAGEMAKER_SESSION,
     model_type: JumpStartModelType = JumpStartModelType.OPEN_WEIGHTS,
+    config_name: Optional[str] = None,
 ) -> BaseDeserializer:
     """Retrieves the default deserializer for the model matching the given arguments.
 
@@ -125,6 +126,7 @@ def retrieve_default(
             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).
+        config_name (Optional[str]): Name of the JumpStart Model config to apply. (Default: None).
     Returns:
         BaseDeserializer: The default deserializer to use for the model.
 
@@ -146,4 +148,5 @@ def retrieve_default(
         tolerate_deprecated_model=tolerate_deprecated_model,
         sagemaker_session=sagemaker_session,
         model_type=model_type,
+        config_name=config_name,
     )

src/sagemaker/enums.py

@@ -40,3 +40,12 @@ class RoutingStrategy(Enum):
     """The endpoint routes requests to the specific instances that have
     more capacity to process them.
     """
+
+
+class Tag(str, Enum):
+    """Enum class for tag keys to apply to models."""
+
+    OPTIMIZATION_JOB_NAME = "sagemaker-sdk:optimization-job-name"
+    SPECULATIVE_DRAFT_MODEL_PROVIDER = "sagemaker-sdk:speculative-draft-model-provider"
+    FINE_TUNING_MODEL_PATH = "sagemaker-sdk:fine-tuning-model-path"
+    FINE_TUNING_JOB_NAME = "sagemaker-sdk:fine-tuning-job-name"

src/sagemaker/environment_variables.py

@@ -37,6 +37,7 @@ def retrieve_default(
     sagemaker_session: Session = DEFAULT_JUMPSTART_SAGEMAKER_SESSION,
     instance_type: Optional[str] = None,
     script: JumpStartScriptScope = JumpStartScriptScope.INFERENCE,
+    config_name: Optional[str] = None,
 ) -> Dict[str, str]:
     """Retrieves the default container environment variables for the model matching the arguments.
 
@@ -68,6 +69,7 @@ def retrieve_default(
             variables specific for the instance type.
         script (JumpStartScriptScope): The JumpStart script for which to retrieve environment
             variables.
+        config_name (Optional[str]): Name of the JumpStart Model config to apply. (Default: None).
     Returns:
         dict: The variables to use for the model.
 
@@ -91,4 +93,5 @@ def retrieve_default(
         sagemaker_session=sagemaker_session,
         instance_type=instance_type,
         script=script,
+        config_name=config_name,
     )

src/sagemaker/estimator.py

@@ -274,7 +274,10 @@ def __init__(
                 AWS services needed. If not specified, the estimator creates one
                 using the default AWS configuration chain.
             tags (Optional[Tags]):
-                Tags for labeling a training job. For more, see
+                Tags for labeling a training job. These won't be propagated to Models,
+                Endpoints during :meth:`~sagemaker.estimator.EstimatorBase.deploy`. The
+                :meth:`~sagemaker.estimator.EstimatorBase.deploy` takes in a seperate
+                tags parameter. For more on tags, see
                 https://docs.aws.amazon.com/sagemaker/latest/dg/API_Tag.html.
             subnets (list[str] or list[PipelineVariable]): List of subnet ids. If not
                 specified training job will be created without VPC config.